Generating column offset corrections for image sensors
View Patent ↗An image sensor includes multiple photoactive pixels and multiple dark reference pixels arranged in rows and columns to form a pixel array. A dark signal is read out of one or more dark reference pixels in each column and used to determine a column offset for one or more columns in the pixel array. An offset window is used for each column in the pixel array to define an acceptable maximum dark signal and an acceptable minimum dark signal for each column. The dark signals from each column are analyzed to determine if there are any dark signals outside the offset window. If any of the dark signals are outside the offset window, the dark signal or signals can be compensated for or discarded.
1. A method for determining outlier dark signals when producing column offset corrections used to compensate for column fixed pattern noise in an image sensor, wherein the image sensor includes a plurality of photoactive pixels and a plurality of dark reference pixels arranged in rows and columns to form a pixel array, the method comprising:
obtaining a global offset window for all of the columns in the pixel array, wherein the global offset window defines an acceptable maximum dark signal and an acceptable minimum dark signal for all of the columns;
reading out a dark signal from a given number of dark reference pixels in each column;
determining whether one or more of the dark signals from each column are outside the global offset window, and if so, compensating for each dark signal outside the global offset window;
determining a column offset correction for each column using dark signals within the global offset window; and
after a given number of dark signals have been read out of the pixel array, determining a local offset window for each column in the pixel array, wherein the local offset window defines an acceptable maximum dark signal and an acceptable minimum dark signal for each column.
2. The method of claim 1 , further comprising:
reading out a dark signal from a given number of dark reference pixels in each column;
determining whether one or more of the dark signals from each column are outside a respective local offset window, and if so, compensating for each dark signal outside the local offset window.
3. The method of claim 1 , wherein the acceptable maximum dark signal is based on previously measured maximum dark signal and the acceptable minimum dark signal is based on a previously measured minimum dark signal.
4. The method of claim 1 , wherein compensating for each dark signal outside the global offset window includes substituting a previous dark signal for the dark signals determined to be outside the global offset window.
5. The method of claim 1 , wherein the given number of dark pixels changes dynamically based on noise in the dark reference pixels.