Method for compression of optical observation data based on analysis of differential structure

Abstract

A method for compression of high-contrast images based on analysis of their differential structures is suggested. The concept of the method, software implementation and additional factors that increase the efficiency of the method are considered. The compression method is applicable to static and dynamic (video) images obtained, in particular, by optical sensors. The principles of compression we propose allow one to create multi-threaded computing solutions at both the hardware-software and hardware levels. The degree of image compression based on differential structure analysis is comparable to the JPEG compression degree, but the image quality is higher by numerical criteria (MSE, PSNR, SNR). The method requires a relatively small number of operations for compression, which reduces the processing load, and can be successfully used at spacecrafts for obtaining and transmitting images.