An automatic ROI coding algorithm based on multi-resolution target detection

Region of interest (ROI) coding technique could effectively compress the image data while the important information could be saved well, and itpsilas very suitable to the compression of remote sensing images. The ROI needs to be defined in advance to the existing ROI coding algorithms, while the ROI of remote sensing images cannot be defined by the users in some conditions. So the application of the ROI coding algorithm is restricted in the remote sensing image compression. To solve this problem, an automatic ROI coding algorithm based on multi-resolution target detection is proposed. According to the edge information of targets in high frequency sub-bands, the ROI could be detected automatically. Then, after ROI regions of different resolutions are matched, the multi-ROI could be identified with different priorities based on the matching degree. Numerical experiments are conducted on a real remote sensing image, and experimental results prove the effectiveness of the proposed algorithm.