Radially-based chroma noise reduction for cameras

A method of reducing noise in an image using an image processor comprising:

• receiving image data from an image sensor, wherein the image data is representative of an image and comprises a plurality of pixels and a center point, wherein each pixel has a color value, and wherein the center point is representative of the center of the image sensor's light intensity falloff function;
• blurring each pixel within at least a first portion of the plurality of pixels, wherein the blurring of a single pixel within the first portion of the plurality of pixels comprises:
• computing a blurred value for the single pixel using a formula that only considers other pixels within the first portion of the plurality of pixels that are within a predetermined distance from the single pixel and that possess a color value that differs from the color value of the single pixel by less than a threshold value, wherein the threshold value for the single pixel is determined at least in part by the distance between the single pixel and the center point;
• transforming the color value of each pixel in the first portion of the plurality of pixels to be equal to the computed blurred value for the pixel;
• and
• storing the transformed color value of each pixel in the first portion of the plurality of pixels in a memory.

An apparatus, comprising:

• an image sensor;
• memory operatively coupled to the image sensor for receiving image data, wherein the image data is representative of an image and comprises a plurality of pixels and a center point, wherein each pixel has a value of interest, and wherein the center point is representative of the center of the image sensor's light intensity falloff function;
• and
• a processor unit operatively coupled to the memory, the memory comprising instructions for causing the processor unit to:
• read a first portion of the image data corresponding to a first plurality of pixels, wherein one pixel within the first plurality of pixels is a first center pixel;
• compute a first blurred value for the first plurality of pixels, wherein the first blurred value is based at least in part on the value of interest of each pixel in the first plurality of pixels having a value of interest that differs from the value of interest of the first center pixel by less than a threshold value, wherein the threshold value for the first center pixel is determined at least in part by the distance between the first center pixel and the center point;
• transform the value of interest of the first center pixel to be equal to the computed first blurred value;
• and
• store the transformed value of interest of the first center pixel in the memory.