Study On Synthesis Of Fusion Validity Distribution Of Bimodal Infrared Images

The infrared polarization and infrared intensity images are respectively imaged by using the difference polarization and intensity information of the thermal infrared target,which have some complementary information on targets,such as different luminance,contour,and texture of the target,and have been widely used in target detection and tracking,object recognition,and man-made target discrimination.How to effectively and reasonably select the fusion algorithm to better synthesize the complementary information of the source images is of great significance for improving the detection and recognition capabilities of the bimodal electro-optical imaging and obtaining high-quality fused images in the actual environment.At present,the choice of the bimodal infrared image fusion algorithms is mostly by analyzing the fusion effect on the significant difference feature types in advance,and the final fusion algorithm is determined to perform the fusion processing.However,the change of the actual scene often leads to the occurrence of uncertainties in the amplitude variation of many types of significant differences at the same time,resulting in an uncertain change on the algorithm fusion effect.This makes it difficult to ensure multiple types of differences with better fusion effect,bringing about the fusion failure of the subjectively chosen fusion algorithm.Therefore,it is need to have quantitative analysis about the change of the fusion effect caused by the variation of the difference feature's amplitude,this paper deeply studies the fusion algorithm's validity changing in the fusion processing by the changing amplitude of the difference feature:Based on the analysis of the imaged characteristics of the bimodal images,the amplitudes of the main different types of features are extracted;the concept of the fusion validity of the different feature amplitude is proposed,and the validity of fusion algorithm can be calculated at each different feature amplitude to construct the fusion validity distribution.By using the possibility distribution synthesis theory,different rules are used to synthesize the distributions to get the results of the fusion effectiveness of multi-difference feature amplitudes.In order to verify the efficiency of the fusion validity distribution and distribution synthesis results,this paper first verifies that the fusion validity distribution can truly reflect the variation of the fusion effect with the amplitude.Secondly,according to the amplitudes of multiple sets of bimodal infrared images as the cut-set level,it is verified from the subjective and objective experiments that the algorithm chosen is the fusion algorithm with the best fusion effect.Finally,a fusion method based on synthetic distribution is proposed.The correlation experiment proves that the proposed method has a good fusion effect.The side illustrates the effectiveness of the distribution of the fusion efficiency of the structure and distribution synthesis results.