Over-exposure causes color errors

Over-exposure produces saturation and this also produces color artifacts. We illustrate the effect here by making a sensor that matches the PowerPoint illustration in Psych 221.

In this case, we arrange the color filters so that the G response is larger than the R and B. We set the illuminant correction. Then we allow the saturation to occur and use the same illuminant correction scaling. The white patch becomes purplish, as in the teaching slide.

Copyright Imageval Consulting, LLC, 2015

Contents

ieInit

Set up the default scene and the oi

oi = oiCreate;
scene = sceneCreate;
oi = oiCompute(oi,scene);

Now, to make the effect in the PowerPoint we scale the green sensitivity

% Create a sensor
sensor = sensorCreate;
sensor = sensorSetSizeToFOV(sensor,sceneGet(scene,'fov'),oi);

% Adjust the sensor so that G is more sensitive
f = sensorGet(sensor,'filter transmissivities');
f(:,1) = f(:,1)*.2;
f(:,3) = f(:,3)*.5;
sensor = sensorSet(sensor,'filter transmissivities',f);

% Auto exposure to get the good timing
sensor = sensorSet(sensor,'auto exposure','on');

Compute the sensor response

% With auto exposure, it looks good
sensor = sensorCompute(sensor,oi);
eTime = sensorGet(sensor,'exposure time');
% ieAddObject(sensor); sensorWindow('scale',true);

ip = ipCreate;
% ip = ipSet(ip,'correction method illuminant','gray world');
ip = ipCompute(ip,sensor);
ieAddObject(ip); ipWindow;

% This is how we scale the RGB for this case.
fprintf('The transform from sensor to display RGB\n')
ipGet(ip,'Combined transform')

The transform from sensor to display RGB

ans =

  3×3 single matrix

   1.0e-15 *

    0.2682   -0.0247    0.0129
    0.0199    0.0926   -0.0295
   -0.0685   -0.0042    0.2633

Now increase the exposure and recompute

% Fix the transform.
ip = ipSet(ip,'transform method','current');

% Increase the exposure time and recompute
% Notice that the whitest surfaces becomes purplish in this overexposed case
% The gray surfaces, which are not over exposed, are still gray.
sensor = sensorSet(sensor,'auto exposure','off');
sensor = sensorSet(sensor,'exposure time',3*eTime);
sensor = sensorCompute(sensor,oi);
ip = ipCompute(ip,sensor);
ieAddObject(ip); ipWindow;

Published with MATLAB® R2024a