Research On Image Noise Suppression Method Of Active Laser Imaging

With the development of information technology,people need clearer images.Laser active imaging system has been widely applied because of its high quality of its imaging system.But the image produced by the laser active imaging system can also be affected by the noise contained in the system itself and the outside world.The image will be polluted by noise.The lack of clear images will affect the observer's correct understanding of the information and content of the image.Although the research on image denoising has made great progress in recent years,most of the methods still can not be applied to the image of active laser imaging,so it is of great significance to study the image denoising algorithm of active laser imaging.Laser active images usually contain noise and affect people's observation of images.Because of its transmission characteristics,speckle noise is the main noise.The range gated imaging system is developed from the laser active imaging system.Compared with the laser active imaging system,the range gated imaging system also uses the laser as the light source,but it overcomes the backward scattering light in the atmosphere of the laser active imaging system,and makes the imaging quality greatly improved.Speckle is the main noise in the range gated imaging system.The main work of this paper is to remove noise,especially for speckle noise.In this paper,two denoising algorithms are studied,one is the non local mean algorithm of the representative algorithm in the space domain,and the other is the representative algorithm of the frequency domain.In the study of non local means denoising algorithm,considering the influence of kernel function on the weight value,the kernel function is improved to make the weight allocation more reasonable.According to the original single Euclidean distance to measure the similarity between pixels,the pixels close to the edge are not enough to match the pixels,and the similarity between pixels is measured by the dual standard of Euclidean distance and gray distance.In the study of wavelet transform de-noising algorithm,the phenomenon of ringing of hard threshold at threshold in threshold de-noising algorithm is analyzed,and the soft threshold is contracted to the wavelet coefficients,and the phenomenon of blurring of the edge details in the image is not completely preserved.A new threshold function is proposed in this paper,which inhibits the ringing of the ringing.Image,which makes the coefficients more accurate when the wavelet is reconstructed.Finally,based on the above two improved algorithms,a denoising algorithm combined with non local mean and wavelet transform is proposed,and the image denoising simulation experiments and the corresponding indexes are calculated.Experimental results show that the algorithm can remove most speckle noise effectively and has better denoising effect?...