Research On The Method Of Forward Motion Blur Detection

Motion is one of the main reasons that lead to blurring of the image. During the exposure time of the camera, owing to the relative motion between the camera and the object, the forward motion blurred image is obtained. Because of this, the image restoration technology is gradually developed, and is widely used in all walks of life. The purpose of image restoration is to restore the degraded image to the original image as much as possible. However, the key to the restoration of forward motion blurred image is the determination of the point spread function, that is the direction of motion and the size of the fuzzy image. Therefore, it has an important practical significance to the accurate identification of motion blur image parameters, and it is also the main research content of this paper.In this paper, we analyze the current situation of the research on the identification of motion direction and the fuzzy scale. We improve the method of fuzzy angle detection based on spectrum and we propose an improved fuzzy scale detection algorithm based on two differential auto correlation method. The specific research contents are as follows:(1) For the detection of fuzzy direction, we improve the algorithm of fuzzy angle detection based on spectrum in this paper. Firstly, we take the modulus of the spectrum of the blurred image, and then take the logarithm operation. Secondly, the Log operator is used to detect the edge, and the two value image is obtained. Finally, through the Hough transform to extract a number of straight lines, and the angle of motion blur is obtained by calculating the average value of the angle of a number of lines. Through a lot of simulation experiments, it is proved that this method is of high precision and small error. By comparison with other methods, the results show that the improved algorithm has better robustness and stability.(2) For the detection of fuzzy scale, we propose a second differential auto correlation method which is based on the combination of column difference and sobel operator. First, we use the first order differential operators to carry out the differential of the fuzzy image. Then the sobel operator is used to detect the edge. After that, solve the auto-correlation of gradient image and on the basis of the average value of the results to draw the point spread function. By calculating the distance between the zero frequency and the negative spike, the size of the vague scale is obtained. This paper takes MATLAB as the experimental platform. The feasibility and adaptability of the algorithm are proved by several simulation experiments on different images. Compared with other detection algorithms, it is proved that this method has good stability, high precision and small errors.(3) A comprehensive experiment for the detection of forward motion blur parameters. Firstly, we seek angle for the blur image, then rotate the image and then seek the blur scale. Aiming at the problem of image rotation, a method for finding fuzzy scale by cyclic rotation is proposed in this paper. That is to rotate the image in the range of the detected angle plus-minus 1 and rotate once every 0.1 degrees and to detect its corresponding fuzzy scale at the same time. The experimental results show that this method can not only verify the accuracy of the angle detection, but also prove that the scale detection algorithm can detect the motion blur length in the range of positive and negative 0.5 degrees in the horizontal direction.