| Synthetic Aperture Radar(SAR)is widely used in missile guidance.The amount of imaging data of missile-borne SAR in the flat,large front,down and last stages of the missile is quite large.In order to reduce the amount of information,SAR data can be processed by reducing the image resolution to the corresponding resolution before transmitting and storing in different application scenarios.Even so,a large amount of channel resources still need to be occupied when the seeker transfers SAR data to the flight control computer or the ground control center.Therefore,it is of great significance to study the algorithm of SAR image compression.In this thesis,some research work focused on the compression algorithm suitable for missile-borne radar image,including the intra-frame compression algorithm and the inter-frame compression algorithm for the radar image.This thesis carried out the theoretical study and experiments every stage of image compression,developed a set of local optimal combination algorithm,based on which,the intra-frame compression algorithm and the inter-frame compression algorithm for radar image were completed,besides,the corresponding experiments were carried out,which verifying the research results of the radar image compression algorithm in this paper.The following is the main research content of this thesis:(1)Based on the analysis of the principle and characteristics of SAR imaging,it is found that there exists intra-frame redundancy and inter-frame redundancy in SAR image.By comparing the physical meaning and the gray histogram of the SAR image to the natural image,it is found that SAR image is not exactly the same as that of the natural image.In order to compress SAR image effectively,according to its own characteristics,the algorithm based on wavelet transform is used to study the intra-frame compression algorithm of the radar image after studying and comparing the advantages and limitations of the compression algorithms based on discrete cosine transform(DCT),the compression algorithm based on wavelet transform.Among the four compression algorithms,Differential Pulse Code Modulation(DPCM),conditional replenishment,conditional subsampling and motion compensation,the most powerful motion compensation algorithm is selected for radar image inter-frame compression algorithm.In this thesis,the evaluation standard of image compression coding is introduced in detail.For the evaluation of image compression coding parameters,the compression ratio was selected as the evaluation index in this paper.For the evaluation of image fidelity standard,the PSNR in the objective evaluation was selected as the evaluation standard of this thesis.For the evaluation of the time complexity of theimage compression coding algorithm,the total codec time was selected as the evaluation criterion in this thesis.The selection of the evaluation index of image compression coding provides a theoretical basis for the research and selection of the research object compression algorithm in this thesis.(2)The implementation of intra-frame compression algorithm for radar image based on wavelet transform was introduced in detail.After that,the three stages of the compression model were introduced in detail:wavelet transform,quantization and entropy coding.For the stage of wavelet transform,the lifting wavelet transform was adopted and the Db9/7 wavelet basis was used.For the quantization stage,Set Partitioning in Hierarchical Trees(SPIHT)with better performance was adopted.For entropy coding,the optimal arithmetic coding algorithm was used.Finally,a set of intra-frame image compression algorithms,which is suitable for this research object was completed.(3)The implementation steps of the inter-frame prediction compression algorithm based on motion compensation were introduced in detail,and the algorithms used in each step were introduced and selected.In this thesis,the study of motion estimation was carried out according to the experimental environment and the way of realization of the research objects.And the motion estimation algorithm based on block matching was adopted.In the study of finding the best matching block,the absolute error and criterion,the three step search algorithm and the half pixel estimation precision were used.In the study of motion vector prediction,the motion vector time domain prediction was adopted.Finally,the scheme of inter-frame prediction compression algorithm based on motion compensation was completed.(4)In this thesis,the experiment is carried out by the MATLAB platform on the PC with the Windows7 operating system.Three radar maps,which are 256*256,512*512 and 1024*1024,were compressed and simulated using wavelet transform based intra-frame compression algorithm for radar image.It can ensure that the PSNR is more than 28dB,and the PSNR is 28.8388dB,28.3077dB and 28.6403dB and the compression ratio is 25.0616,18.0789,and 13.6517.The time consuming of the algorithm is 1.8096s,5.6004s and 20.9665s.In this thesis,SPIHT based on the wavelet transform was compared with EZW based on the wavelet transform.The compression ratio is increased by 45.10%and the time consumption is reduced by 41.63%when the peak signal-to-noise ratio loss is 2.34%.The innovation of this thesis is to combine the intra-frame compression algorithm based on wavelet transform and SPIHT algorithm with the inter-frame compression algorithm based on inter-frame prediction based on motion compensation,which can not only give full play to the superiority of the former to radar image inter-frame compression,but also give full play to the advantage of the latter,the inter-frame compression algorithm,which is similar to the H.26x series compression standard,to continuous-frame radar image inter-frame compression.In the first experiment,five continuous frame radar images of the size of 256*256 were compressed and simulated using image compression algorithm for inter-frame prediction of radar based on motion compensation.It can be guaranteed that the average PSNR is 32.7896dB and the compression ratio is 16.4695.The total time consuming of the algorithm is 11.4817s.Compared with the intra-frame compression algorithm based on wavelet transform,in the case of PSNR decreasing 0.6736dB,the compression ratio is increased by 6.4233,and the time consuming of the algorithm is reduced by 3.8907s.In the case of time consuming increased 1.135s,the PSNR increased 0.3111dB,and the compression ratio increased by 0.4655 compared with the H.265 algorithm.Through the comparison of the results and the experimental radar image of the first experiment and the second experiment,the differents between images in the distribution of the background pixel value of the radar map and the target object will result in the different correlation between the pixels of the image.The worse the correlation is,the worse the compression effect will be.The image compression algorithm for inter-frame prediction of radar based on motion compensation meets the compression requirements of the radar image in the project to ensure that the peak signal to noise ratio is not less than 28dB,and the compression ratio is greater than 10 in this thesis. |
PDF Full Text Request