[libcamera-devel] [PATCH 1/7] utils: tuning: libtuning: Implement the core of libtuning
Laurent Pinchart
laurent.pinchart at ideasonboard.com
Fri Oct 14 01:49:07 CEST 2022
Hi Paul,
Thank you for the patch.
On Thu, Oct 06, 2022 at 09:00:59PM +0900, Paul Elder via libcamera-devel wrote:
> Implement the core of libtuning, our new tuning tool infrastructure. It
> leverages components from raspberrypi's ctt that could be reused for
raspberrypi should be written Raspberry Pi.
> tuning tools for other platforms.
>
> Signed-off-by: Paul Elder <paul.elder at ideasonboard.com>
> ---
> utils/tuning/libtuning/__init__.py | 9 +
> utils/tuning/libtuning/average_functions.py | 21 +
> utils/tuning/libtuning/generators/__init__.py | 0
> .../tuning/libtuning/generators/generator.py | 12 +
> utils/tuning/libtuning/gradient.py | 111 +++
> utils/tuning/libtuning/image.py | 272 ++++++++
> utils/tuning/libtuning/libtuning.py | 191 +++++
> utils/tuning/libtuning/macbeth.py | 654 ++++++++++++++++++
> utils/tuning/libtuning/macbeth_ref.pgm | 5 +
> utils/tuning/libtuning/modules/__init__.py | 0
> utils/tuning/libtuning/modules/module.py | 41 ++
> utils/tuning/libtuning/parsers/__init__.py | 0
> utils/tuning/libtuning/parsers/parser.py | 18 +
> utils/tuning/libtuning/smoothing.py | 21 +
> utils/tuning/libtuning/utils.py | 198 ++++++
> 15 files changed, 1553 insertions(+)
OK, this will be a fun review :-)
I assume the code was copied from ctt and then refactored. I won't ask
you to split that in two (or more) commits as that would be painful to
do. It would make the refactoring easier to review, but forcing me to
review the code anew has advantages too, as I haven't read it in details
yet.
Nonetheless, could you at least shortly describe in the commit message
how loosely or tightly related this patch and the current implementation
in ctt are ?
Another thing I'm curious about is if you envision moving ctt to
libtuning. Patches 3/7 to 6/7 show how support for the Raspberry Pi IPA
module can be implemented, so I assume you have considered this could be
a way forward, but could you explain your opinion with a bit more
details ?
> create mode 100644 utils/tuning/libtuning/__init__.py
> create mode 100644 utils/tuning/libtuning/average_functions.py
> create mode 100644 utils/tuning/libtuning/generators/__init__.py
> create mode 100644 utils/tuning/libtuning/generators/generator.py
> create mode 100644 utils/tuning/libtuning/gradient.py
> create mode 100644 utils/tuning/libtuning/image.py
> create mode 100644 utils/tuning/libtuning/libtuning.py
> create mode 100644 utils/tuning/libtuning/macbeth.py
> create mode 100644 utils/tuning/libtuning/macbeth_ref.pgm
> create mode 100644 utils/tuning/libtuning/modules/__init__.py
> create mode 100644 utils/tuning/libtuning/modules/module.py
> create mode 100644 utils/tuning/libtuning/parsers/__init__.py
> create mode 100644 utils/tuning/libtuning/parsers/parser.py
> create mode 100644 utils/tuning/libtuning/smoothing.py
> create mode 100644 utils/tuning/libtuning/utils.py
Could you add a README.md file in utils/tuning/ to list the dependencies
? So far I have noticed cv2, numpy, pyexiv2 and rawpy. The last two are
not packaged by most distributions, so I think I'll work on replacing
them at some point, but there's no urgency. They're nicely isolved in
libtuning/image.py, let's make sure they are not exposed outside of that
file.
Update: I had a closer look at rawpy and pyexiv2. They respectively wrap
libraw and libexiv2, which are standard dependencies. libraw is a bit
overkill if all we need is reading the raw data from DNG files, but it
does the job. pyexiv2 bothers me a bit more, while it is also a simple
wrapper around exiv2, the fact that the upstream git tree contains
pre-compiled binaries instead of wiring up the compilation of the
corresponding .cpp file (which is also in the git tree) through setup.py
makes me question the practices followed by the project.
Anyway, a possible replacement is something we'll look at later. It's
annoying that there's no defacto standard Python library for this task,
and writing our own will likely not solve the fragmentation issue :-)
> diff --git a/utils/tuning/libtuning/__init__.py b/utils/tuning/libtuning/__init__.py
> new file mode 100644
> index 00000000..63f3c8f9
> --- /dev/null
> +++ b/utils/tuning/libtuning/__init__.py
> @@ -0,0 +1,9 @@
Please add an SPDX header to every file. A copyright notice should also
be added.
> +from libtuning.utils import *
> +from libtuning.libtuning import *
> +
> +from libtuning.image import *
> +from libtuning.macbeth import *
> +
> +from libtuning.average_functions import *
> +from libtuning.gradient import *
> +from libtuning.smoothing import *
Do we want to alias all the contents of all modules in the libtuning
namespace ? Wouldn't it be cleaner to keep the module hierarchy ?
> diff --git a/utils/tuning/libtuning/average_functions.py b/utils/tuning/libtuning/average_functions.py
> new file mode 100644
> index 00000000..4220e481
> --- /dev/null
> +++ b/utils/tuning/libtuning/average_functions.py
The two other similar files are called gradient.py and smoothing.py, I
would call this average.py.
> @@ -0,0 +1,21 @@
> +import libtuning as lt
Not used.
In the comment block that will contain the SPDX tag and copyright
header, a one line (or one sentence) description of what the file
contains would be helpful.
> +
> +import numpy as np
> +
> +
> +# @brief Wrapper for np averaging functions so that they can be duck-typed
> +class Average(object):
> + def __init__(self, params: list = []):
> + self.params = params
> + return
> +
> + def __average__(self, np_array):
https://docs.python.org/3/reference/lexical_analysis.html#reserved-classes-of-identifiers:
Any use of __*__ names, in any context, that does not follow
explicitly documented use, is subject to breakage without warning.
Let's not risk that.
> + raise NotImplementedError
> +
> + def average(self, np_array):
> + return self.__average__(np_array)
> +
> +
> +class Mean(Average):
> + def __average__(self, np_array):
> + return np.mean(np_array)
> diff --git a/utils/tuning/libtuning/generators/__init__.py b/utils/tuning/libtuning/generators/__init__.py
> new file mode 100644
> index 00000000..e69de29b
> diff --git a/utils/tuning/libtuning/generators/generator.py b/utils/tuning/libtuning/generators/generator.py
> new file mode 100644
> index 00000000..51dd03de
> --- /dev/null
> +++ b/utils/tuning/libtuning/generators/generator.py
> @@ -0,0 +1,12 @@
> +from pathlib import Path
> +
> +
> +class Generator(object):
> + def __init__(self):
> + return
> +
> + def __write__(self, output_file: Path, output_dict: dict, output_order: list):
> + raise NotImplementedError
> +
> + def write(self, output_path: str, output_dict: dict, output_order: list):
> + return self.__write__(Path(output_path), output_dict, output_order)
> diff --git a/utils/tuning/libtuning/gradient.py b/utils/tuning/libtuning/gradient.py
> new file mode 100644
> index 00000000..ebf4f20e
> --- /dev/null
> +++ b/utils/tuning/libtuning/gradient.py
> @@ -0,0 +1,111 @@
> +import libtuning as lt
> +
> +import math
> +
> +
> +# @brief Gradient for how to allocate pixels to sectors
> +# @description There are no parameters for the gradients as the domain is the
> +# number of pixels and the range is the number of sectors, and
> +# there is only one curve that has a startpoint and endpoint at
> +# (0, 0) and at (#pixels, #sectors). The exception is for curves
> +# that *do* have multiple solutions for only two points, such as
> +# gaussian, and curves of higher polynomial orders if we had them.
> +#
> +# todo There will probably be a helper in the Gradient class, as I have a
> +# feeling that all the other curves (besides Linear and Gaussian) can be
> +# implemented in the same way.
> +class Gradient(object):
> + # @param remainder Mode of handling remainder
> + def __init__(self, remainder: lt.remainder = lt.remainder.DistributeFront):
> + self.remainder = remainder
> + return
> +
> + # @brief Distribute pixels into sectors (only in one dimension)
> + # @param domain Number of pixels
> + # @param sectors Number of sectors
> + # @return A list of number of pixels in each sector
> + def __distribute__(self, domain: list, sectors: list) -> list:
> + raise NotImplementedError
> +
> + def distribute(self, domain: list, sectors: list, ) -> list:
> + return self.__distribute__(domain, sectors)
> +
> +
> +class Circular(Gradient):
> + def __distribute__(self, domain, sectors):
> + raise NotImplementedError
> +
> +
> +class Exponential(Gradient):
> + def __distribute__(self, domain, sectors):
> + raise NotImplementedError
> +
> +
> +class Gaussian(Gradient):
> + def __distribute__(self, domain, sectors):
> + raise NotImplementedError
> +
> +
> +class Hyperbolic(Gradient):
> + def __distribute__(self, domain, sectors):
> + raise NotImplementedError
> +
> +
> +class Linear(Gradient):
> + def __distribute__(self, domain, sectors):
> + size = domain / sectors
> + rem = domain % sectors
> +
> + if rem == 0:
> + return [int(size) for i in range(sectors)]
> +
> + size = math.ceil(size)
> + rem = domain % size
> + output_sectors = [int(size) for i in range(sectors - 1)]
> +
> + # Not sure if there's a use case for the first two, and even for
> + # the next two, not sure what to do because we have to shrink the
> + # number of sectors for the divisible ones, and then put remainder
> + # into the remaining sectors, but what if it divides nicely into
> + # the smaller number of sectors? Then the sectors for the remainder
> + # pixels will be empty. I'm leaving them unimplemented for now. Or
> + # we can remove them if we don't think they're necessary.
