Laser interference image quality evaluation method

Abstract

The invention discloses a laser interference image quality evaluation method. The method comprises the steps of acquiring a laser interference image and a corresponding reference image; using the sameconvolution network to respectively carry out convolution pooling operation of multiple times of different level feature extraction on the laser interference image and the reference image in sequence; and calculating the similarity between the feature vectors obtained when the laser interference image and the reference image are subjected to the same convolution pooling, and carrying out weightedcalculation on all the similarities to obtain an image quality score. According to the method, the convolutional network is introduced into laser interference image quality evaluation, the distortiondegree of the interference image is measured by utilizing the similarity of the output features of the reference image and the interference image in each convolutional layer of the convolutional network, and the hierarchy of the features extracted by the convolutional network and the sensitivity to occlusion are fully utilized. Besides, the similarity values corresponding to all convolution pooling are subjected to weighted calculation, the final evaluation score conforms to the actual subjective perception of human eyes, the reliability is high, the positions of a target and a light spot donot need to be detected, and the application scene is wide.

Description

technical field [0001] The invention belongs to the field of image processing, and more particularly relates to a laser interference image quality evaluation method. Background technique [0002] Laser is used in the interference of photoelectric imaging system because of its monochromaticity, directivity, high brightness and other characteristics. In the process of photoelectric countermeasures, imaging devices such as optical CCD or CMOS are easily affected by laser interference, resulting in a significant decline in image quality. Detection and recognition performance of information processing systems. In-depth analysis of the laser interference effect of the image, especially the impact of laser interference on target detection performance, and a quantitative index system can provide technical support and theoretical basis for the research and development of laser interference systems. [0003] At present, the full-reference image quality evaluation algorithm is simple...

AI Technical Summary

Problems solved by technology

[0004] The invention provides a method for evaluating the quality of laser interference images, which is used to solve the technical problem that the existing laser interference image quality evaluation needs to know the position information of interference spots and targets in advance and has a narrow application range.

Images