Research On The Remote Sensing Stereo Pair Compression Algorithm Based On The End Scan Threshold

With the development of satellite remote sensing technology,the scale for people to get spatial data is increasing.The contradiction between the acquire and transmission of remote sensing data is increasingly prominent, this makes the remote sensing image real-time compression becomes a problem needs to be solved immediately. Now the compression technology still haven’t reached the demand of good real-time compression.So the research to remote sensing image real-time compression algorithm has very important practical significance.This thesis uses single remote sensing image and remote sensing stereo pairs as research objects to study the real-time compression technology of the remote sensing image,and then we put forward a compression algorithm of remote sensing stereo image based on the end scan threshold.First of all,this thesis studies the wavelet transform and the lifting scheme.Because the wavelet transform has very good time/frequency characteristics and can also match the human visual system,so it is widely used in the image compression coding area.Because of the traditional wavelet transform refers to floating point operation, computational complexity is very high,so it is not suitable for being used in real-time compression of remote sensing. Therefore we need to use the lifting scheme and integer transform to making up the defects.Then,we use the single remote sensing image as research object to prepare for the reference image compression of the remote stereo pair.To analyze to SPIHT algorithm,it has the problem of rescanning and large storage number in coding.In order to solve these problems,we come up with a improved algorithm.In this algorithm,we bringing in the idea of the end scan threshold and end scan output bit. By pre-setting the end scan threshold and end output bit to solve the problem of long time rescanning and large storage space. At the same time, after the image is compressed by wavelet, the important information mainly exist in the low-frequency part and this offers us possibility to code separately to high-frequency and low-frequency part. Using DPCM to compress low-frequency main information to make them lossless.For the high-frequency part,we use the improved algorithm based on the end scan threshold to promote the compressionratio.Finally,introducing the compression algorithm used in single remote sensing image to remote sensing stereo pair compression.Combining the adaptive disparity estimation and the disparity compensation method of overlapping blocks,we put forward a compression scheme of remote sensing stero pair based on end scan threshold.To illustrate the effectiveness of the improved algorithm,we test the compression performance of the single remote sensing images. When the improved algorithm is compared with SPIHT & JPEG2000,image’s peak signal to noise ratio increased by 2-3dB and the time of coding and decoding decreased by 50% more or less.In the performance test of the synthesis compression algorithm,the scheme of this papaer has improved the predicted image’s peak signal to noise ratio compared with the former 2 ways.And the combined PSNR has improved by 3-4dB.