论文部分内容阅读
针对激光主动成像图像特点及实际应用需要,提出了一种基于同态滤波与双数复值小波变换级联的图像降噪算法。首先通过同态滤波将乘性散斑噪声变换为加性噪声;然后用基于改进Qshift滤波器的双树复值小波对含噪图像进行分解,通过Bayes自适应阈值法修正小波系数;最后再进行相应的逆变换得到去噪图像。该算法具有近似平移不变性、多方向选择性及精确重构性,采用信噪比(SNR)、峰值信噪比(PSNR)和运行时间作为算法去噪性能的评价标准进行实验。实验结果表明该算法能够有效抑制图像中的散斑噪声,计算效率高,且很好地保护了图像细节。
According to the characteristics of laser active imaging and its practical application, an image denoising algorithm based on homomorphic filter and double complex valued wavelet transform cascade is proposed. Firstly, the speckle noise is transformed into additive noise by homomorphic filtering. Then the noisy image is decomposed by the double-tree complex valued wavelet based on the improved Qshift filter, and the wavelet coefficients are modified by Bayes adaptive thresholding method. Finally, The corresponding inverse transform gets the denoised image. The algorithm has approximate translational invariance, multi-directional selectivity and exact reconstruction. The SNR, PSNR and running time are taken as the evaluation criteria for the denoising performance of the algorithm. Experimental results show that the proposed algorithm can effectively suppress the speckle noise in the image and has high computational efficiency and good image detail protection.