Research On Synthesizing Effectiveness Distribution Of Different Features And Multi-attributes Of Bimodal Infrared Image

The infrared light intensity image and the infrared polarization image are imaged based on the radiation intensity information and polarization information of the infrared target,respectively.These two types of images have many differences in characteristics such as brightness,edges,and texture details,so they have a lot of complementary information.The same fusion algorithm has different fusion capabilities for different difference features.Therefore,studying the dyna-mic changes of multiple attribute types,amplitude,frequency,and other attributes of dual-modal infrared images and the correlation between feature attributes a-re the key to difference-driven fusion.There is a problem that multiple algorithms in the dual-mode infrared image difference-drive-n fusion conflict with each other,the fusion effect is poor,or even fails.Finally,a coordination mechanism between the multi-attributes of different features and the effective driving relationship between the fusion algorithms is established.Based on this,this paper conducts research on the fusion effectiveness distribution synthesis of the difference features and multiple attributes of dual-modal infrared images,analyzes the deep connection between the difference features and multiple attributes and the fusion algorithm and rules,and realizes the difference features through the method of possibility distribution synthesis.The purpose of attribute adaptive selection of fusion algorithm or fusion rule provides new ideas for the realization of difference feature driven fusion.The main research contents of this article are as follows:(1)Multi-attribute characterization of differential features: Explore the imaging characteristics of dual-modal infrared images and the types of differential features,and clarify the characteristics of different attributes of differential features.Analyze the relationship between the amplitude of the difference feature and the two attributes of frequency.Among them,the non-parametric estimation meth-od is used to construct the probability distribution of the difference feature frequency.The optimal frequency is selected by comparing the similarity measures between the different frequency distribution structure and the original cumulative distribution function The attribute construction method lays the foundation for the realization of the distribution synthesis of the validity of multi-attribute fusion of difference features.(2)Distribution structure of multi-attribute fusion effectiveness of different features: Explain the characteristics of the fusion algorithm and fusion rules,use the distance measure to characterize the fusion effectiveness of different features,use the stability index of the fusion effectiveness to evaluate the distribution of the fusion effectiveness of different features,and use the differences Feature types and amplitude attributes drive the selection of the optimal fusion algorithm for fusion,to achieve the distribution structure of the difference feature amplitude fusion efficiency and the difference feature frequency fusion effectiveness distribution structure.Finally,according to the difference feature weight function,the distribution of the difference feature multiple attribute fusion effectiveness distribution is constructed..(3)The establishment of validity distribution synthesis of multi-attribute fusion of difference features: clarify the meaning of possibility distribution synthesis and explore the applicability of the possibility distribution synthesis in multi-attribute fusion of difference features.On this basis,a fusion efficiency distribution-n synthesis of similar attributes with multiple attributes is established,and the mapping results of the fusion validity distribution synthesis of the amplitude attributes and frequency attributes of the similar characteristics with differences are explored.Establish the fusion efficiency distribution synthesis of heterogeneous difference feature multi-attribute,and propose algorithm-driven,rule-driven distribution synthesis of fusion feature weight difference function weighting function.The optimal fusion strategy is selected by the heterogeneous difference feature w-eight function-driven fusion to evaluate the optimal The fusion strategy optimize-s the feasibility of the combination method to meet the complex and changing needs of multiple attributes and different attributes,and provides a new method for the implementation of the optimal fusion strategy driven by the different attributes and multiple attributes.