Systems and methods for generating high dynamic range images

A method for determining a pixel value in a high dynamic range image from N input images of different brightness, where N is an integer greater than two, comprising:

• forming a list of pairs of input pixel intensities, each of the input pixel intensities corresponding to the same location in the N input images, and wherein all pairs in the list of pairs, except for the first pair in the list of pairs, comprise one of the input pixel intensities in the preceding pair in the list of pairs and one of the input pixel intensities not represented in any of the preceding pairs, each of the N input images having lower dynamic range than the high dynamic range image;
• for each pair in the list of pairs, determining relative combination weights for each of the input pixels intensities, each of the relative combination weights being a continuous function of both of the input pixel intensities in the pair, the function having predefined values for a plurality of pairs of predefined input pixel intensities, the predefined values being defined by consideration of at least one quality metric;
• applying a normalization condition to the relative combination weights to determine absolute combination weights,
• obtaining N brightness parameters respectively associated with the N input images;
• correcting for sensor-induced errors in at least one of the brightness parameters;
• and
• calculating the pixel value in the high dynamic range image as a weighted average of the input pixel intensities using the absolute combination weights, wherein the step of calculating comprises scaling, based on the brightness parameters, the input pixel intensities to a common image brightness.

A system for generating a high dynamic range image, comprising:

• a processor;
• a memory communicatively coupled to the processor and including an image storage for storing first and second input pixel intensities from a first input image and second input image of lower dynamic range than the high dynamic range image;
• a combination weight module, including machine readable instructions stored in the memory, that when executed by the processor, perform the function of determining a first combination weight for the first input pixel intensity and a second combination weight for the second input pixel intensity, each of the first and second combination weights being a continuous function of the first and second input pixel intensities, the function having predefined values for a plurality of pairs of a predefined first pixel intensity and a predefined second pixel intensity, the predefined values being defined by consideration of at least one quality metric;
• and
• a pixel value calculation module, including machine readable instructions stored in the memory, that when executed by the processor, perform the function of (a) obtaining a first brightness parameter associated with the first input image and a second brightness parameter associated with the second input image, (b) correcting for sensor-induced errors in at least one of the first and second brightness parameters, (c) scaling, based on the first and second brightness parameters, the first and second input pixel intensities to a common brightness, and (d) calculating a pixel value in a high dynamic range image as a weighted average of the scaled first and second input pixel intensities using the first and second combination weights.

A system for generating a high dynamic range image, comprising:

• a processor;
• a memory communicatively coupled to the processor and including an image storage for storing input pixel intensities from N input images, N being an integer greater than two, each of the N input images having lower dynamic range than the high dynamic range image;
• a combination weight module including machine readable instructions stored in the memory, that when executed by the processor, perform the functions of (a) forming a list of pairs of input pixel intensities, each of the input intensities corresponding to the same location in the N input images, and wherein all pairs in the list of pairs, except for the first pair in the list of pairs, comprises one of the input pixel intensities in the preceding pair in the list of pairs and one of the input pixel intensities not represented in any of the preceding pairs, (b) determining relative combination weights for the input pixel intensities, each of the relative combination weights for a pair in the list of pairs being a continuous function of both of the input pixel intensities in the pair, the function having predefined values for a plurality of pairs of predefined input pixel intensities, the predefined values being defined by consideration of at least one quality metric, and (c) applying a normalization condition to the relative combination weights to determine absolute combination weights;
• and
• a pixel value calculation module including machine readable instructions stored in the memory, that when executed by the processor, perform the functions of (a) obtaining N brightness parameters respectively associated with the N input images, (b) correcting for sensor-induced errors in at least one of the brightness parameters, (c) scaling, based on the brightness parameters, the input pixel intensities to a common brightness, and (d) calculating a pixel value in a high dynamic range image as a weighted average of the scaled input pixel intensities using the absolute combination weights.