Abstract: To study the quality effect of the remote sensing image brought by the image compression process, several algorithms of objective quality assessment about the visual and geometric quality of the remote sensing image are discussed. A sub-pixel level corner detection algorithm based on the Harria corner detection algorithm is improved. Two compression algorithms of JPEG2000 and SPIHT are used to verify these assessment algorithms. The experimental results show that the image compression rate can be no larger than 16 if you want to distinguish the details by your eyes, and it can be no larger than 8 if you need satisfactory precision in computer vision applications.

Key words: image compression, image quality assessment, corner detection algorithm, compression rate