A New Blind Evaluation Method of Infrared Image Quality
Wenjun Lu · 2021 IEEE International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI) · 2021
Blind evaluation of digital image quality has always been one of the most challenging problems in artificial intelligence and computer vision. Because of the limitation of imaging mechanism, the overall quality of infrared reconnaissance image is lower than that of visible image. It will seriously restrict the efficiency of subsequent image processing, seriously affect the performance of detection efficiency. The contours, textures and point features that can represent the characteristics of infrared radiation are analyzed and extracted, and a simple infrared image quality evaluation model combining mixed features and target saliency is constructed. Taking real infrared images as samples and subjective evaluation as the benchmark, the relative comparative verification is carried out. The results show that this method can achieve better evaluation performance than state-of-the-art methods.