> +
> + # Also not sure if there's a use case for the last two, since
> + # distributing the remaining pixels means that only one sector will
> + # be smaller than all the rest which will be the same size, so we
> + # can't actually split it between *two* sectors.
> +
> + # If we eliminate all six of these cases we could use a simpler
> + # parameter as opposed to an entire enum.
> +
> + if self.remainder == lt.remainder.Append:
> + raise NotImplementedError
> + elif self.remainder == lt.remainder.Prepend:
> + raise NotImplementedError
> + elif self.remainder == lt.remainder.Midpend:
> + raise NotImplementedError
> + elif self.remainder == lt.remainder.Endpend:
> + raise NotImplementedError
> + elif self.remainder == lt.remainder.DistributeFront:
> + output_sectors.append(rem)
> + elif self.remainder == lt.remainder.DistributeBack:
> + output_sectors.insert(0, rem)
> + elif self.remainder == lt.remainder.DistributeMiddle:
> + raise NotImplementedError
> + elif self.remainder == lt.remainder.DistributeEdges:
> + raise NotImplementedError
> + else:
> + raise ValueError
> +
> + return output_sectors
> +
> +
> +class Logarithmic(Gradient):
> + def __distribute__(self, domain, sectors):
> + raise NotImplementedError
> +
> +
> +class Parabolic(Gradient):
> + def __distribute__(self, domain, sectors):
> + raise NotImplementedError
> diff --git a/utils/tuning/libtuning/image.py b/utils/tuning/libtuning/image.py
> new file mode 100644
> index 00000000..51185a69
> --- /dev/null
> +++ b/utils/tuning/libtuning/image.py
> @@ -0,0 +1,272 @@
> +import binascii
> +import numpy as np
> +from pathlib import Path
> +import pyexiv2 as pyexif
> +import rawpy as raw
> +import re
> +
> +import libtuning.utils as utils
> +
> +
> +class Image:
> + def __init__(self, path: Path):
> + self.path = path
> + self.name = path.name
> + self.alsc_only = False
> + self.color = -1
> + self.lux = -1
> +
> + # May raise KeyError as there are too many to check
> + def _loadMetadataExif(self):
The Python coding style recommends snake_case.
> + # RawPy doesn't load all the image tags that we need, so we use py3exiv2
> + metadata = pyexif.ImageMetadata(self.path)
> + metadata.read()
> +
> + self.ver = 100 # random value
> + # The DNG and TIFF/EP specifications use different IFDs to store the
> + # raw image data and the Exif tags. DNG stores them in a SubIFD and in
> + # an Exif IFD respectively (named "SubImage1" and "Photo" by pyexiv2),
> + # while TIFF/EP stores them both in IFD0 (name "Image"). Both are used
> + # in "DNG" files, with libcamera-apps following the DNG recommendation
> + # and applications based on picamera2 following TIFF/EP.
> + #
> + # This code detects which tags are being used, and therefore extracts the
> + # correct values.
> + try:
> + self.w = metadata['Exif.SubImage1.ImageWidth'].value
> + subimage = "SubImage1"
> + photo = "Photo"
We tend to standardize on single quotes for strings in our Python code
base.
> + except KeyError:
> + self.w = metadata['Exif.Image.ImageWidth'].value
> + subimage = "Image"
> + photo = "Image"
> + self.pad = 0
> + self.h = metadata[f'Exif.{subimage}.ImageLength'].value
> + white = metadata[f'Exif.{subimage}.WhiteLevel'].value
> + self.sigbits = int(white).bit_length()
> + self.fmt = (self.sigbits - 4) // 2
> + self.exposure = int(metadata[f'Exif.{photo}.ExposureTime'].value * 1000000)
> + self.againQ8 = metadata[f'Exif.{photo}.ISOSpeedRatings'].value * 256 / 100
> + self.againQ8_norm = self.againQ8 / 256
> + self.camName = metadata['Exif.Image.Model'].value
> + self.blacklevel = int(metadata[f'Exif.{subimage}.BlackLevel'].value[0])
> + self.blacklevel_16 = self.blacklevel << (16 - self.sigbits)
> +
> + # Channel order depending on bayer pattern
> + # The key is the order given by exif, where 0 is R, 1 is G, and 2 is B
> + # The second value of the value is the index where the color can be
> + # found, where the first is R, then G, then G, then B.
> + # The first value of the value is probably just for consistency with
> + # the brcm loader.
> + bayer_case = {
> + '0 1 1 2': (0, (lt.color.R, lt.color.GR, lt.color.GB, lt.color.B)),
> + '1 2 0 1': (1, (lt.color.GB, lt.color.R, lt.color.B, lt.color.GR)),
> + '2 1 1 0': (2, (lt.color.B, lt.color.GB, lt.color.GR, lt.color.R)),
> + '1 0 2 1': (3, (lt.color.GR, lt.color.R, lt.color.B, lt.color.GB))
> + }
> + cfa_pattern = metadata[f'Exif.{subimage}.CFAPattern'].value
> + self.pattern = bayer_case[cfa_pattern][0]
> + self.order = bayer_case[cfa_pattern][1]
The order is only used to index self.channels (in alsc.py only at the
moment). Wouldn't it be better to reorder the channels instead so that
self.channels will always be in the same order ?
> +
> + def _readImageDng(self):
> + raw_im = raw.imread(str(self.path))
> + raw_data = raw_im.raw_image
> + shift = 16 - self.sigbits
> + c0 = np.left_shift(raw_data[0::2, 0::2].astype(np.int64), shift)
> + c1 = np.left_shift(raw_data[0::2, 1::2].astype(np.int64), shift)
> + c2 = np.left_shift(raw_data[1::2, 0::2].astype(np.int64), shift)
> + c3 = np.left_shift(raw_data[1::2, 1::2].astype(np.int64), shift)
> + self.channels = [c0, c1, c2, c3]
> +
> + def loadDng(self):
> + try:
> + self._loadMetadataExif()
> + except Exception as e:
> + utils.eprint(f'Failed to load metadata from {self.path}: {e}')
> + return False
> +
> + try:
> + self._readImageDng()
> + except Exception as e:
> + utils.eprint(f'Failed to load image data from {self.path}: {e}')
> + return False
> +
> + return True
> +
> + @staticmethod
> + def baToByte(ba):
> + total = 0
> + for i in range(len(ba)):
> + total += 256**i * b[i]
> + return total
> +
> + def _loadMetadataBrcm(self, buf):
> + self.ver = baToByte(buf[4:5])
> + self.w = baToByte(buf[0xd0:0xd2])
> + self.h = baToByte(buf[0xd2:0xd4])
> + self.pad = baToByte(buf[0xd4:0xd6])
> + self.fmt = buf[0xf5]
> + self.sigbits = 2 * self.fmt + 4
> + self.pattern = buf[0xf4]
> + self.exposure = baToByte(buf[0x90:0x94])
> + self.againQ8 = baToByte(buf[0x94:0x96])
> + self.againQ8_norm = self.againQ8 / 256
> + camName = buf[0x10:0x10 + 128]
> + camName_end = camName.find(0x00)
> + self.camName = buf[0x10:0x10 + 128][:camName_end].decode()
> +
> + bayer_case = {
> + 0: (lt.color.R, lt.color.GR, lt.color.GB, lt.color.B),
> + 1: (lt.color.GB, lt.color.R, lt.color.B, lt.color.GR),
> + 2: (lt.color.B, lt.color.GB, lt.color.GR, lt.color.R),
> + 3: (lt.color.GR, lt.color.R, lt.color.B, lt.color.GB),
> + # arbitrary order for greyscale casw
> + 128: (lt.color.R, lt.color.GR, lt.color.GB, lt.color.B)
> + }
> + self.order = bayer_case[self.pattern]
> +
> + # manual blacklevel - not robust
> + if 'ov5647' in self.camName:
> + self.blacklevel = 16
> + else:
> + self.blacklevel = 64
> + self.blacklevel_16 = self.blacklevel << (6)
> +
> + def _readImageBrcm(self, buf):
> + # Check if data is 10 or 12 bits
> + if self.sigbits == 10:
> + # Calculate length of scanline
> + lin_len = ((((((self.w + self.pad + 3) >> 2)) * 5) + 31) >> 5) * 32
> +
> + # Stack scan lines into matrix
> + raw = np.array(raw).reshape(-1, lin_len).astype(np.int64)[:self.h, ...]
This won't compile, the function parameter is called buf, not raw.
I think I would drop support for BRCM images, that's legacy code for
Raspberry Pi. Let's use DNG only going forward.
> +
> + # Separate 5 bits in each package, stopping when w is satisfied
> + ba0 = raw[..., 0:5 * ((self.w + 3) >> 2):5]
> + ba1 = raw[..., 1:5 * ((self.w + 3) >> 2):5]
> + ba2 = raw[..., 2:5 * ((self.w + 3) >> 2):5]
> + ba3 = raw[..., 3:5 * ((self.w + 3) >> 2):5]
> + ba4 = raw[..., 4:5 * ((self.w + 3) >> 2):5]
> +
> + # Assemble 10 bit numbers
> + ch0 = np.left_shift((np.left_shift(ba0, 2) + (ba4 % 4)), 6)
> + ch1 = np.left_shift((np.left_shift(ba1, 2) + (np.right_shift(ba4, 2) % 4)), 6)
> + ch2 = np.left_shift((np.left_shift(ba2, 2) + (np.right_shift(ba4, 4) % 4)), 6)
> + ch3 = np.left_shift((np.left_shift(ba3, 2) + (np.right_shift(ba4, 6) % 4)), 6)
> +
> + # Interleave bits
> + mat = np.empty((self.h, self.w), dtype=ch0.dtype)
> + mat[..., 0::4] = ch0
> + mat[..., 1::4] = ch1
> + mat[..., 2::4] = ch2
> + mat[..., 3::4] = ch3
> +
> + # There is some leaking memory somewhere in the code. This code
> + # here seemed to make things good enough that the code would run
> + # for reasonable numbers of images, however this is techincally
> + # just a workaround. (sorry)
> + ba0, ba1, ba2, ba3, ba4 = None, None, None, None, None
> + del ba0, ba1, ba2, ba3, ba4
> + ch0, ch1, ch2, ch3 = None, None, None, None
> + del ch0, ch1, ch2, ch3
> +
> + # Same as before but 12 bit case
> + elif self.sigbits == 12:
> + lin_len = ((((((self.w + self.pad + 1) >> 1)) * 3) + 31) >> 5) * 32
> + raw = np.array(raw).reshape(-1, lin_len).astype(np.int64)[:self.h, ...]
