Research On The Restoration Of Focal Plane Image Based On Double Digital Cameras

Image is an important way for humans to obtain information.However,in the imaging process,many factors will interfere and degenerate the quality of images.Therefore,restoring image to restore its original information is of great significance.The existing restoration algorithms are mostly based on degradation models.When the degradation information is unknown or the estimation is not accurate,the quality of image restoration is poor.So proposed a focal plane image restoration algorithm based on dual digital cameras.According to the phase difference theory,solved the wavefront disturbance information received in the target imaging process,and inverted the degradation model of the image to achieve the purpose of restoring the ideal image.The main contents of this article are as follows:(1)Researching on the phase difference theory,to determine the basic model of the image restoration algorithm.Use the improved BFGS optimization algorithm and Wolfe-Powell guideline for solving the model parameters.(2)Analyze and correct the errors of the restoration algorithm.For image mismatch problem,sub-pixel image registration is performed by combining phase correlation method,least square fitting and gradient method.For the defocus error,the defocus compensation coefficient is introduced to make the adaptive defocus error correction to the algorithm.For noise problem and the ill-posed problem of the algorithm,noise models with different illumination levels were introduced,and regularity methods were used for pathological suppression.(3)Through the restoration of the image,we analyze the effects of the different correct model and the value of defocus on the image,and compared with the classical algorithm,to test the feasibility and the advantage of the algorithm.The experimental data shows that the algorithm can solve the problem of blurred image and shadow image caused by image mismatch and other errors,and improve the accuracy of image restoration.When the focus range is 0.8~1.2 ?,the ideal focal plane image can be recovered.When the defocus distance is 1 ?,the image sharpness is the highest and the details are most obvious.Compared with the classical image restoration algorithm,this method has better effect on random noise suppression and does not need to degenerate prior information.It can overcome the interference of higher-order phase difference,which is superior to other image restoration algorithms.