Research On Evaluation Methods Of Short-wave Infrared And Medium-wave Infrared Image Fusion Systems

Image fusion technology has been widely used in the field of image processing for its complementary advantages to multi-source image information technology. For decades, researchers intensely studied the fusion algorithm, and developed various multi-Band fusion systems, among which shortwave and medium wave infrared image fusion system have outstanding performance in the field of military night vision and target detection. Evaluation is a very important link in that it can promote the development of the image fusion technology. But so far, there are not enough researches on fusion system evaluation methods to form a complete evaluation model. Current researches on image fusion evaluation mostly focus on the evaluation of the quality of fused images. This paper evaluates the shortwave and medium wave infrared image fusion system from two aspects:noise suppression and target detection probability.Firstly, according to the general process of fusion imaging, four major aspects that affect the image quality are discussed. Then, the spectral distribution of typical environments and objects are analyzed, and a modified matching factor based on the target structure is proposed.Secondly, common noises in the fusion system are analyzed, fusion noise gain factor is proposed to measure the noise suppression, and the fuse results of different noise input are analyzed.Thirdly, amplitude similarity and phase similarity between source image and fused image are considered. Based on that, an evaluation factor of fused image is proposed. Then, the existing target detection probability is amended by using this factor.Finally, a program to assess the image fusion system is designed using MATLAB simulation tool. The program includes the integration of image quality evaluation based on phase and amplitude and the integration of fusion algorithm evaluation based on noise suppression capability.