> + ba0 = raw[..., 0:3 * ((self.w + 1) >> 1):3]
> + ba1 = raw[..., 1:3 * ((self.w + 1) >> 1):3]
> + ba2 = raw[..., 2:3 * ((self.w + 1) >> 1):3]
> + ch0 = np.left_shift((np.left_shift(ba0, 4) + ba2 % 16), 4)
> + ch1 = np.left_shift((np.left_shift(ba1, 4) + (np.right_shift(ba2, 4)) % 16), 4)
> + mat = np.empty((self.h, self.w), dtype=ch0.dtype)
> + mat[..., 0::2] = ch0
> + mat[..., 1::2] = ch1
> +
> + else:
> + raise ValueError('BRCM image data must be 10 bit or 12 bits')
> +
> + # Separate bayer channels
> + c0 = mat[0::2, 0::2]
> + c1 = mat[0::2, 1::2]
> + c2 = mat[1::2, 0::2]
> + c3 = mat[1::2, 1::2]
> + self.channels = [c0, c1, c2, c3]
> +
> + def loadBrcm(self):
> + try:
> + with open(self.path, 'rb') as image:
> + f = image.read()
> + except FileNotFoundError:
> + utils.eprint(f'File {self.path} not found')
> + return False
> +
> + if f is None:
> + utils.eprint(f'Failed to open {self.path}')
> + return False
> +
> + b = bytearray(f)
> +
> + # Find end of image followed by BRCM header
> + match = bytearray(b'\xff\xd9 at BRCM')
> + match_str = binascii.hexlify(match)
> + b_str = binascii.hexlify(b)
> +
> + # index is divided by two to go from string to hex
> + indices = [m.start() // 2 for m in re.finditer(match_str, b_str)]
> + if len(indices) == 0:
> + utils.eprint(f'No Broadcom header found in {self.path}')
> + return False
> +
> + start = indices[0] + 3
> + buf = b[start:start + 32768]
> +
> + try:
> + self._loadMetadataBrcm(buf)
> + except Exception as e:
> + utils.eprint(f'Failed to load metadata from {self.path}: {e}')
> + return False
> +
> + buf = b[start + 32768:]
> + try:
> + self._readImageBrcm(buf)
> + except Exception as e:
> + utils.eprint(f'Failed to load image data from {self.path}: {e}')
> + return False
> +
> + return True
> +
> + def getPatches(self, cen_coords, size=16):
> + ret = True
> +
> + # Obtain channel widths and heights
> + ch_w, ch_h = self.w, self.h
> + cen_coords = list(np.array((cen_coords[0])).astype(np.int32))
> + self.cen_coords = cen_coords
> +
> + # Squares are ordered by stacking macbeth chart columns from left to
> + # right. Some useful patch indices:
> + # white = 3
> + # black = 23
> + # 'reds' = 9, 10
> + # 'blues' = 2, 5, 8, 20, 22
> + # 'greens' = 6, 12, 17
> + # greyscale = 3, 7, 11, 15, 19, 23
> + all_patches = []
> + for ch in self.channels:
> + ch_patches = []
> + for cen in cen_coords:
> + # Macbeth centre is placed at top left of central 2x2 patch to
> + # account for rounding Patch pixels are sorted by pixel
> + # brightness so spatial information is lost.
> + patch = ch[cen[1] - 7:cen[1] + 9, cen[0] - 7:cen[0] + 9].flatten()
> + patch.sort()
> + if patch[-5] == (2**self.sigbits - 1) * 2**(16 - self.sigbits):
> + ret = False
> + ch_patches.append(patch)
> +
> + all_patches.append(ch_patches)
> +
> + self.patches = all_patches
> +
> + return ret
> diff --git a/utils/tuning/libtuning/libtuning.py b/utils/tuning/libtuning/libtuning.py
> new file mode 100644
> index 00000000..1b7d1306
> --- /dev/null
> +++ b/utils/tuning/libtuning/libtuning.py
> @@ -0,0 +1,191 @@
> +import libtuning.utils as utils
> +from libtuning.utils import eprint
> +
> +from enum import Enum, IntEnum
> +
> +
> +class color(IntEnum):
Class names should use CamelCase.
> + R = 0
> + GR = 1
> + GB = 2
> + B = 3
> + G = 4
G seems unused.
I would name the class BayerComponent or something similar.
> +
> +
> +class debug(Enum):
> + Plot = 1
> +
> +
> +# @brief What to do with the leftover pixels after dividing them into ALSC
> +# sectors, when the division gradient is uniform
> +# todo Do the first four and last two even make sense?
> +# @var Append Put the leftover pixels in their own smaller sector, after the
> +# uniform sectors
> +# @var Prepend Same as Append, but before the uniform sectors instead of after
> +# @var Midpend Same as Append, but in the center sector (if there are an odd
> +# number of sectors in that dimension) or the center two sectors (if there
> +# are an even number of sectors)
> +# @var Endpend same as midpend, but divided between the first and last sectors
> +# @var DistributeFront Divide the remainder equally (until running out,
> +# obviously) into the existing sectors, starting from the front
> +# @var DistributeBack Same as DistributeFront but starting from the back
> +# @var DistributeMiddle Same as DistributeFront but spreading from the middle
> +# @var DistributeEdges Same as Distribute Middle but spreading from both the
> +# front and back
> +class remainder(Enum):
> + Append = 1
> + Prepend = 2
> + Midpend = 3
> + Endpend = 4
> + DistributeFront = 5
> + DistributeBack = 6
> + DistributeMiddle = 7
> + DistributeEdges = 8
> +
> +
> +# @brief A helper class to contain a default value for a module configuration
> +# parameter
> +class param():
> + # @var Required The value contained in this instance is irrelevant, and the
> + # value must be provided by the tuning configuration file.
> + # @var Optional If the value is not provided by the tuning configuration
> + # file, then the value contained in this instance will be used instead.
> + # @var Hardcode The value contained in this instance will be used
> + class mode(Enum):
> + Required = 0
> + Optional = 1
> + Hardcode = 2
> +
> + # @param name Name of the parameter. Shall match the name used in the
> + # configuration file for the parameter
> + # @param required Whether or not a value is required in the config
> + # parameter of getVal()
> + # @param val Default value (only relevant if mode is Optional)
> + def __init__(self, name: str, required: mode, val=None):
> + self.name = name
> + self.required = required
> + self.val = val
> +
> + def getValue(self, config: dict):
> + if self.required is mode.Hardcode:
> + return self.val
> +
> + if self.required is mode.Required and self.name not in config:
> + raise ValueError(f'Parameter {self.name} is required but not provided in the configuration')
> +
> + return config[self.name] if self.required is mode.Required else self.val
> +
> + def isRequired(self):
> + return self.required is mode.Required
Rename self.required to self.__required and implement this function as a
getter (with a @property decorator). I think I'd turn getValue() and
getInfo() into getters too.
> +
> + # @brief Used by libtuning to auto-generate help information for the tuning
> + # script on the available parameters for the configuration file
> + # todo implement this
> + def getInfo(self):
> + raise NotImplementedError
> +
> +
> +class Camera(object):
> +
> + # External functions
> +
> + def __init__(self, platform_name):
> + self.name = platform_name
> + self.modules = []
> + self.parser = None
> + self.generator = None
> + self.output_order = []
> + self.config = {}
> + self.output = {}
> + return
> +
> + def add(self, module):
> + self.modules.append(module)
> + return
> +
> + def setInputType(self, parser):
> + self.parser = parser
> + return
> +
> + def setOutputType(self, output):
> + self.generator = output
> + return
> +
> + def setOutputOrder(self, modules):
> + self.output_order = modules
> + return
> +
> + # @brief Convert classes in self.output_order to the instances in self.modules
> + def _prepareOutputOrder(self):
> + output_order = self.output_order
> + self.output_order = []
> + for module_type in output_order:
> + modules = [module for module in self.modules if type(module) == module_type]
> + if len(modules) > 1:
> + eprint(f'Multiple modules found for module type "{module.name}"')
> + return False
> + if len(modules) < 1:
> + eprint(f'No module found for module type "{module.name}"')
> + return False
> + self.output_order.append(modules[0])
> +
> + return True
> +
> + def _validateSettings(self):
> + if self.parser is None:
> + eprint('Missing parser')
> + return False
> +
> + if self.generator is None:
> + eprint('Missing generator')
> + return False
> +
> + if len(self.modules) == 0:
> + eprint('No modules added')
> + return False
> +
> + if len(self.output_order) != len(self.modules):
> + eprint('Number of outputs does not match number of modules')
> + return False
> +
> + return True
> +
> + def run(self, argv):
> + args = utils.processArgs(argv, self.name)
> + if args is None:
> + return -1
> +
> + if not self._validateSettings():
> + return -1
> +
> + if not self._prepareOutputOrder():
> + return -1
> +
> + if len(args.config) > 0:
> + self.config, disable = self.parser.parse(args.config, self.modules)
> +
> + for module in disable:
> + if module in self.modules:
> + self.modules.remove(module)
> +
> + for module in self.modules:
> + if not module.validateConfig(self.config):
> + eprint(f'Config is invalid for module {module.name}')
> + return -1
> +
> + images = utils.loadImages(args.input, self.config, self.modules)
> + if images is None:
> + return -1
> +
> + # we need args for input image locations and debug options, and config
> + # for stuff like do_color and luminance_strength
> + for module in self.modules:
> + out = module.process(args, self.config, images, self.output)
> + if out is None:
> + eprint(f'Module {module.name} failed to process, aborting')
> + break
> + self.output[module] = out
> +
> + self.generator.write(args.output, self.output, self.output_order)
> +
> + return 0
> diff --git a/utils/tuning/libtuning/macbeth.py b/utils/tuning/libtuning/macbeth.py
> new file mode 100644
> index 00000000..bfedbe95
> --- /dev/null
> +++ b/utils/tuning/libtuning/macbeth.py
> @@ -0,0 +1,654 @@
> +# SPDX-License-Identifier: BSD-2-Clause
> +#
> +# Copyright (C) 2019, Raspberry Pi Ltd
> +#
> +# (Copied from: ctt_macbeth_locator.py)
> +
> +import cv2
> +import os
> +from pathlib import Path
> +
> +
> +# Reshape image to fixed width without distorting returns image and scale
> +# factor
> +def reshape(img, width):
> + factor = width / img.shape[0]
> + return cv2.resize(img, None, fx=factor, fy=factor), factor
> +
> +
> +# Display image for debugging... read at your own risk...
