[libcamera-devel] [PATCH v2] ipa: raspberrypi: alsc: Limit the calculated lambda values
Laurent Pinchart
laurent.pinchart at ideasonboard.com
Wed Apr 6 11:52:18 CEST 2022
On Wed, Apr 06, 2022 at 08:02:12AM +0100, Naushir Patuck wrote:
> On Tue, 5 Apr 2022 at 15:53, Laurent Pinchart wrote:
> > On Tue, Apr 05, 2022 at 07:57:58AM +0100, Naushir Patuck via
> > libcamera-devel wrote:
> > > Under the right circumstances, the alsc calculations could spread the colour
> >
> > Sounds like the wrong circumstances, not the right ones :-)
> >
> > > errors across the entire image as lambda remains unbound. This would cause the
> > > corrected image chroma values to slowly drift to incorrect values.
> > >
> > > This change adds a config parameter (alsc.lambda_bound) that provides an upper
> > > and lower bound to the lambda value at every stage of the calculation. With this
> > > change, we now adjust the lambda values so that the average across the entire
> > > grid is 1 instead of normalising to the minimum value.
> > >
> > > Signed-off-by: Naushir Patuck <naush at raspberrypi.com>
> > > Reviewed-by: David Plowman <david.plowman at raspberrypi.com>
> > > ---
> > > src/ipa/raspberrypi/controller/rpi/alsc.cpp | 57 +++++++++++++++------
> > > src/ipa/raspberrypi/controller/rpi/alsc.hpp | 1 +
> > > 2 files changed, 43 insertions(+), 15 deletions(-)
> > >
> > > diff --git a/src/ipa/raspberrypi/controller/rpi/alsc.cpp b/src/ipa/raspberrypi/controller/rpi/alsc.cpp
> > > index be3d1ae476cd..a88fee9f6d94 100644
> > > --- a/src/ipa/raspberrypi/controller/rpi/alsc.cpp
> > > +++ b/src/ipa/raspberrypi/controller/rpi/alsc.cpp
> > > @@ -149,6 +149,7 @@ void Alsc::Read(boost::property_tree::ptree const ¶ms)
> > > read_calibrations(config_.calibrations_Cb, params, "calibrations_Cb");
> > > config_.default_ct = params.get<double>("default_ct", 4500.0);
> > > config_.threshold = params.get<double>("threshold", 1e-3);
> > > + config_.lambda_bound = params.get<double>("lambda_bound", 0.05);
> > > }
> > >
> > > static double get_ct(Metadata *metadata, double default_ct);
> > > @@ -610,30 +611,47 @@ static double compute_lambda_top_end(int i, double const M[XY][4],
> > >
> > > // Gauss-Seidel iteration with over-relaxation.
> > > static double gauss_seidel2_SOR(double const M[XY][4], double omega,
> > > - double lambda[XY])
> > > + double lambda[XY], double lambda_bound)
> > > {
> > > + const double min = 1 - lambda_bound, max = 1 + lambda_bound;
> > > double old_lambda[XY];
> > > int i;
> > > for (i = 0; i < XY; i++)
> > > old_lambda[i] = lambda[i];
> > > lambda[0] = compute_lambda_bottom_start(0, M, lambda);
> > > - for (i = 1; i < X; i++)
> > > + lambda[0] = std::clamp(lambda[0], min, max);
> > > + for (i = 1; i < X; i++) {
> > > lambda[i] = compute_lambda_bottom(i, M, lambda);
> > > - for (; i < XY - X; i++)
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + }
> > > + for (; i < XY - X; i++) {
> > > lambda[i] = compute_lambda_interior(i, M, lambda);
> > > - for (; i < XY - 1; i++)
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + }
> > > + for (; i < XY - 1; i++) {
> > > lambda[i] = compute_lambda_top(i, M, lambda);
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + }
> > > lambda[i] = compute_lambda_top_end(i, M, lambda);
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > // Also solve the system from bottom to top, to help spread the updates
> > > // better.
> > > lambda[i] = compute_lambda_top_end(i, M, lambda);
> > > - for (i = XY - 2; i >= XY - X; i--)
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + for (i = XY - 2; i >= XY - X; i--) {
> > > lambda[i] = compute_lambda_top(i, M, lambda);
> > > - for (; i >= X; i--)
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + }
> > > + for (; i >= X; i--) {
> > > lambda[i] = compute_lambda_interior(i, M, lambda);
> > > - for (; i >= 1; i--)
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + }
> > > + for (; i >= 1; i--) {
> > > lambda[i] = compute_lambda_bottom(i, M, lambda);
> > > + lambda[i] = std::clamp(lambda[i], min, max);
> > > + }
> > > lambda[0] = compute_lambda_bottom_start(0, M, lambda);
> > > + lambda[0] = std::clamp(lambda[0], min, max);
> > > double max_diff = 0;
> > > for (i = 0; i < XY; i++) {
> > > lambda[i] = old_lambda[i] + (lambda[i] - old_lambda[i]) * omega;
> > > @@ -653,15 +671,26 @@ static void normalise(double *ptr, size_t n)
> > > ptr[i] /= minval;
> > > }
> > >
> > > +// Rescale the values so that the avarage value is 1.
