Compression Algorithm, Based On The Content Characteristics Of Interference Images

The imaging spectrometer is widely used in remote sensing detection. Form the end of twentieth century, the spatially modulated Fourier transform imaging spectrometer receives more and more attention .Compared to the ordinary image the interference pattern has it's own characters. Based on the characters of the interference pattern two kinds of algorithm is brought forward in this paper for compressing the interference pattern.As the base theory of the dissertation some basic knowledge of image compression is introduced, including how to acquire interference pattern and its charecters of both space correlation and spectral correlation.In this paper we introduced image compression algorithm of difference between spectrum to interference pattern. The algorithm is simple, easily hardware implementation, widely use and can effectively reduce spectral correlation. but it is difficult to increase compression ratio.The paper presents two compression algorithm of interference pattern: interference pattern classification algorithm and interference pattern space compression encode algorithm. The two algorithms are both compress data by reducing space correlation of interference pattern. In interference pattern classification algorithm , we classify a frame interference pattern by comparing the distance of two strip to specified bound(threshold) in turn. If the distance is lower than threshold ,they belong to the same sort. The choice of the threshold is very important. Here we present two methods of classification and deduce both formulas of each threshold. In the fourth chapter we present the compression flow and code-stream structure of interference pattern space compression encode algorithm in detail. The algorithm is lossless, with the features of simple and easily hardware implementation. The compression ratio of this algorithm is depending space correlation of interference pattern, so it is suitable for deep space and ocean remote sense detection.