| 摘要 |
Creating a High Dynamic Range (HDR) image within a digital camera comprising mixing a plurality of images captured at different exposures, registering counterpart pixels together, deriving a normalized exposure level for each image, blending first and second images to generate an intermediate image and, when two images are used, the intermediate image is a mixed output image and, when more than two images are used, the image blending process is repeated using the previous intermediate image instead of the first image and another image instead of the second image until all images are blended when a final intermediate image is the mixed output image. The HDR image may be displayed on the camera display. Luma values are normalized to a specific standard deviation and mean, local motion detected between a reference image and a comparison image, comparison image pixels having local motion clustered into patches, and blended using a look-up table weighting value computed from the binary connected image to produce an output image. Degamma conversion may convert the images from a lower bits-per-pixel format (8-bit) to a higher bits-per-pixel format (16-bit) prior to blending. A ghost remover corrects for movement during series image capture.
|
