Image processing method based on HDR

Abstract

The invention discloses an image processing method based on an HDR. The method comprises the steps of obtaining high-bit-wide image data collected by image collection equipment; conducting bit interception operation on each frame of high-bit-wide image of the image data according to a set sequence to generate at least three low-bit-wide images with a set amount of bit data and different brightness; obtaining image quality parameters of each image, and calculating corresponding fusion weight factors; performing downsampling processing on pixel values and the fusion weight factors of the pixel values, and conducting weight fusion on images in the same downsampling dimension; performing downsampling reverse operation on a weight fusion result to generate a frame of low-bit-wide high dynamic range image. According to the image processing method based on the HDR, the high-bit-wide images can be mapped into the low-bit-wide images, the dynamic range and resolution of an initial image can be reserved, meanwhile the adopted images of the different brightness all come from the same exposure frame, there is no need of registration operation, and the method has the advantages of being simple in calculation, high in instantaneity and the like.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an HDR-based image processing method. Background technique [0002] At present, general image sensors or CMOS image sensors have a high sampling bit width. However, the display sampling bit width of most displays is only 8 bits, which is lower than the sampling bit width of the image sensor. Therefore, in order to adapt to the display bit width of the display, the original high-bit-width image is usually truncated during image processing, for example, the upper 8 bits are truncated to generate 8-bit data. However, such a processing process will cause a loss to the dynamic range of the original image. [0003] The existing HDR algorithm adopts the line interleaved high and low exposure method or uses different exposure methods of multiple frames to obtain the original data of multiple frames with different brightness, and then uses a suitable fusion method to fus...

AI Technical Summary

Problems solved by technology

The way of interlaced row exposure will cause the loss of image resolution, such as the way of high exposure (high brightness) of odd lines and low exposure (low brightness) of even lines, the vertical resolution of high brightness and low brightness images is half of the original resolution

Multiple frames with different exposure methods, because multiple frames are arranged sequentially in time, if the target moves quickly, images of different exposures need to be registered before fusion, image registration will reduce the real-time performance of shooting

Images