| According to the development of information technology and multimedia technology,the information becomes the main character of the modern society.The main media agent of information is image and so digital image is one of the most important information which is broadly used and the study of image information increasingly becomes a science of great attention. Image coding technology is the heart of image communication system,and so image compression and coding significantly affects the development of image communication technology.In radar data processing,because of the very large data size of a radar image,it must be compressed while maintaining the important information contained before it is stored or transmitted.Therefore,research and development of a new image compression method due to its compression format,storage and transmission is a good choice.Wavelet transform coding technique contributes the radar signal processing to develop towards high resolution,super broad band,multi function and intelligent functioning.In chapter 1,the history of radar image processing developmentwith some radar data processing methods,such as matching filter theory,Fourier transform,Short Time Fourier Transform(STFT),Wigner-Ville Distribution(WVD),Self-Adaptive Filter Technology and Wavelet Transform.Then the application of wavelet transform in radar image processing is introduced,with the history of wavelet transform.Some applications such as Radar Target detection,super broadband radar signal processing,clutter spectrum estimation and anti-clutter,SAR imaging and target recognition and radar frequency compression are presented in the succeeding article.Finally, wavelet base selection and wavelet coefficient coding of wavelet image coding technology is presented.Due to the demand of no watchman in radar network and radar station,how to reliably transmit radar video frequency over the narrow band becomes the problem.In the present radar processing devices,analog processing is used in most radars.But bandwidth used is large, waveform distortion is large,and functioning is unstable.Besides,the signal can be transmitted through the digital communication devices only after it is digitized.If the digitized signal is not compressed,and the bandwidth used is so wide such as 4MHz,8 bit sampling 2MHz analog radar signal,it is necessary 32 Mbps digital communication bandwidth.If 2 channels of video signals are transmitted,64 Mbps is needed.But the present day small radar of 2Mbps bandwidth cannot satisfy with the demand and so the radar image must be compressed before transmission.In image coding field,since the wavelet transform has the ability of description for unstable signals,it causes the wavelet transform based method better image description than discrete cosine transform.Wavelet technology is very suitable for radar image compression.For taking the best advantages for relativity of the nearest periodic signals,radar signals can be compressed by using 2-dimensinal wavelet transform in order to obtain the better compression result.Firstly,the signal is decomposed different frequency portion through the orthogonal filter. Then the image is quantized.Then the quantized image is coded with Huffman encoding.Besides, it has less error and it can control bit rate exactly during the transmission,and so it is satisfied with the demand of fixed bit rate.Error bit during the transmission does not need to stop the whole transmission because it can recover from the later consecutive similar bits,and so masking cannot occur in wavelet transform like in JPEG.In chapter 2,Basis of wavelet analysis is introduced together with wavelet function,scaling function,wavelet packet and types of wavelet transforms,multi resolution analysis(MRA)and Mallat algorithm in mathematical equations.And then,EZW analysis and SPIHT analysis are also presented together with the comparison between these two algorithms.It is learnt that SPIHT method is simpler than EZW and it also has higher efficiency.So SPIHT method is chosen for radar image compression in this thesis.In chapter 3,structure of image transmission system,history and development of image transmission technology,network transmission technology and main transmission networks are introduced.Then some of the international image coding standards are presented.Then various types of computer networks due to their size,speed and topology are introduced.Finally, communication protocols and some important protocols are introduced.The local area network is the most suitable for the radar image transmission system because its structure is simple,its speed is fast and programming is also simple.In chapter 4,the 7 layer OSI model is introduced.Then the comparison between OSI model and TCP/IP protocol is discussed.Due to the simplest structure,less error and faster speed,TCP/IP is applied for the transmission of images between the server and client model.TCP/IP protocol and LAN network structure is used in this thesis.Finally,an experiment is shown and proved that the images can be transmitted through the network with less error and faster speed.TCP/IP is a protocol suite that is composed of many different communication protocols.TCP sets up a connection at both ends and guarantees reliable delivery of the full message sent.It also tests for errors and requests retransmission if necessary.IP is similar to the network layer of OSI model.It tells the address of sender and receiver to each router on the route.In the chapter 5,the details of the step by step sub processes of the radar image compression and transmission are mainly presented.Firstly,analog radar singal is transformed into digital signal by using the A/D converter.The image is converted into the MxN format in order to be convenient for compression.The image format is PGM which is similar to BMP.Then the image is decomposed by using orthogonal wavelet filters group.After 3 level decomposition,the image is compressed using SPIHT algorithm.SPIHT algorithm has 3 main sub processes:initialization, sorting,and thinning.Then the image is coded with arithmetic coding.Then it is transmitted through the LAN using server/client model file transfer process.According to the results,real time radar image compression and transmission is realized.In the last chapter,the conclusion and future works are presented.The experiment results have shown that SPIHT coding has better performance than Embedded Zero Tree(EZW)algorithm in many aspects.Under the low bit rate conditions,the Peak Signal to Noise Ratio(PSNR)values of the SPIHT algorithm are above 30 dB.Besides,the algorithm of SPIHT is simpler than EZW,and so it is faster in processing which means that it is more suitable for real time image coding.Radar image compression for low bit rates while maintaining the high PSNR is achieved using SPIHT algorithm.Radar image transmission through the LAN network using TCP/IP protocol also fulfils the transmission of radar images in faster speed with less error.Due to the fast connection speed of the LAN,real time radar image transmission is realized within 3 seconds of time interval for each of the real time radar images.The experiments have proved that SPIHT has better efficiency and image transmission through the LAN network is realized for real time transmission. |
PDF Full Text Request