> +def represent(img, name='image'):
> + # if type(img) == tuple or type(img) == list:
> + # for i in range(len(img)):
> + # name = 'image {}'.format(i)
> + # cv2.imshow(name, img[i])
> + # else:
> + # cv2.imshow(name, img)
> + # cv2.waitKey(0)
> + # cv2.destroyAllWindows()
> + # return 0
> + """
> + code above displays using opencv, but this doesn't catch users pressing 'x'
> + with their mouse to close the window.... therefore matplotlib is used....
> + (thanks a lot opencv)
> + """
Let's write comments as comments, not as strings.
> + grid = plt.GridSpec(22, 1)
> + plt.subplot(grid[:19, 0])
> + plt.imshow(img, cmap='gray')
> + plt.axis('off')
> + plt.subplot(grid[21, 0])
> + plt.title('press \'q\' to continue')
> + plt.axis('off')
> + plt.show()
> +
> + # f = plt.figure()
> + # ax = f.add_subplot(211)
> + # ax2 = f.add_subplot(122)
> + # ax.imshow(img, cmap='gray')
> + # ax.axis('off')
> + # ax2.set_figheight(2)
> + # ax2.title('press \'q\' to continue')
> + # ax2.axis('off')
> + # plt.show()
> +
> +
> +def draw_macbeth_results(img, coords_fit):
> + # Extract data from coords_fit and plot on original image
> + if show and coords_fit is not None:
> + copy = img.copy()
> + verts = coords_fit[0][0]
> + cents = coords_fit[1][0]
> +
> + # Draw circles at vertices of macbeth chart
> + for vert in verts:
> + p = tuple(np.round(vert).astype(np.int32))
> + cv2.circle(copy, p, 10, 1, -1)
> +
> + # Draw circles at centres of squares
> + for i in range(len(cents)):
> + cent = cents[i]
> + p = tuple(np.round(cent).astype(np.int32))
> +
> + # Draw black circle on white square, white circle on black square
> + # an grey circle everywhere else.
> + if i == 3:
> + cv2.circle(copy, p, 8, 0, -1)
> + elif i == 23:
> + cv2.circle(copy, p, 8, 1, -1)
> + else:
> + cv2.circle(copy, p, 8, 0.5, -1)
> + copy, _ = reshape(copy, 400)
> + represent(copy)
> +
> +
> +def find_macbeth(img, mac_config):
> + small_chart = mac_config['small']
> + show = mac_config['show']
> +
> + # Catch the warnings
> + warnings.simplefilter("ignore")
> + warnings.warn("runtime", RuntimeWarning)
> +
> + # Reference macbeth chart is created that will be correlated with the
> + # located macbeth chart guess to produce a confidence value for the match.
> + script_dir = Path(os.path.realpath(os.path.dirname(__file__)))
> + macbeth_ref_path = script_dir.joinpath('macbeth_ref.pgm')
> + ref = cv2.imread(str(macbeth_ref_path), flags=cv2.IMREAD_GRAYSCALE)
> + ref_w = 120
> + ref_h = 80
> + rc1 = (0, 0)
> + rc2 = (0, ref_h)
> + rc3 = (ref_w, ref_h)
> + rc4 = (ref_w, 0)
> + ref_corns = np.array((rc1, rc2, rc3, rc4), np.float32)
> + ref_data = (ref, ref_w, ref_h, ref_corns)
> +
> + # Locate macbeth chart
> + cor, mac, coords, ret = get_macbeth_chart(img, ref_data)
> +
> + # Following bits of code try to fix common problems with simple techniques.
> + # If now or at any point the best correlation is of above 0.75, then
> + # nothing more is tried as this is a high enough confidence to ensure
> + # reliable macbeth square centre placement.
> +
> + for brightness in [2, 4]:
> + if cor >= 0.75:
> + break
> + img_br = cv2.convertScaleAbs(img, alpha=brightness, beta=0)
> + cor_b, mac_b, coords_b, ret_b = get_macbeth_chart(img_br, ref_data)
> + if cor_b > cor:
> + cor, mac, coords, ret = cor_b, mac_b, coords_b, ret_b
> +
> + # In case macbeth chart is too small, take a selection of the image and
> + # attempt to locate macbeth chart within that. The scale increment is
> + # root 2
> +
> + # These variables will be used to transform the found coordinates at
> + # smaller scales back into the original. If ii is still -1 after this
> + # section that means it was not successful
> + ii = -1
> + w_best = 0
> + h_best = 0
> + d_best = 100
> +
> + # d_best records the scale of the best match. Macbeth charts are only looked
> + # for at one scale increment smaller than the current best match in order to avoid
> + # unecessarily searching for macbeth charts at small scales.
> + # If a macbeth chart ha already been found then set d_best to 0
> + if cor != 0:
> + d_best = 0
> +
> + for index, pair in enumerate([{'sel': 2 / 3, 'inc': 1 / 6},
> + {'sel': 1 / 2, 'inc': 1 / 8},
> + {'sel': 1 / 3, 'inc': 1 / 12},
> + {'sel': 1 / 4, 'inc': 1 / 16}]):
> + if cor >= 0.75:
> + break
> +
> + # Check if we need to check macbeth charts at even smaller scales. This
> + # slows the code down significantly and has therefore been omitted by
> + # default, however it is not unusably slow so might be useful if the
> + # macbeth chart is too small to be picked up to by the current
> + # subselections. Use this for macbeth charts with side lengths around
> + # 1/5 image dimensions (and smaller...?) it is, however, recommended
> + # that macbeth charts take up as large as possible a proportion of the
> + # image.
> + if index >= 2 and (not small_chart or d_best <= index - 1):
> + break
> +
> + w, h = list(img.shape[:2])
> + # Set dimensions of the subselection and the step along each axis
> + # between selections
> + w_sel = int(w * pair['sel'])
> + h_sel = int(h * pair['sel'])
> + w_inc = int(w * pair['inc'])
> + h_inc = int(h * pair['inc'])
> +
> + loop = ((1 - pair['sel']) / pair['inc']) + 1
> + # For each subselection, look for a macbeth chart
> + for i in range(loop):
> + for j in range(loop):
> + w_s, h_s = i * w_inc, j * h_inc
> + img_sel = img[w_s:w_s + w_sel, h_s:h_s + h_sel]
> + cor_ij, mac_ij, coords_ij, ret_ij = get_macbeth_chart(img_sel, ref_data)
> +
> + # If the correlation is better than the best then record the
> + # scale and current subselection at which macbeth chart was
> + # found. Also record the coordinates, macbeth chart and message.
> + if cor_ij > cor:
> + cor = cor_ij
> + mac, coords, ret = mac_ij, coords_ij, ret_ij
> + ii, jj = i, j
> + w_best, h_best = w_inc, h_inc
> + d_best = index + 1
> +
> + # Transform coordinates from subselection to original image
> + if ii != -1:
> + for a in range(len(coords)):
> + for b in range(len(coords[a][0])):
> + coords[a][0][b][1] += ii * w_best
> + coords[a][0][b][0] += jj * h_best
> +
> + if not ret:
> + return None
> +
> + coords_fit = coords
> + if cor < 0.75:
> + eprint(f'Warning: Low confidence {cor:.3f} for macbeth chart in {img.path.name}')
> +
> + if show:
> + draw_macbeth_results(img, coords_fit)
> +
> + return coords_fit
> +
> +
> +# @brief Compute coordinates of macbeth chart vertices and square centres,
> +# @return (max_cor, best_map_col_norm, fit_coords, success)
> +#
> +# Also returns an error/success message for debugging purposes. Additionally,
> +# it scores the match with a confidence value.
> +#
> +# Brief explanation of the macbeth chart locating algorithm:
> +# - Find rectangles within image
> +# - Take rectangles within percentage offset of median perimeter. The
> +# assumption is that these will be the macbeth squares
> +# - For each potential square, find the 24 possible macbeth centre locations
> +# that would produce a square in that location
> +# - Find clusters of potential macbeth chart centres to find the potential
> +# macbeth centres with the most votes, i.e. the most likely ones
> +# - For each potential macbeth centre, use the centres of the squares that
> +# voted for it to find macbeth chart corners
> +# - For each set of corners, transform the possible match into normalised
> +# space and correlate with a reference chart to evaluate the match
> +# - Select the highest correlation as the macbeth chart match, returning the
> +# correlation as the confidence score
> +#
> +# todo: clean this up
> +def get_macbeth_chart(img, ref_data):
Please move this function above find_macbeth(), as find_macbeth() calls
it.
