Research On Composite Compression Algorithm Of Large Size Optical Image Sequence

It is an important branch of the image processing field that the technology of the imagecompress coding,and having an extensive application in transmission,remotesensing,medical,military affairs,and etc every field.With the technology progressing,whilestudying the algorithm of the image compressing,not merely how to won the more greatercompression rate of the image is considered,it is even more be cared that the visual effect ofthe image comeback.How to have concurrently or in pursue a best balance between them,it isthe greatest difficult problem to be faced that the technology of image compresses atpresent.This paper in allusion to the large size image sequences of optical measure,improvingand optimizing it based on the algorithms of image compression which advanced atpresent,have researched on the technology of the Regions of Interest distilled and automaticsearch and tail、layered compression of Back Ground as well as composite compression withcoding stream structure and so on.The mostly works and innovations are concluded asfollows:1.Introduced and summarized the image encoding technology which researching statusand developing trend in the international and at home, elucidated the purpose and significanceof this research,explained the principles and applications of the image coding technology,targeted introduced several commonly used encoding methods and image quality evaluatecriterion,and laid the foundation for the following program design and the formation of thesystem;2.Designed a scheme which automatically extracted and lossless compression in theregion of interest of image sequences.Using the (5,3) integer wavelet to multi-resolutionanalysis for the original image, then based on the watershed transform algorithm and thedouble cross search algorithm to achieve the ROI of automatic detection and tracking,andgenerating the ROI template,finally using the huffman algorithm to lossless encoding.Theprogram has a high searching accuracy and the efficiency which is higher nearly40%thanconventional methods, the treated ROI area fully meet the image processing accuracy;3.Designed a high proportion lossy compression scheme for the background area andmeet the human visual characteristics.According to the characteristics of the backgroundregion,and based on wavelet coiffient partion algorithm,wavelet coefficients, the joint (9,7) wavelet and bandelet transform layered compression to low-frequency texture and highfrequency details for the background area.Not only achieved a high proportioncompression,but also the refer detail informations have the greatest extent retained, and sothe image can to to, the entire image’s redundancy minimize,the rate-distortion to aminimum,and meet the needs of hunman eyes;4.Build a optical measurement image sequence encoding and decoding system,and hadoptimized the system.Organized the multi-frame continuous compression images to optical toform a continuous image data stream. In allusion to the character of the large size imagesequences,have proposed a optimize scheme of the quantization step of the self adaptativequantization coding、a operation method of adopting the multiplication of integers and shift tosimilar to the floating-point and a control algorithm of the compression rate counted on thebasis of the code passway,reduced the coding time,improved the computational speed,andincreased efficiency of compressing.This research can be better solve the compression problems of the large size imagesequences of optical measure,can compress the images based on protecting the sensitive areasin the picture,provide a convenience to the large size image sequences of optical measure forsaving、transmiting and broadcasting,and improve the use efficiency of the images.Thisalgorithm is easy to coding and realize, have a high efficiency of compression,can be appliedin the digital images procession and compression extensively.