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

For an image, people only care about the specific parts of the image. Such as UAV reconnaissance images, the military target(military vehicles, tanks, bunkers), which users care about, takes up a small area of the image. To reduce the storage space of image data and recover the military target with high quality, a new method, which called separation compression method, proposed in this thesis. Separating Region of Interest(ROI) and Region of Background(BG) from the UAV reconnaissance image, then compressing the interested military target with little loss or without loss, and compressing the uninterested regions with much loss.1、This thesis analyzes two image compression methods based on Region of Interest of JPEG2000, and compares the advantages and disadvantages of the two methods. The compression methods based on Region of Interest in research, realizing the recovery quality of ROI is higher than BG, but they compress the image as a whole. According to the characteristics and compression requirements of UAV reconnaissance images, a separation compression method proposed in this thesis.2 、 This thesis analyzes the commonly used edge detection method. According to the characteristics of military reconnaissance, choosing an improved Canny algorithm to find the edge of military target in UAV reconnaissance images, then judging the regular closed curve. Determining ROI of UAV reconnaissance images at last.3、This thesis deeply studies separation compression method. According to the established area of military target, choosing a suitable separation method to separate ROI and BG. After separation, we can choose suitable compression method freely. After comparing EZW and SPIHT, select the improved SPIHT algorithm to compress ROI and BG with different compression ratio.4、Emulating the method using MATLAB language in MATLAB7.0 environment. The results of experiments show that the method proposed in this thesis can not only maintain the quality of ROI, but also can adjust the relative quality of ROI and BG freely. The method solves the contradiction between high compression ratio and high quality.