> + (ref, ref_w, ref_h, ref_corns) = ref_data
> +
> + """
> + the code will raise and catch a MacbethError in case of a problem, trying
> + to give some likely reasons why the problem occred, hence the try/except
> + """
> + try:
> + """
> + obtain image, convert to grayscale and normalise
> + """
> + src = img
> + src, factor = reshape(src, 200)
> + original = src.copy()
> + a = 125 / np.average(src)
> + src_norm = cv2.convertScaleAbs(src, alpha=a, beta=0)
> + """
> + This code checks if there are seperate colour channels. In the past the
> + macbeth locator ran on jpgs and this makes it robust to different
> + filetypes. Note that running it on a jpg has 4x the pixels of the
> + average bayer channel so coordinates must be doubled.
> +
> + This is best done in img_load.py in the get_patches method. The
> + coordinates and image width, height must be divided by two if the
> + macbeth locator has been run on a demosaicked image.
> + """
> + if len(src_norm.shape) == 3:
> + src_bw = cv2.cvtColor(src_norm, cv2.COLOR_BGR2GRAY)
> + else:
> + src_bw = src_norm
> + original_bw = src_bw.copy()
> + """
> + obtain image edges
> + """
> + sigma = 2
> + src_bw = cv2.GaussianBlur(src_bw, (0, 0), sigma)
> + t1, t2 = 50, 100
> + edges = cv2.Canny(src_bw, t1, t2)
> + """
> + dilate edges to prevent self-intersections in contours
> + """
> + k_size = 2
> + kernel = np.ones((k_size, k_size))
> + its = 1
> + edges = cv2.dilate(edges, kernel, iterations=its)
> + """
> + find Contours in image
> + """
> + conts, _ = cv2.findContours(edges, cv2.RETR_TREE,
> + cv2.CHAIN_APPROX_NONE)
> + if len(conts) == 0:
> + raise MacbethError(
> + '\nWARNING: No macbeth chart found!'
> + '\nNo contours found in image\n'
> + 'Possible problems:\n'
> + '- Macbeth chart is too dark or bright\n'
> + '- Macbeth chart is occluded\n'
> + )
> + """
> + find quadrilateral contours
> + """
> + epsilon = 0.07
> + conts_per = []
> + for i in range(len(conts)):
> + per = cv2.arcLength(conts[i], True)
> + poly = cv2.approxPolyDP(conts[i], epsilon * per, True)
> + if len(poly) == 4 and cv2.isContourConvex(poly):
> + conts_per.append((poly, per))
> +
> + if len(conts_per) == 0:
> + raise MacbethError(
> + '\nWARNING: No macbeth chart found!'
> + '\nNo quadrilateral contours found'
> + '\nPossible problems:\n'
> + '- Macbeth chart is too dark or bright\n'
> + '- Macbeth chart is occluded\n'
> + '- Macbeth chart is out of camera plane\n'
> + )
> +
> + """
> + sort contours by perimeter and get perimeters within percent of median
> + """
> + conts_per = sorted(conts_per, key=lambda x: x[1])
> + med_per = conts_per[int(len(conts_per) / 2)][1]
> + side = med_per / 4
> + perc = 0.1
> + med_low, med_high = med_per * (1 - perc), med_per * (1 + perc)
> + squares = []
> + for i in conts_per:
> + if med_low <= i[1] and med_high >= i[1]:
> + squares.append(i[0])
> +
> + """
> + obtain coordinates of nomralised macbeth and squares
> + """
> + square_verts, mac_norm = get_square_verts(0.06)
> + """
> + for each square guess, find 24 possible macbeth chart centres
> + """
> + mac_mids = []
> + squares_raw = []
> + for i in range(len(squares)):
> + square = squares[i]
> + squares_raw.append(square)
> + """
> + convert quads to rotated rectangles. This is required as the
> + 'squares' are usually quite irregular quadrilaterls, so performing
> + a transform would result in exaggerated warping and inaccurate
> + macbeth chart centre placement
> + """
> + rect = cv2.minAreaRect(square)
> + square = cv2.boxPoints(rect).astype(np.float32)
> + """
> + reorder vertices to prevent 'hourglass shape'
> + """
> + square = sorted(square, key=lambda x: x[0])
> + square_1 = sorted(square[:2], key=lambda x: x[1])
> + square_2 = sorted(square[2:], key=lambda x: -x[1])
> + square = np.array(np.concatenate((square_1, square_2)), np.float32)
> + square = np.reshape(square, (4, 2)).astype(np.float32)
> + squares[i] = square
> + """
> + find 24 possible macbeth chart centres by trasnforming normalised
> + macbeth square vertices onto candidate square vertices found in image
> + """
> + for j in range(len(square_verts)):
> + verts = square_verts[j]
> + p_mat = cv2.getPerspectiveTransform(verts, square)
> + mac_guess = cv2.perspectiveTransform(mac_norm, p_mat)
> + mac_guess = np.round(mac_guess).astype(np.int32)
> + """
> + keep only if candidate macbeth is within image border
> + (deprecated)
> + """
> + in_border = True
> + # for p in mac_guess[0]:
> + # pptest = cv2.pointPolygonTest(
> + # img_con,
> + # tuple(p),
> + # False
> + # )
> + # if pptest == -1:
> + # in_border = False
> + # break
> +
> + if in_border:
> + mac_mid = np.mean(mac_guess,
> + axis=1)
> + mac_mids.append([mac_mid, (i, j)])
> +
> + if len(mac_mids) == 0:
> + raise MacbethError(
> + '\nWARNING: No macbeth chart found!'
> + '\nNo possible macbeth charts found within image'
> + '\nPossible problems:\n'
> + '- Part of the macbeth chart is outside the image\n'
> + '- Quadrilaterals in image background\n'
> + )
> +
> + """
> + reshape data
> + """
> + for i in range(len(mac_mids)):
> + mac_mids[i][0] = mac_mids[i][0][0]
> +
> + """
> + find where midpoints cluster to identify most likely macbeth centres
> + """
> + clustering = cluster.AgglomerativeClustering(
> + n_clusters=None,
> + compute_full_tree=True,
> + distance_threshold=side * 2
> + )
> + mac_mids_list = [x[0] for x in mac_mids]
> +
> + if len(mac_mids_list) == 1:
> + """
> + special case of only one valid centre found (probably not needed)
> + """
> + clus_list = []
> + clus_list.append([mac_mids, len(mac_mids)])
> +
> + else:
> + clustering.fit(mac_mids_list)
> + # try:
> + # clustering.fit(mac_mids_list)
> + # except RuntimeWarning as error:
> + # return(0, None, None, error)
> +
> + """
> + create list of all clusters
> + """
> + clus_list = []
> + if clustering.n_clusters_ > 1:
> + for i in range(clustering.labels_.max() + 1):
> + indices = [j for j, x in enumerate(clustering.labels_) if x == i]
> + clus = []
> + for index in indices:
> + clus.append(mac_mids[index])
> + clus_list.append([clus, len(clus)])
> + clus_list.sort(key=lambda x: -x[1])
> +
> + elif clustering.n_clusters_ == 1:
> + """
> + special case of only one cluster found
> + """
> + # print('only 1 cluster')
> + clus_list.append([mac_mids, len(mac_mids)])
> + else:
> + raise MacbethError(
> + '\nWARNING: No macebth chart found!'
> + '\nNo clusters found'
> + '\nPossible problems:\n'
> + '- NA\n'
> + )
> +
> + """
> + keep only clusters with enough votes
> + """
> + clus_len_max = clus_list[0][1]
> + clus_tol = 0.7
> + for i in range(len(clus_list)):
> + if clus_list[i][1] < clus_len_max * clus_tol:
> + clus_list = clus_list[:i]
> + break
> + cent = np.mean(clus_list[i][0], axis=0)[0]
> + clus_list[i].append(cent)
> +
> + """
> + represent most popular cluster centroids
> + """
> + # copy = original_bw.copy()
> + # copy = cv2.cvtColor(copy, cv2.COLOR_GRAY2RGB)
> + # copy = cv2.resize(copy, None, fx=2, fy=2)
> + # for clus in clus_list:
> + # centroid = tuple(2*np.round(clus[2]).astype(np.int32))
> + # cv2.circle(copy, centroid, 7, (255, 0, 0), -1)
> + # cv2.circle(copy, centroid, 2, (0, 0, 255), -1)
> + # represent(copy)
> +
> + """
> + get centres of each normalised square
> + """
> + reference = get_square_centres(0.06)
> +
> + """
> + for each possible macbeth chart, transform image into
> + normalised space and find correlation with reference
> + """
> + max_cor = 0
> + best_map = None
> + best_fit = None
> + best_cen_fit = None
> + best_ref_mat = None
> +
> + for clus in clus_list:
> + clus = clus[0]
> + sq_cents = []
> + ref_cents = []
> + i_list = [p[1][0] for p in clus]
> + for point in clus:
> + i, j = point[1]
> + """
> + remove any square that voted for two different points within
> + the same cluster. This causes the same point in the image to be
> + mapped to two different reference square centres, resulting in
> + a very distorted perspective transform since cv2.findHomography
> + simply minimises error.
> + This phenomenon is not particularly likely to occur due to the
> + enforced distance threshold in the clustering fit but it is
> + best to keep this in just in case.
> + """
> + if i_list.count(i) == 1:
> + square = squares_raw[i]
> + sq_cent = np.mean(square, axis=0)
> + ref_cent = reference[j]
> + sq_cents.append(sq_cent)
> + ref_cents.append(ref_cent)
> +
> + """
> + At least four squares need to have voted for a centre in
> + order for a transform to be found
> + """
> + if len(sq_cents) < 4:
> + raise MacbethError(
> + '\nWARNING: No macbeth chart found!'
