Research On The Lossless ROI Encoding Technology Of The Visual Measuring Image

In the large-scale precision measurement process, the visual measuring image isthe only known quantity got by the vision measuring system, and it has the directimpact on the measuring result. But the storage and transmission of the vast amountsof visual measuring image data become difficult problems, which hindered the fullautomation of measurement system, meanwhile the special application features of thevisual measuring image proposed the new requirements to the image compressiontechnology. This paper carried out a detailed research on the lossless ROI encodingtechnology of the visual measuring image, and the specific studies is as follows:1. Based on the in-depth analysis of the more successful vision measuringsystems and the image ROI encoding technology, combined with the applicationfeatures of the visual measuring image, this paper proposed the lossless ROI encodingprogram and the precautions, and developed the algorithm evaluation standards whichis suitable for visual measuring image.2. Summarized the precautions in the visual measuring image encoding process,and established complete extraction method of ROI and ROI feature points, by theROI extraction method analysis of the various types of image.3. The lossless ROI encoding scheme of visual measuring image based on thewavelet transformation was proposed. The encoding area divided method made thescheme suitable for the arbitrary size visual measuring image and increased thealgorithm universality. The improved listless MNLS algorithm not only saved thememory space of the three lists of the original SPIHT algorithm but also made full useof the correlation in the brother nodes. The proved method of encoding the ROIcoefficient signs and the way to deal with “0”encoding stream reduced the encodingstream expenditure in the lossless ROI encoding process, the loss method that is onlyencoding a few effective bit-planes in the background area guaranteed the most of thebackground information and increased the whole compression ratio.4. Research emphatically the visual measuring image encoding method based onthe wavelet packet transformation. Fully summarized the wavelet packet coefficientsfeatures and proposed the selection criteria of the best wavelet packet basis, which issuitable for visual measuring image. To the conflict between father and son nodes inthe encoding process of wavelet packet coefficients, this paper proposed theexchanged position method, and encoding results is close to the encoding methodbased on wavelet transformation, but it took a long time. So as to improve the coding efficiency, the Multi-SPIHT coding method was proposed based on the waveletpacket transformation. It adopted the way to parallel process of the multiple spatialorientation trees, ensured the encoding quality and saved two-third of the coding time.5. From the aspect of measurement accuracy experiment, the compression ratioand the coding time experiment, and the comparative experiment to V-STARS, in thelarge-scale three-dimensional control field, verified the Multi-SPIHT lossless ROIencoding of the visual measuring image based on wavelet packet transformation. Theexperiment results demonstrated feasibility, validity and accuracy of the algorithmproposed by this paper.