Research On Channel Coding Adopted For Panorama Image On Region Of Interest

Vision image has been applied in the military, public etc, which is a very important information source. The high quality, the high efficiency transmission and receiving of vision image also have been one of the popular research topics in the communication. The information of panorama image is more than that of common image, the image code of panorama image is longer than that of panorama image, and, when vision environment have been known ordinarily, the importance of the information in the image are different. Owing to this, can treat to the different regions of a image with different methods in the transmission of panorama image. Usually the important regions of panorama image are called region of interest, the not important regions of panorama image are called background region. In this thesis, the studies focus on the channel code of region of interest in panorama image.Aiming at the characteristics of panorama image, the image coding method according to LDPC channel coding method is adopted in the thesis, based on the analysis of all kinds of the image coding. The method is used to code the panorama image on ROI. First, the structure methods of LDPC codes are studied in the article. The algorithms which can encode quickly are introduced. And then LDPC decoding algorithms are researched, a improved BP decoding algorithm is proposed. Finally, ROI is protected with unequal error protection method. The image is decoded with the improved BP decoding algorithm. The image transmission is completed. The transmission method has good performance.That is suitable to image transmission with ROI.The performance of improved BP decoding algorithm is better than that of the BP-Based algorithm. An appropriate correction factor can enhance the rate of decoding convergence, and have better performance than BP decoding algorithm. But decoding number of times isn't increased.Using LDPC code which has long code length and low code rate on ROI and using LDPC code which has short code length and high code rate on background, which cause that the image quality of ROI is better than the image quality of background, when the whole image is preserved.