> >
> > s/avarage/average/
> >
> > > +static void reaverage(double *ptr, size_t n)
> >
> > Could you please use a span ?
>
> Sure, no problem.
>
> > > +{
> > > + double sum = 0;
> > > + for (size_t i = 0; i < n; i++)
> > > + sum += ptr[i];
> >
> > This could become
> >
> > double = std::accumulate(data.begin(), data.end(), 0.0);
> >
> > > + double ratio = 1 / (sum / n);
> > > + for (size_t i = 0; i < n; i++)
> > > + ptr[i] *= ratio;
> >
> > And possibly
> >
> > std::for_each(data.begin(), data.end(), [ratio](double &n){ n *= ratio; });
> >
> > but I'm not entirely sure that's more readable :-)
> >
> > for (double &d : data)
> > d *= ratio;
> >
> > works for me too.
>
> I'd probably prefer the latter for readability.
>
> > > +}
> > > +
> > > static void run_matrix_iterations(double const C[XY], double lambda[XY],
> > > double const W[XY][4], double omega,
> > > - int n_iter, double threshold)
> > > + int n_iter, double threshold, double lambda_bound)
> > > {
> > > double M[XY][4];
> > > construct_M(C, W, M);
> > > double last_max_diff = std::numeric_limits<double>::max();
> > > for (int i = 0; i < n_iter; i++) {
> > > - double max_diff = fabs(gauss_seidel2_SOR(M, omega, lambda));
> > > + double max_diff = fabs(gauss_seidel2_SOR(M, omega, lambda, lambda_bound));
> > > if (max_diff < threshold) {
> > > LOG(RPiAlsc, Debug)
> > > << "Stop after " << i + 1 << " iterations";
> > > @@ -675,10 +704,8 @@ static void run_matrix_iterations(double const C[XY], double lambda[XY],
> > > << last_max_diff << " to " << max_diff;
> > > last_max_diff = max_diff;
> > > }
> > > - // We're going to normalise the lambdas so the smallest is 1. Not sure
> > > - // this is really necessary as they get renormalised later, but I
> > > - // suppose it does stop these quantities from wandering off...
> > > - normalise(lambda, XY);
> > > + // We're going to normalise the lambdas so the total average is 1.
> > > + reaverage(lambda, XY);
> >
> > That seems to make sense, although I wonder how it will affect the next
> > ALSC iteration. I suppose it doesn't matter too much, but have you
> > checked if this change has an effect on the number of iterations of the
> > SOR (I'm mostly curious about the impact on the normal case, not the
> > "right circumstances") ?
>
> This change has no direct impact on the number of iterations run. It limits
> the amount of error "bleeding" across cells through the iterations.
As the code iterates up to n_iter, but stops when the delta is below the
threshold, I thought that restricting the lambda values could result in
breaking from the loop later, as avoid the "bleeding" means changes can
spread more slowly and the delta could stay above the threshold for
longer. I could be wrong of course.
> In general,
> we want to ensure our calibration table is a good fit and our auto correction
> should not deviate much from that table.
>
> > Reviewed-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> >
> > I'll wait for your reply to the above questions before applying, please
> > let me know if you want to send a v3 or if I should make modifications
> > locally.
>
> Happy for you to merge with the suggested changes if you have them
> available locally :)
Not yet, but I can easily fix that.
> > > }
> > >
> > > static void add_luminance_rb(double result[XY], double const lambda[XY],
> > > @@ -737,9 +764,9 @@ void Alsc::doAlsc()
> > > compute_W(Cb, config_.sigma_Cb, Wb);
> > > // Run Gauss-Seidel iterations over the resulting matrix, for R and B.
> > > run_matrix_iterations(Cr, lambda_r_, Wr, config_.omega, config_.n_iter,
> > > - config_.threshold);
> > > + config_.threshold, config_.lambda_bound);
> > > run_matrix_iterations(Cb, lambda_b_, Wb, config_.omega, config_.n_iter,
> > > - config_.threshold);
> > > + config_.threshold, config_.lambda_bound);
> > > // Fold the calibrated gains into our final lambda values. (Note that on
> > > // the next run, we re-start with the lambda values that don't have the
> > > // calibration gains included.)
> > > diff --git a/src/ipa/raspberrypi/controller/rpi/alsc.hpp b/src/ipa/raspberrypi/controller/rpi/alsc.hpp
> > > index 9616b99ea7ca..d1dbe0d1d22d 100644
> > > --- a/src/ipa/raspberrypi/controller/rpi/alsc.hpp
> > > +++ b/src/ipa/raspberrypi/controller/rpi/alsc.hpp
> > > @@ -41,6 +41,7 @@ struct AlscConfig {
> > > std::vector<AlscCalibration> calibrations_Cb;
> > > double default_ct; // colour temperature if no metadata found
> > > double threshold; // iteration termination threshold
> > > + double lambda_bound; // upper/lower bound for lambda from a value of 1
> > > };
> > >
> > > class Alsc : public Algorithm
--
Regards,
Laurent Pinchart
More information about the libcamera-devel
mailing list