Algorithmic Research of Image Blur Detection

As a powerful information transfer tool, image is widely used in various fields, however, the image acquired in the process of collecting images can be blurred due to various reasons, which reduces the effectiveness of information transfer. To address this problem, an image blur detection method based on Laplace operator re-blur theory is proposed. The method utilizes Gaussian blur to re-blur the original image, and then uses the Laplace operator to perform edge detection on the original image and the re-blur image, and defines the computed variance of the Laplace operator as the blur metric of the image, which the ratio is used as an objective score of the original image blur to determine the degree of image blur. The experiments are based on the open source dataset developed by the Laboratory for Image and Video Engineering at the University of Texas at Dallas, abbreviated as LIVE dataset and the effectiveness of the algorithm for image blur detection is analyzed using the Spearman’s Rank Order Correlation Coefficient (SROCC) and Pearson’s Linear Correlation Coefficient (PLCC) evaluation functions. The experimental results show that the method is effective in detecting the blur of the image.