> + '\nNot enough squares found'
> + '\nPossible problems:\n'
> + '- Macbeth chart is occluded\n'
> + '- Macbeth chart is too dark of bright\n'
> + )
> +
> + ref_cents = np.array(ref_cents)
> + sq_cents = np.array(sq_cents)
> + """
> + find best fit transform from normalised centres to image
> + """
> + h_mat, mask = cv2.findHomography(ref_cents, sq_cents)
> + if 'None' in str(type(h_mat)):
> + raise MacbethError(
> + '\nERROR\n'
> + )
> +
> + """
> + transform normalised corners and centres into image space
> + """
> + mac_fit = cv2.perspectiveTransform(mac_norm, h_mat)
> + mac_cen_fit = cv2.perspectiveTransform(np.array([reference]), h_mat)
> + """
> + transform located corners into reference space
> + """
> + ref_mat = cv2.getPerspectiveTransform(
> + mac_fit,
> + np.array([ref_corns])
> + )
> + map_to_ref = cv2.warpPerspective(
> + original_bw, ref_mat,
> + (ref_w, ref_h)
> + )
> + """
> + normalise brigthness
> + """
> + a = 125 / np.average(map_to_ref)
> + map_to_ref = cv2.convertScaleAbs(map_to_ref, alpha=a, beta=0)
> + """
> + find correlation with bw reference macbeth
> + """
> + cor = correlate(map_to_ref, ref)
> + """
> + keep only if best correlation
> + """
> + if cor > max_cor:
> + max_cor = cor
> + best_map = map_to_ref
> + best_fit = mac_fit
> + best_cen_fit = mac_cen_fit
> + best_ref_mat = ref_mat
> +
> + """
> + rotate macbeth by pi and recorrelate in case macbeth chart is
> + upside-down
> + """
> + mac_fit_inv = np.array(
> + ([[mac_fit[0][2], mac_fit[0][3],
> + mac_fit[0][0], mac_fit[0][1]]])
> + )
> + mac_cen_fit_inv = np.flip(mac_cen_fit, axis=1)
> + ref_mat = cv2.getPerspectiveTransform(
> + mac_fit_inv,
> + np.array([ref_corns])
> + )
> + map_to_ref = cv2.warpPerspective(
> + original_bw, ref_mat,
> + (ref_w, ref_h)
> + )
> + a = 125 / np.average(map_to_ref)
> + map_to_ref = cv2.convertScaleAbs(map_to_ref, alpha=a, beta=0)
> + cor = correlate(map_to_ref, ref)
> + if cor > max_cor:
> + max_cor = cor
> + best_map = map_to_ref
> + best_fit = mac_fit_inv
> + best_cen_fit = mac_cen_fit_inv
> + best_ref_mat = ref_mat
> +
> + """
> + Check best match is above threshold
> + """
> + cor_thresh = 0.6
> + if max_cor < cor_thresh:
> + raise MacbethError(
> + '\nWARNING: Correlation too low'
> + '\nPossible problems:\n'
> + '- Bad lighting conditions\n'
> + '- Macbeth chart is occluded\n'
> + '- Background is too noisy\n'
> + '- Macbeth chart is out of camera plane\n'
> + )
> + """
> + Following code is mostly representation for debugging purposes
> + """
> +
> + """
> + draw macbeth corners and centres on image
> + """
> + copy = original.copy()
> + copy = cv2.resize(original, None, fx=2, fy=2)
> + # print('correlation = {}'.format(round(max_cor, 2)))
> + for point in best_fit[0]:
> + point = np.array(point, np.float32)
> + point = tuple(2 * np.round(point).astype(np.int32))
> + cv2.circle(copy, point, 4, (255, 0, 0), -1)
> + for point in best_cen_fit[0]:
> + point = np.array(point, np.float32)
> + point = tuple(2 * np.round(point).astype(np.int32))
> + cv2.circle(copy, point, 4, (0, 0, 255), -1)
> + copy = copy.copy()
> + cv2.circle(copy, point, 4, (0, 0, 255), -1)
> +
> + """
> + represent coloured macbeth in reference space
> + """
> + best_map_col = cv2.warpPerspective(
> + original, best_ref_mat, (ref_w, ref_h)
> + )
> + best_map_col = cv2.resize(
> + best_map_col, None, fx=4, fy=4
> + )
> + a = 125 / np.average(best_map_col)
> + best_map_col_norm = cv2.convertScaleAbs(
> + best_map_col, alpha=a, beta=0
> + )
> + # cv2.imshow('Macbeth', best_map_col)
> + # represent(copy)
> +
> + """
> + rescale coordinates to original image size
> + """
> + fit_coords = (best_fit / factor, best_cen_fit / factor)
> +
> + return(max_cor, best_map_col_norm, fit_coords, True)
> +
> + """
> + catch macbeth errors and continue with code
> + """
> + except MacbethError as error:
> + eprint(error)
> + return(0, None, None, False)
> diff --git a/utils/tuning/libtuning/macbeth_ref.pgm b/utils/tuning/libtuning/macbeth_ref.pgm
> new file mode 100644
> index 00000000..9b9f4920
> --- /dev/null
> +++ b/utils/tuning/libtuning/macbeth_ref.pgm
> @@ -0,0 +1,5 @@
PGM files are nice, you can add comments, so SPDX is possible too :-)
> +P5
> +# Reference macbeth chart
> +120 80
> +255
> + !#!"
#!"&&$#$#'"%&#+2///..../.........-()))))))))))))))))))(((-,*)'(&)#($%(%"###""!%""&"&&!$" #!$ !"! $&**" !#
5.,%
+,-5"0<HBAA54" %##((()*+,---.........+*)))))))))))))))-.,,--+))('((''('%'%##"!""!"!""""#!
! %?/v??z:????L??????c?,!#""%%''')**+)-../..../.-*)))))))))))))**,,)**'(''&'((&&%%##$! !!!! ! !
!
5*"-)&7(1.75Rnge`\`$ ""!"%%%'')())++--/---,-..,-.,++**))))())*)*)''%'%&%&'&%%"""""
!
!! $&$$&##(+*,,/10122126545./66402006486869650*.1.***)*+)()&((('('##)('&%%&%$$$#$%$%$ (((*))('((('('(&%V
0;>>;@@>@AAAACBCB=&<?????????????????<5x???????????????|64RYVTSRRRMMNLKJJLH+&0gijgdeffmmnpnkji`#3????????????????bY!
3FHHIIIHIJIIJHIII@#??????????????????=7}????????????????:5Wcbcbdcb`^^`^^_^Y,'6????????????????r'<????????????????l%
2FHHIIHJJJJJJIIJI?%;?????????????????>7|????????????????;8Xfeeegeccb`^aba]Z+)<????????????????r)>????????????????q#
3GHIIIIJIIJJIHIJI@&5?????????????????=8~????????????????;8Zgghggedbdcbda^\Z+(;????????????????y)9????????????????z"
3GIIJJJJJKJJJJJJJ@'4?????????????????>9|????????????????=8Zhighgeeeedeca__[/)B????????????????v&:????????????????|#
3GJJIIJKKKJJJKKJK@&6?????????????????>9~????????????????<8Yghegggffihccab^\/*C????????????????z'9?????????????????$
6IKJJMMMKMKKMKKMLC&2?????????????????@9?????????????????<9Yghhhhijiegdcebc^0)G????????????????(7?????????????????% 6JLMMNMMKMMNMMMMMD&2?????????????????@:~????????????????=9Xfghhjiigdgddedc`1)M????????????????}(:??????????????¾?&
"8LNOONNOMONNMMNOND'3?????????????????@;????????????????=:Ziiigheegegegggdc1,Q????????????????~)8?????????????????%
# "9NNNPPPQOOOOONNOOD'0??????????????????;????????????????=;[iigeeegghgdedgea0-P????????????????(8???????????????Ý'
"#$:NNOQPPRPQPOOPQPPD*1?????????????????A;?????????????????;:Yfghgghgghghhdggc3.\????????????????~);???????????????¤( &%%;OQQQRSSRPQQQQSQQF)3?????????????????B<?????????????????=:Wfhghhhihggghfhee4/f?????????????????*:???????????????ä&
%%%?RSSSSSTTTTSSSTTRE)5?????????????????B=?????????????????@:Ygiihhiiiihihiiif72p????????????????}(9???????????????Ʃ'
#%&?TUTTTUUQSTTTTTVSF*3?????????????????F>?????????????????A;[ghjiihiiiihihije50r?????????????????)6???????????????ƫ&
&#%?SVVVUUUUUTUUVVUUG*5?????????????????F=?????????????????A;Yhijiiijjiiiiijje81t????????????????~)5???????????????ư' '$$=OQRRQQPRSRSSSSSSG+6?????????????????D@??????????????????;Wefgggggfffgeeefc41x????????????????{*5?????????????????(
&&&'++++,,*-,-00-0100*-SUX\]]`_ffgiooopo=;X\bedbadbca`]\]ZZ;;<::8:;9983433110/-,...1//12410/..--+)"",---,-./,,.-/-0-(
&&%+/0103322011223233)(34534767::;;==:=B9;BFGEEGIKJKIJGIJCD=<:76566554111/0/1.*+00233300/00//..,+*#")(*)++,++))*++**'! !&$*w????¼???????????1-_addc`ceccdccedbb?A|????????????????B>=>?@@?====;<:;:<:11r?????????????????+.?????????????????( !'%*z???????????????ɠ42gjmllklomooonpopmHG?????????????????D>AEDEFEECEECCCDDEC46????????????????0:???????????????Ѿ,!!&&,|???????????????ʡ61inknnoopoppoqqrqoEE?????????????????FACGFFFFFFDFDDDDDDC57??????????????????09?????????????????+!"%%-~???????????????ʡ42inopppppoqqqrrsrnAB?????????????????C?DGGGGFFFFDFFDDEDC48??????????????????1;?????????????????+!!"#*|???????????????ʡ62imoppppqqqqrtrqtrGD?????????????????H?CGGGGGGGGFFFFFFDB38??????????????????1<???????????????Խ, !)}???????????????ˢ63mooppqqqqqqrrtvtoDH?????????????????JACHHGGHGGFFFDDGGFD29??????????????????3>???????????????,
$){???????????????ˢ53jpppqprqrrrttuvuo>H?????????????????JAFHHHHHGGHGGFGGFFE28??????????????????3:???????????????ڽ- "*{???????????????̣53loqpqsqrrrtrutsvrAH?????????????????HCGHIHHHHHHGFGHGGGD5;??????????????????28?????????????????,
+}???????????????ʡ52mqoqpqrttttttuurpFI?????????????????OCEHHIHHHHGHGGFFIGF8<??????????????????48???????????????ۿ,
