Research On The Compression Algorithm Of UAV Reconnaissance Image Based On Region Of Interest

Region of Interest(ROI) image coding technique means to compress the interesting regions in an image with little loss or without loss, and to compress the uninteresting regions (region of background,BG) with much loss.Based on this idea a high compression ratio can be obtained and the important information can be preserved in the image. Scince people always feel more interested in some particular areas of an UAV reconnaissance image and require higher quality of these areas than other areas in the real application, the ROI image coding teehnique is applied to the compression of UAV reconnaissance images in this paper.First, the research background and the developments are discussed in this paper. Secondly, this paper elaborates comprehensively on the wavelet transform theory, analyzes the distribution characteristic of the image wavelet coefficients after wavelet transform and then discusses two embedded coding algorithms based on wavelet transform in detail : the EZW algorithm and the SPIHT algorithm. After that, this paper deeply studys the two ROI coding algorithms based on bit plane shift technique in JPEG2000 and the ROI mask generation technology, puts forward an improved bit plane shift algorithm and an improved SPIHT algorithm according to the characteristics of wavelet coefficients after bit plane shift, then realizes the total compression method using C++ language in VC++ environment.The results of experiments show that the compression method proposed in this paper can not only maintain the ROI image quality very well,but also can obtain good BG image quality in low bit rate(less than 0.5bpp). Finally, this paper also puts forward two kinds of ROI image coding algorithm which are not based on bit plane shift and realized by MATLAB simulation software.