唐彩虹　张赞波

摘 要： 激光遥感图像乘性噪声的降噪处理中，传统解决乘性噪声的图像降噪技术存在较多的缺陷，故根据乘性噪声的非线性和时变性，提出基于TV模型的激光遥感图像乘性噪声综合降噪方法，在不破坏图像基本结构的情况下进行降噪。该方法使用TV模型针对乘性噪声的非线性和时变性进行能量变分，获取乘性噪声能量极限，综合使用规则化方法和边界约束法对能量极限进行迭代求解，从激光遥感图像的边缘处开始向内进行乘性噪声平滑降噪，很好地保留了图像边缘信息。实验对不同降噪方法进行了对比分析，结果显示，所提降噪方法具备良好的降噪性能。

关键词： 激光遥感图像； 乘性噪声； TV模型； 迭代求解； 图像降噪

中图分类号： TN249?34； TP391 文献标识码： A 文章编号： 1004?373X（2017）20?0098?03

Abstract： The traditional image noise reduction technology of the multiplicative noise has various disadvantages in multiplicative noise reduction of laser remote sensing image. According to the nonlinear and time?varying characteristics of multiplicative noise， the multiplicative noise comprehensive denoising method based on TV model is proposed for laser remote sensing image， which can denoise the image while maintaining the basic structure of the image. In the method， the TV model is used to conduct energy variation for nonlinear and time?varying characteristics of multiplicative noise to obtain the energy limitation of multiplicative noise， the regularization method and boundary constraint method are comprehensively used to conduct iterative solution for energy limitation， and the multiplicative noise is smoothly denoised from the edge of the laser remote sensing image to its interior to preserve the image edge information better. The contrastive analysis of different denoising methods is performed in experiment. The results show that the proposed method has perfect performance of noise reduction.

Keywords： laser remote sensing image； multiplicative noise； TV model； iterative solution； image noise reduction

0 引 言

乘性噪聲一般由信道特征的不规则变化所产生，是制约图像质量的首要问题。激光遥感图像所用的无线电信道中乘性噪声尤其严重，具备非线性和时变性。传统解决乘性噪声的图像降噪技术存在较多的缺陷，如噪声源头检测技术采用结构缺陷检测方法对激光发射源的设备结构进行预判，列举几项源头特征以指导激光遥感图像的生成，却对激光质量造成了一定影响，甚至导致出现新型噪声[1]。时域平均滤波方法只能缓解激光遥感图像的局部乘性噪声，自适应滤波方法在时变性信号的处理中作用不大，故需要新的激光遥感图像乘性噪声降噪技术为日后各行业的应用做出正确指引。……