A Remote Sensing Image Compression Algorithm Based on Adaptive Threshold

Because different wavelet sub-bands contain different image information after a RSI (Remote Sensing Image) being transformed by wavelet, a RSI compression method based on adaptive threshold has been proposed. According to entropy theory, the amount of information a wavelet sub-band contains can be expressed by its entropy, and adaptive threshold of each wavelet sub-band was set depending on the its entropy. To reduce much computation of the entropy, the relation between entropy and other statistic value for each wavelet sub-band was analyzed. And we found the average absolute value had clear and steady relation with the entropy. By curve fitting, the mathematical expression of adaptive threshold for RSI compression was achieved. Experimental results demonstrated that the method had the adaptivity that the image with simple texture could be compressed with high CR (Compress Ratio) and the image with complex texture could be compressed with low CR, and both of the two kinds of compressed images had a good quality.