基于冗余字典的声纳图像斑点噪声去除算法研究通信与信息系统专业论文.docxVIP

优秀毕业论文 精品参考文献资料 AbstraclAbstract Abstracl Abstract Due to the particularity of underwater environment，sonar imaging quality is seriously affected by the absorption，scattering and reverberation，especially the backward scattering and reverberation that make strong noise，seriously degrading the quality of the sonar image，which makes it difficulty tO the subsequent image processing and applications．Therefore，how to construct de。noising algorithm according tO the characteristics of underwater image has great significance Firstly,this paper analyzes the influence factors and imaging characteristics of sonar image．Compared with the traditional optical image，sonar image has characteristics of serious speckle，low resolution，edge blur，low contrast，etc According to the imaging mechanism，the statistical distribution model of speckle is obtained，and established the speckle Rayleigh distribution multiple model Subsequently,according to the sonar image’s characteristics，the method based on K．SVD dictionary learning is applied in the sonar image de—noising．Considering the similarity characteristics of the sonar imaging feature and the single target of sonar image，One way is obtaining global dictionary from few high quality sonar image，then de—noising by using the dictionary．Another way is directly training in noised image and obtaining the adaptive dictionary,then reconstructing the image Experimental results show that two kinds of processing methods in edge preserving and speckle reduction achieve good performance Finally,this paper proposed adaptive dictionary learning with multi—resolution characteristics based the research of over-complete dictionary and wavelet analysis The new dictionary not only has the characteristics of over—complete dictionary,but also inherited the characteristics of the wavelet analysis．Experiments show that compared with K—SVD training dictionary,the proposed method has better performance to speckle reduction and significantly improve the time ef