Shearlet-based hard-thresholding for interfered infrared image denoising

An infrared imaging device is effected by external electromagnetic interference(EMI), and images appear salt and pepper noise (not excluded blind element and flash point of infrared focal plane arrays) and striping noise. It causes an serious effect on the target detection and tracking. Median filter can not filter out the noise. It urgently needs to resolve the problem. This paper presents a denoising method based on the shearlet transform, and takes hard-thresholding method, to eliminate the infrared image denoising. Simulation results show that this method effectively eliminates the noise, and retains the original image information. At the same time, in order to show superiority, the method is compared with the median filtering denoising, db wavelet denoising, curvelet denoising method in entropy, PSNR, LSCR and so on, which shows advantages.