(|???????????????ʢ41krqpqqqrrtrtuvtuoEH?????????????????PBHHIIIHIIHIHGHGHHE7<??????????????????58?????????????????* (z???????????????ʡ63kpqprqqstttutrvvoFO?????????????????LEHHIIHIHHHIGHGIHGF4=??????????????????5<?????????????????* 'z???????????????ȡ62lppqrqrrrtttuttvpAG?????????????????MGHIIIIHIIIHHIIJHHG4<??????????????????4<?????????????????+ !){???????????????Ƞ62jopqqqqqrtttutttrEH?????????????????OHFIIIIIJIIIIHIHIHI7>??????????????????5;?????????????????, !)z???????????????Ɵ53lppqqrqrtttuuuutsFI?????????????????RHGJIJHJKJJJIIIIIIH9>??????????????????5;?????????????????+ !({???????????????Ŝ41joppprqrrrutttvvrIH?????????????????THCJJJJJIJIJJIJJJIH7=??????????????????5;?????????????????+ (u?????????????????65gjlmmmnoopnpprpqoIH?????????????????OIBIJJJIJJJJIIIHHHG89??????????????????29???????????????ʾ' "&,-*)-01/,0/12102-+04448789<>>??AFAD at DBCIJNRWTSUXT[WUQUOKFEBBABA?>>=<<;;67942:<<<>9999864565363&(13335422./1/-+..+ !"&$$""$"&$%'()(''*+-0124688:<>>??A>?EBCHKOLJLNOSQOXQQVMLACGHGHIGFHGDCCBB@??7432233210111.,++,++%(++)*(''%%%$$#%&$#
")0/001120024455520+-U]`addcdhefeekecYGFJRXYYVWWZWVXXVZTOBF}????????????????K7Ybccddfeg`^]^]\[Z[*)OTTPPQPOKOLLJJLIK
!1;:9:<<===;=???A at 9*/?????????????????FJmxyxwyzzzxyzzz{zxLO?????????????????]=??????????????????.-???????????????y#
!!2><=;==>=<<>@@@@A9-0?????????????????IKnz||{|{||{}}~}}{zLO?????????????????]>??????????????????..????????????????~%
$2==;<>>?===>@A at AB;+1?????????????????JJo{|y{||}{||}}}}}yMT?????????????????_>??????????????????-.????????????????}#
%2<=;=<@?>==>?A at AA9+3?????????????????FMlz{{y|}}}}||}|}}{MT?????????????????d>??????????????????-,????????????????#
%1<<<;==<<=>?A?@AA:,3?????????????????INo{{y{||||}|}}|~}{RT?????????????????d=??????????????????/-????????????????}#
!$0<<<=<<==>A@@>@AA:-2?????????????????HInzz{{||{{}~~}}|}zMR?????????????????d=??????????????????++????????????????~#
"$/;<==>;===@@@@>AA:+2?????????????????KHn||y|||||{}~}|}|xMS?????????????????d=??????????????????+,????????????????}#
! "/:<=>@<<>=@@@@@AA;-3?????????????????MFs||{{{y}z}}|}|}}yMW?????????????????c>??????????????????,)????????????????|!
!1;>?>><<>@>>=>ABB;,0?????????????????LHr{|{|}|y|}}}}}zNX?????????????????c???????????????????()????????????????z#
$/;;<=;<>>=>>>@@BB:,1?????????????????IInyz||||||{||}{~|{NV?????????????????c;?????????????????('????????????????}#
$0:<==<;>@>>>>@ABB:,/?????????????????HLlx|}y{y{|y{|}}}}yMR?????????????????d>~?????????????????*(???????????????y"
!&3:;<<;==@@=>AABBA;-3?????????????????KLqz{|||y{}|}{}|~{zRQ?????????????????c9w?????????????????)'????????????????y"
!%1<<;=>===<=@@ABBC<.5?????????????????IIlz{|}~~~|}{||~}}zMU?????????????????d;p?????????????????)$???????????????x"
$2===<==@=<>=ABBBC?/0?????????????????IGkz}}{||}{||y||}zyOV?????????????????c7o?????????????????'&~??????????~?z"
#"#/;<:<<?>;===@?AAA>07?????????????????GGgwxz{yyxyzzyz{yuuHO?????????????????\8v?????????????????'$w~~}|||{~|{zxxxxv!
"""'*+(+)*))()+,,.../0398;=<=>DCCDDCBBDHBCJMMLMPNPOJPKPSJDICCNMPONMNNOKHIFDBHE3/46433323.....*+,)(
!##!!!!!$#$$#$#&"
!!"(+**,,*+.//1478:<:33ACDFGGIIHIJLPKNMQFIPTTRVXVUXUUTXUSTNEGGFDEFAA>==;94877520-,))*(((('&$#!!"
&%'FQPQR]dq??????????=F?????????????????QN?????????????????LE????znki^[YTPUOS;.%-/12322221/10//,/
%#0??????????????????@Q?????????????????QM?????????????????KE?????????????????H01NNQOQQOOMNNLKLJGB
'&/??????????????????AW?????????????????OL?????????????????KE?????????????????F-,PQQPQPPQPOONMNNKE
''0??????????????????CZ?????????????????RM?????????????????JE?????????????????F,*NSQPPQOOOOMNNMKID
('2??????????????????D[?????????????????QK?????????????????IF?????????????????F,*NPPPPPPNOONMMMJIF!
'(2??????????????????F]?????????????????RL?????????????????HD????????????????F+%MPPPPOOONONNMMKID )*4??????????????????D^?????????????????PL?????????????????IC?????????????????F+&NPOOOPPOONMMKMKHD
**6??????????????????D_?????????????????QJ?????????????????FC~????????????????F,'MPOOOOONONNKKIIIG
,+7??????????????????D^?????????????????QI?????????????????EB|????????????????E+&MONOOONNNNKMJKJHH
,-8??????????????????D]?????????????????PI?????????????????HE????????????????C,#LOOOONONNNKKKMKJF
,*6??????????????????Ca?????????????????MH?????????????????IF?????????????????D*%KONOMNMMKMKJJJIJE
,,6??????????????????B^?????????????????MG?????????????????HB}????????????????D+&LONOOONNMMMMKLKIA
,,6??????????????????A\?????????????????MF?????????????????IE????????????????E+&LNNMONNMMKKKKKIHF
--6??????????????????A[?????????????????KF?????????????????JC????????????????F*&LMONMNMNKKJMKJJIF
**5??????????????????>W?????????????????KE?????????????????F?}????????????????C*%KONNNJKKKMKJKJKID
,*4??????????????????<W?????????????????MA?????????????????GCx????????????????B)%HKLKKJJJKIHIHHFGC!
()*q????????????????o39v|}wwwwwwrqtuspn=9^gadcfgce`dbUY[\^>;DIJDB?FEGE=7>8634.(&&(%&*&%%'+*)+*#%(
)''03364443233222243/-+133423333423766645789:><<<;<;<?=?;<<:78673/001113--.-+*)&&#"&$#%&""$!!
))+rbPpAD9-*******+*++)++--.//./.0/21453469:=;98<;<>=;><7766666741012.-13/-+-/(''&&&%%&$.%0()-%-#-#'
# &
(
%
)))h?n?YQg?7(*))))*)**,--....../0/0001357666::;;>?>AA866666666656565300/20/.-*)(('((&&%)d=yoP?<???F?QFx;?2?1?0
))*RQ.0*,,5*(*))))*,**,+/.../...02/22224456468;:>BB;>;:76666666666755303033/,.-*(())('&')#)"##(+$+ *
# ) ) &
> diff --git a/utils/tuning/libtuning/modules/__init__.py b/utils/tuning/libtuning/modules/__init__.py
> new file mode 100644
> index 00000000..e69de29b
> diff --git a/utils/tuning/libtuning/modules/module.py b/utils/tuning/libtuning/modules/module.py
> new file mode 100644
> index 00000000..e45a8751
> --- /dev/null
> +++ b/utils/tuning/libtuning/modules/module.py
> @@ -0,0 +1,41 @@
> +# @var hr_name Human-readable module name
> +# @var name Name of the module. Should match the standard name eg. 'alsc'
> +class Module(object):
> + def __init__(self):
> + self.hr_name = "Base Module"
> + self.name = "module"
> + self.options = {}
> +
> + # todo: I don't think we need these and the options member variable
> + def setValue(self, key, value):
> + self.options[key] = value
> +
> + def appendValue(self, key, value):
> + if key not in self.options:
> + self.options[key] = []
> + if not isinstance(self.options[key], list):
> + raise TypeError(f'Options "{key}" in module "{self.name}" is not a list')
> + self.options[key].append(value)
> +
> + def __validateConfig__(self, config: dict) -> bool:
> + if self not in config:
> + eprint(f'No config found for {self.name}')
> + return False
> + return True
> +
> + def __process__(self, args, config: dict, images: list, outputs: dict) -> dict:
> + raise NotImplementedError
> +
> + def validateConfig(self, config: dict) -> bool:
> + return self.__validateConfig__(config)
> +
> + # @brief Do the module's processing
> + # @param args argparse arguments
> + # @param config Full configuration from the input configuration file
> + # @param images List of images to process
> + # @param outputs The outputs of all modules that were executed before this
> + # module. Note that this is an input parameter, and the
> + # output of this module should be returned directly
> + # @return Result of the module's processing
> + def process(self, args, config: dict, images: list, outputs: dict) -> dict:
> + return self.__process__(args, config, images, outputs)
> diff --git a/utils/tuning/libtuning/parsers/__init__.py b/utils/tuning/libtuning/parsers/__init__.py
> new file mode 100644
> index 00000000..e69de29b
> diff --git a/utils/tuning/libtuning/parsers/parser.py b/utils/tuning/libtuning/parsers/parser.py
> new file mode 100644
> index 00000000..f1e6e629
> --- /dev/null
> +++ b/utils/tuning/libtuning/parsers/parser.py
> @@ -0,0 +1,18 @@
> +class Parser(object):
> + def __init__(self):
> + return
> +
> + def __parse__(self, config_file):
> + raise NotImplementedError("__parse__() must be implemented")
> +
> + # @brief Parse a config file into a config dict
> + # @details The config dict shall have one key 'general' with a dict value
> + # for general configuration options, and all other entries shall
> + # have the module as the key with its configuration options (as a
> + # dict) as the value. The config dict shall prune entries that are
> + # for modules that are not in @a modules.
