Research Of Lifting Wavelet Algorithm And Its Realization On TDS642EVM Board

In recent years, with the development of information and network, theapplied area of video and image obtains great extension. As a result, thetransmission of image gains great attention. The chiefly question of imagetransmission is a quantity of data, but the channel of network transmissionis relatively narrow. To resolve the question, we must compress the data ofimage transmitted in the network. So all kinds of different compressionalgorithms constantly come forth.The first generation coding technology already arrives the limit by thedevelopment of about half century. The rising of second generation codingtechnology that thinks over the vision characteristic of people and paysattention to the image structure leads people to a brand-new field. Thetraditional Fourier transform contacts time region with frequency region,and puts up the character that is difficult to put up in time region infrequency region. Through time-frequency transform, it wipes off theredundancy between data, but Fourier transform can not express some localcharacters in image exactly, for example, the edges and lines. Wavelettransform called mathematic microscope is able to take on the task oftime-frequency transform, and can observe image in any scale. It utilizeshigh distinguishing rate and low distinguishing rate to observe detail andwhole of image. Wavelet transform can separate the high frequencycomponent from low frequency component of image energy, and the energymostly converges low frequency part. It fits vision characteristic of eye verywell, so can be used to compress image effectively.The fast algorithm of discrete wavelet transform is all through researchhotspot. Swelden brought forward Lifting Scheme that did not depend onFourier transform in 1995.Because its complexity is a half of traditionalMallat algorithm, it becomes the mainstream way of calculating discretewavelet transform. Lifting Scheme offers a new and faster realization wayto the first generation wavelet transform. At the same time, Lifting Schemeis not only the basic tool of constructing the second generation wavelettransform, but also makes the construction of the first generation not dependon the structure of Fourier transform., and reduces the difficulty ofconstruction of the first generation. It has been proved that Lifting Schemecan realize all the first generation wavelet transforms.On the basis of research of Lifting wavelet technology and DSPtechnology, this paper empolders a sort of long-distance monitoring system.This system is the offspring of uniting of digital signals technology, imagecompression coding technology and network communication technology. Inhardware aspect, we select the TDS642EVM(multiplex real-time imagepressing platform) produced by TI company, and use C language to compileand debug the code in Code Compose Studio as the software developmentenvironment. While the debugged program is running, the video imagecaptured from video device is compressed by using Lifting waveletalgorithm and then sent to network. The image can be received at the sink'sPC computer and be decompressed to show. It realized the aim ofmonitoring. This system's volume is very small. Its dependability is high.Its stability is good. Its real time is strong and its use of work is low. So itcan be applied abroad in the outdoor environment to monitor.Although there are lots of network monitoring system productions, most ofthem are based on the operating system. The reliability and stability of theproductions are not satisfied, and the hardware occupies large space. Thehardware's cost is high and its operation is complex, it is not suitable forapplying in outdoor environment. So the application field is narrower.The Lifting wavelet transform algorithm is the kernel of the system.Lifting wavelet transform algorithm as a new algorithm has someadvantages including high compression ratio, fast compression rate andoccupying small bandwidth. The algorithm is not international standard. Itis hard to decompress so it is secure. As monitoring system has the securityproblem which should be concerned about, the system should have thefunction of secrecy. Although the digital communication technology hasbeen well developed, the current technology still has shortcomings insecurity. Therefore, wavelet transform algorithm, which has the doublefunction of encryption and saving bandwidth, is selected to meet the need.This system is developed after studying the technology of DSP and usingthe result of studying the wavelet technology in the Windows environment.Through the studying of wavelet compression theory, the author comparedthe compression effect of Lifting wavelet algorithm with the compressioneffect of Embedded Zero Tree Wavelet algorithm, and then selected theLifting wavelet algorithm which compression effect is better and transplantthe algorithm into TDS642EVM. Through compare and selection ofhardware devices performance, after selecting the empoldering board,hardware emulator TDS510 and software empoldering tool, the authorconstructed the hardware platform and configured the software empolderingenvironment, and the algorithm is realized on TDS642EVM.Because the system is faced to application, we must consider thequestion of its popularization. The main aspect which affects production'spopularization is its price, and that the primary part in the price is cost ofhardware. The solving of the question is that the other schoolmatesempolder the DSP board which has the function of disposing image all bythemselves and the board has had rudiment. Debugged for some time theboard will succeed finally and its price will reduce. In the way the systemcan occupy some room in the market competition, and can popularize easily.The system's code needs optimize. Its function needs to be perfectedfurther, for example the transmission of audio frequency signal and thecontrol of pedestal. The challenge and question will be the direction ofeffort for the future.