Image sensor with hybrid binning
An image sensor may perform hybrid pixel binning. In pixel binning, pixel values from multiple pixels are combined into a single representative binning value. In a hybrid pixel binning scheme, different pixel groups may be binned in different ways in a single image sensor. When the range of values in a pixel group is low (indicating a flat surface), a mean or median binning scheme may be used. When the range of values in a pixel group is high (indicating an edge), a spatial weighting binning scheme may be used. When a pixel group has an intermediate range, a blend of the median/mean and spatial weighting may be used to avoid undesired blinking in the binning output. The hybrid binning scheme may reduce noise while still preserving high-frequency detail.
1. A method of operating an image sensor comprising a plurality of imaging pixels, wherein the plurality of imaging pixels comprises a plurality of pixel groups, wherein each imaging pixel in the plurality of imaging pixels has a corresponding pixel value, and wherein the method comprises, for each pixel group:
determining a range of the pixel values of that pixel group; and
based at least on the determined range, determining a single binning output for that pixel group, wherein determining the single binning output comprises determining the single binning output based at least on a mean of the pixel values of that pixel group, one or more tuning thresholds, a noise type indicator, or one or more noise parameters.
2. The method defined in claim 1 , wherein determining the single binning output comprises determining the single binning output based at least on a median of the pixel values of that pixel group.
3. The method defined in claim 1 , wherein determining the single binning output comprises determining the single binning output based at least on the mean of the pixel values of that pixel group.
4. The method defined in claim 1 , wherein determining the single binning output comprises determining the single binning output based at least on a weighted average of the pixel values of that pixel group.
5. The method defined in claim 1 , wherein determining the single binning output comprises determining the single binning output based at least on the one or more tuning thresholds.
6. The method defined in claim 1 , wherein determining the single binning output comprises determining the single binning output based at least on the noise type indicator.
7. The method defined in claim 1 , wherein determining the single binning output comprises determining the single binning output based at least on the one or more noise parameters.
8. The method defined in claim 1 , wherein determining the single binning output comprises determining a noise prediction parameter that is a function of the one or more noise parameters and a median of the pixel values of that pixel group.
9. The method defined in claim 8 , wherein determining the single binning output comprises determining an intermediate value that is a function of the range, the noise prediction parameter, a first tuning threshold of the one or more tuning thresholds, and a second tuning threshold of the one or more tuning thresholds.
10. The method defined in claim 9 , wherein determining the single binning output comprises using the median as the single binning output when:
the first tuning threshold is equal to a maximum value for the first tuning threshold; or
the range is less than or equal to the first tuning threshold multiplied by the noise prediction parameter.
11. The method defined in claim 10 , wherein determining the single binning output comprises using a weighted average of the pixel values of that pixel group as the single binning output when:
the second tuning threshold is equal to 0; or
the range is greater than the second tuning threshold multiplied by the noise prediction parameter.
12. The method defined in claim 11 , wherein determining the single binning output comprises using a sum of the median and the intermediate value multiplied by the median subtracted from the weighted average when:
the first tuning threshold is not equal to the maximum value for the first tuning threshold;
the range is not less than or equal to the first tuning threshold multiplied by the noise prediction parameter;
the second tuning threshold is not equal to 0; and
the range is not greater than the second tuning threshold multiplied by the noise prediction parameter.
13. The method defined in claim 1 , wherein the imaging pixels in each respective pixel group are a single color.
14. The method defined in claim 1 , wherein each pixel group of the plurality of pixel groups comprises four pixels of a single color and wherein the four pixels of the single color are arranged in two rows and two columns.
15. The method defined in claim 1 , wherein the plurality of imaging pixels comprises red imaging pixels, green imaging pixels, and blue imaging pixels in a Bayer mosaic pattern.
16. A method of operating an image sensor comprising a plurality of imaging pixels, wherein the plurality of imaging pixels comprises a plurality of pixel groups, wherein each imaging pixel in the plurality of imaging pixels has a corresponding pixel value, and wherein the method comprises:
for a first pixel group of the plurality of pixel groups:
determining a first range of the pixel values of the first pixel group; and
using a median of the pixel values of the first pixel group as a binning output for the first pixel group; and
for a second pixel group of the plurality of pixel groups:
determining a second range of the pixel values of the second pixel group, wherein the second range is greater than the first range; and
using a weighted average of the pixel values of the second pixel group as a binning output for the second pixel group.
17. The method defined in claim 16 , further comprising:
for a third pixel group of the plurality of pixel groups:
determining a third range for the pixel values of the third pixel group, wherein the third range is between the first and second ranges;
determining a median of the pixel values of the third pixel group;
determining a weighted average of the pixel values of the third pixel group; and
using a binning output for the third pixel group that is a function of both the median and the weighted average.
18. The method defined in claim 16 , wherein the plurality of imaging pixels comprises red imaging pixels, green imaging pixels, and blue imaging pixels in a Bayer mosaic pattern.
19. The method defined in claim 18 , wherein each pixel group in the plurality of pixel groups includes imaging pixels of a single color.
20. An image sensor, comprising:
an array of imaging pixels comprising a plurality of pixel groups, wherein the array of imaging pixels comprises red imaging pixels, green imaging pixels, and blue imaging pixels in a Bayer mosaic pattern; and
binning circuitry configured to receive pixel values for each imaging pixel in the array of imaging pixels and, for each pixel group:
determine a range of the pixel values of that pixel group; and
based at least on the determined range, determine a single binning output for that pixel group.