基于卡尔曼滤波的高速运动模糊图像自适应复原方法OA

The noise interference of high-speed motion blur image is dynamic and changeable,which brings some challenges to accurate restoration.The adaptive Kalman filtering can estimate and correct the parameters of the Kalman filtering model according to the real-time change of the noise in the image,optimize the filtering design and reduce the filtering error,and ensure the accuracy of the Kalman filtering.Therefore,a high-speed motion blur image adaptive restoration method based on Kalman filtering is studied.A high-speed motion blur image degradation model is constructed,and initial filtering restoration processing is performed on the degradation model by basic Kalman filtering.The weighted coefficient is integrated into the basic Kalman filtering to obtain the adaptive Kalman filtering.According to the dynamic change of the noise characteristics of the high-speed motion blur image,the noise vector of the image is corrected in real time,and the state prediction value of the image is updated dynamically to obtain the restored image with high consistency with the real high-speed moving image.Taking speed skating as an example.The proposed method is used to restore the high-speed motion blur image of speed skating.The experiments show that the feature similarity and edge preservation effect of the restored speed skating image are good,the noise interference in the image is filtered out effectively,and the overall quality of the image is improved significantly.