> + # @param config (str) Path to config file
> + # @param modules (list) List of modules
> + # @return (dict, list) Configuration and list of modules to disable
> + def parse(self, config_file: str, modules: list) -> (dict, list):
> + return self.__parse__(config_file)
> diff --git a/utils/tuning/libtuning/smoothing.py b/utils/tuning/libtuning/smoothing.py
> new file mode 100644
> index 00000000..b05e5c75
> --- /dev/null
> +++ b/utils/tuning/libtuning/smoothing.py
> @@ -0,0 +1,21 @@
> +import libtuning as lt
Not used.
> +
> +import cv2
> +
> +
> +# @brief Wrapper for cv2 smoothing functions so that they can be duck-typed
> +class Smoothing(object):
> + def __init__(self, params: list = []):
> + self.params = params
> + return
> +
> + def __smoothing__(self, src, ksize):
> + raise NotImplementedError
> +
> + def smoothing(self, src, ksize):
> + return self.__smoothing__(src, ksize)
> +
> +
> +class MedianBlur(Smoothing):
> + def __smoothing__(self, src, ksize):
> + return cv2.medianBlur(src.astype('float32'), ksize).astype('float64')
> diff --git a/utils/tuning/libtuning/utils.py b/utils/tuning/libtuning/utils.py
> new file mode 100644
> index 00000000..63ddedc9
> --- /dev/null
> +++ b/utils/tuning/libtuning/utils.py
Ah, the good ol' utils, where we put everything that doesn't fit
elsewhere :-)
> @@ -0,0 +1,198 @@
> +import argparse
> +import decimal
> +import math
> +import numpy as np
> +import os
> +from pathlib import Path
> +import re
> +import sys
> +
> +from libtuning.image import Image
> +from libtuning.macbeth import find_macbeth
> +
> +# Utility functions
> +
> +
> +def eprint(*args, **kwargs):
> + print(*args, file=sys.stderr, **kwargs)
> +
> +
> +def getModuleByName(modules, name):
> + for module in modules:
> + if module.name == name:
> + return module
> + return None
> +
> +
> +# @brief Round value while keeping the maximum number of decimal points
> +# @param limits Tuple of [min, max] acceptable values
> +# @description Prevents rounding such that significant figures are lost
> +# todo Bikeshed this name
> +def roundWithSigfigs(val, limits: tuple):
> + decimal_points = abs(decimal.Decimal(str(limits[-1])).as_tuple().exponent)
> + lshift = 10**(decimal_points - 1)
> + adjust = 10**(-decimal_points)
> +
> + # We need the division to get rid of stray floating points
> + # todo Any better solution?
> + lower_bound = adjust * 10 * 5 * lshift / lshift
> + upper_bound = adjust * 10 * 95 * lshift / lshift
> +
> + out = val
> + out = np.where((lshift * out) % 1 <= lower_bound, out + adjust, out)
> + out = np.where((lshift * out) % 1 >= upper_bound, out - adjust, out)
> +
> + return out
> +
> +
> +# Private utility functions
> +
> +
> +def _listImageFiles(directory):
> + d = Path(directory)
> + files = [d.joinpath(f) for f in os.listdir(d)
> + if re.search(r'\.(jp[e]g$)|(dng$)|(brcm$)', filename)]
> + files.sort()
> + return files
> +
> +
> +def _parseImageFilename(fn: Path):
> + result = re.search(r'^(alsc_){0,1}(\d+)[kK]_(\d+){0,1}[lLuU].\w{3,4}$', fn.name)
Unless I'm mistaken, {0,1} can be written ?, which would be more
readable.
> + if result is None:
> + eprint(f'The file name of {fn.name} is incorrectly formatted')
> + return None, None, None
> +
> + color = int(result.group(2))
> + alsc_only = result.group(1) is not None
> + lux = None if alsc_only else int(result.group(3))
> +
> + return color, lux, alsc_only
> +
> +
> +def _loadDngImage(path: Path):
> + image = Image(path)
> +
> + if not image.loadDng():
> + return None
> +
> + return image
> +
> +
> +def _loadBrcmImage(path: Path):
> + image = Image(path)
> +
> + if not image.loadBrcm():
> + return None
> +
> + return image
> +
> +
> +def _locateMacbeth(image: Image, config: dict):
> + # Find macbeth centres
> + av_chan = (np.mean(np.array(image.channels), axis=0) / (2**16))
> + av_val = np.mean(av_chan)
> + if av_val < image.blacklevel_16 / (2**16) + 1 / 64:
> + eprint(f'Image {image.path.name} too dark')
> + return None
> +
> + macbeth = find_macbeth(av_chan, config['general']['macbeth'])
> +
> + if macbeth is None:
> + eprint(f'No macbeth chart found in {image.path.name}')
> + return None
> +
> + mac_cen_coords = macbeth[1]
> + if not image.getPatches(mac_cen_coords):
> + eprint(f'Macbeth patches have saturated in {image.path.name}')
> + return None
> +
> + return macbeth
This seems like it would be better placed in macbeth.py.
> +
> +
> +# todo Implement this from check_imgs() in ctt.py
> +def _validateImages(images):
> + return True
> +
> +
> +# Public utility functions
> +
> +
> +def processArgs(argv, platform_name):
This function and the next one seems like they would be better placed in
libtuning.py.
> + parser = argparse.ArgumentParser(description=f'Camera Tuning for {platform_name}')
> + parser.add_argument('-i', '--input', type=str, required=True,
> + help='''Directory containing calibration images (required).
> + Images for ALSC must be named "alsc_{Color Temperature}k_1[u].dng",
> + and all other images must be named "{Color Temperature}k_{Lux Level}l.dng"''')
> + parser.add_argument('-o', '--output', type=str, required=True,
> + help='Output file (required)')
> + # It is not our duty to scan all modules to figure out their default
> + # options, so simply return an empty configuration if none is provided.
> + parser.add_argument('-c', '--config', type=str, default='',
> + help='Config file (optional)')
> + # todo check if we really need this or if stderr is good enough, or if we
> + # want a better logging infrastructure with log levels
> + parser.add_argument('-l', '--log', type=str, default=None,
> + help='Output log file (optional)')
> + return parser.parse_args(argv)
> +
> +
> +def loadImages(input_dir: str, config: dict, modules: list) -> list:
> + files = _listImageFiles(input_dir)
> + if len(files) == 0:
> + eprint(f'No images found in {input_dir}')
> + return None
> +
> + # todo Should this match by name instead of type?
> + has_alsc = any(isinstance(m, modules.ALSC) for m in modules)
> + # todo Is there any use case for multiple ALSC modules?
> + has_only_alsc = has_alsc and len(modules) == 1
Instead of passing the modules to this function, I think the caller
should figure out what images it needs, and pass that explicitly as an
argument.
> +
> + # todo Should this be separated into two lists for alsc_only?
> + images = []
> + for f in files:
> + color, lux, alsc_only = _parseImageFilename(f)
> + if color is None:
> + continue
> +
> + # Skip alsc image if we don't have an alsc module
> + if alsc_only and not has_alsc:
> + eprint(f'Skipping {fn.name} as this tuner has no ALSC module')
fn isn't defined.
> + continue
> +
> + # Skip non-alsc image if we have only an alsc module
> + if not alsc_only and has_only_alsc:
> + eprint(f'Skipping {fn.name} as this tuner only has an ALSC module')
> + continue
> +
> + # Load image
> + if re.search(r'.dng$', f.name):
> + image = _loadDngImage(f)
> + else:
> + image = _loadBrcmImage(f)
> +
> + if image is None:
> + eprint(f'Failed to load image {fn.name}')
> + continue
> +
> + # Populate simple fields
> + image.alsc_only = alsc_only
> + image.color = color
> + image.lux = lux
> +
> + if 'blacklevel' in config['general']:
> + image.blacklevel_16 = config['general']['blacklevel']
> +
> + if alsc_only:
> + continue
> +
> + # Handle macbeth
> + macbeth = _locateMacbeth(params)
params is undefined.
I think it would be useful if you could run the code and fix these
issues.
> + if macbeth is None:
> + continue
> +
> + images.append(image)
> +
> + if not _validateImages(images):
> + return None
> +
> + return images
I haven't reviewed everything as I still need to wrap my head around the
pieces. Hopefully reviewing the next patches will help there.
Regardless, I think the above review comments should give you enough to
address for v2.
--
Regards,
Laurent Pinchart
More information about the libcamera-devel
mailing list