Infrared Image Denoising Algorithm Based on Least Square Fitting
In order to remove the impulse noise in the infrared image and more effectively preserve and restore the edge information and texture features of the image,an infrared image denoising algorithm based on least square fitting is proposed.Noise detection is per-formed according to the prior model of impulse noise and the local deviation of pixels,and pixels that meet the prior conditions of noise and have large local deviation are recognized as noise.For each noisy pixel,its neighboring noise free pixels are used for least square fit-ting,and the median value of the fitting function is used as the intensity estimation value of the current noisy pixel.Experimental results show that compared with some existing methods,the visual effect of image denoised by the proposed method is clearer,the PSNR and FSIM obtained by the proposed method are higher,being 1.5 dB and 0.7％higher than those of the existing methods,respectively,thus,the proposed method has better denoising performance.
infrared image denoisinglocal deviationleast square fittingfeature similarity index
万方数据
维普
