Effective Pretreatment in Identification of Motion-Blur Direction

A motion-blur image has special blurred information which can be dealt to estimate its blurred direction. There are several methods to identify the direction such as based on directional derivation and weighted average [1]. The blurred image can´t recognized smoothly and timely in the area of industrial machine vision. This paper provides an effective pretreatment method to indentify the motion-blur direction related to the high frequency of the image. This method satisfies the high-speed requirements in industrial machine vision. Find the interval of the direction, then used the optimized method based on direction of differential coefficient to find the direction of the motion-blur image. The merit of this method is increased the convergence speed and save the time complexity.