Research On Digital Holographic Denoising Technology Based On Multi-Look And Block Matching Filtering

Digital holography combines traditional optical holography and digital image processing technology,which can record and display three-dimensional information of objects,and has been widely used in many fields.Because the laser source has high coherence,it irradiates objects on optically rough surfaces,causing random interference of scattered light and forming speckle noise.Unlike non-coherent additive noise,speckle noise is a multiplicative noise and is very difficult to suppress.The existence of speckle noise seriously affects the quality of the reconstructed image.Therefore,the removal of speckle noise is essential in the field of digital holography.In this paper,the cause of speckle noise is first studied.From the statistical characteristics of speckle,it is known that speckle contrast is inversely proportional to N,where N is the number of speckle patterns.Therefore,multi-look digital holography can be used to suppress speckle noise.Optical multi-look digital holography requires a light path to collect multiple images.The device is complex and time-consuming.In digital multi-look,a random resampling mask is used to simulate a moving diffuse reflector.The mask mainly includes a fair constraint mask and a random overlapping mask.The two masks can reduce the speckle contrast of the image.The digital filtering algorithm is studied.The non-local mean filtering is based on the idea of weighted average.The algorithm is applied to digital holographic images,and the integral image method is used to accelerate the algorithm speed.The algorithm speed is increased by 10 times or more;Based on the non-local mean filtering,the 3D block matching filtering algorithm is studied,and the algorithm is optimized to reduce its complexity.Finally,the non-local mean filtering is compared with the 3D block matching filtering algorithm,and the peak signal-to-noise ratio and structural similarity are calculated,quantitative evaluation of the algorithm's speckle suppression effect.The speckle contrast in multi-look digital holography is constrained by the ideal curve.After the contrast reaches saturation,the number of images increases and the denoising performance is basically unchanged.3D block matching filtering algorithm reduces denoising performance due to incorrect grouping when noise is large.Combine the twomasks in the multi-look digital holography with the three-dimensional block matching filter algorithm to form a multi-look block matching filter algorithm,and evaluate the noise suppression effect according to the evaluation function.The combination of the fair constraint mask and the digital filtering algorithm has a certain degree of improvement in the denoising effect.When combined with the three-dimensional block matching filtering algorithm,the peak signal-to-noise ratio can be increased by 1d B,it can still maintain good visual effects when noise is large.The random overlapping mask and digital filtering algorithm have better joint denoising effect.Combined with the non-local mean algorithm,the peak signal-to-noise ratio is increased by 2.78 d B;combined with the three-dimensional block matching filtering algorithm,the peak signal-to-noise ratio can be increased by 4.06 d B,and the structural similarity has reached 0.99.The similarity between the filtered image and the original image is very high.Finally,the speckle contrast of the denoised image is calculated.The joint algorithm breaks the ideal limit of speckle contrast in multi-look.At the same time,it also solves the problem of incorrect grouping in the three-dimensional block matching filtering algorithm.Achieved a good visual effect when the image signal-to-noise ratio is relatively low,the structural similarity reached 0.99.Therefore,the joint algorithm can effectively suppress speckle noise.