Updating a fixed pattern noise matrix

A method for processing image data, the method comprising:

• updating a fixed pattern noise matrix by the method comprising:
• calculating a first difference between pixel values of a target image in an electronic-image stabilized video stream and pixel values of a first different image in the electronic-image stabilized video stream, thereby obtaining a first differential matrix;
• calculating a second difference between pixel values of the target image in the electronic-image stabilized video stream and pixel values of a second different image in the electronic-image stabilized video stream, thereby obtaining a second differential matrix;
• identifying a set of candidate positions for fixed pattern noise in the fixed pattern noise matrix by:
• locating a first set of positions in the first differential matrix at which a difference of pixel values deviates from a predetermined value,
• locating a second set of positions in the second differential matrix at which a difference of pixel values deviates from the predetermined value,
• finding a set of overlapping positions between the first set of positions and the second set of positions, and
• adjusting the set of overlapping positions based on stabilization data of the target image, wherein the adjusted set of overlapping positions is used as the set of candidate positions for fixed pattern noise;
• and
• updating, for each position in the set of candidate positions in the fixed pattern noise matrix, a fixed pattern noise value at the position, wherein the updated fixed pattern noise value at each position in the set of candidate positions is based on a value at a corresponding non-adjusted position in the first and/or second differential matrix;
• and reducing fixed pattern noise from the image data by subtracting the fixed pattern noise values of the fixed pattern noise matrix from pixel values of the image data.

A non-transitory computer-readable storage medium having stored thereon instructions for implementing a method, when executed on a device having processing capabilities, for updating a fixed pattern noise matrix, the method comprising:

• calculating a first difference between pixel values of a target image in an electronic-image stabilized video stream and pixel values of a first different image in the electronic-image stabilized video stream, thereby obtaining a first differential matrix;
• calculating a second difference between pixel values of the target image in the electronic-image stabilized video stream and pixel values of a second different image in the electronic-image stabilized video stream, thereby obtaining a second differential matrix;
• identifying a set of candidate positions for fixed pattern noise in the fixed pattern noise matrix by:
• locating a first set of positions in the first differential matrix at which a difference of pixel values deviates from a predetermined value,
• locating a second set of positions in the second differential matrix at which a difference of pixel values deviates from the predetermined value,
• finding a set of overlapping positions between the first set of positions and the second set of positions, and
• adjusting the set of overlapping positions based on stabilization data of the target image, wherein the adjusted set of overlapping positions is used as the set of candidate positions for fixed pattern noise;
• and
• updating, for each position in the set of candidate positions in the fixed pattern noise matrix, a fixed pattern noise value at the position, wherein the updated fixed pattern noise value at each position in the set of candidate positions is based on a value at a corresponding non-adjusted position in the first and/or second differential matrix.

A camera comprising:

• an image sensor configured to capture a video stream;
• a motion sensor;
• and
• a processor configured to:
• electronically image stabilize the captured video stream based on data received from the motion sensor, and
• update a fixed pattern noise matrix using a method comprising:
• calculating a first difference between pixel values of a target image in an electronic-image stabilized video stream and pixel values of a first different image in the electronic-image stabilized video stream, thereby obtaining a first differential matrix;
• calculating a second difference between pixel values of the target image in the electronic-image stabilized video stream and pixel values of a second different image in the electronic-image stabilized video stream, thereby obtaining a second differential matrix;
• identifying a set of candidate positions for fixed pattern noise in the fixed pattern noise matrix by:
• locating a first set of positions in the first differential matrix at which a difference of pixel values deviates from a predetermined value,
• locating a second set of positions in the second differential matrix at which a difference of pixel values deviates from the predetermined value,
• finding a set of overlapping positions between the first set of positions and the second set of positions, and
• adjusting the set of overlapping positions based on stabilization data of the target image, wherein the adjusted set of overlapping positions is used as the set of candidate positions for fixed pattern noise;
• and
• updating, for each position in the set of candidate positions in the fixed pattern noise matrix, a fixed pattern noise value at the position, wherein the updated fixed pattern noise value at each position in the set of candidate positions is based on a value at a corresponding non-adjusted position in the first and/or second differential matrix.