一种新的多指标融合适配性分析方法

摘要

When the sensed images are noisy,the matching probability of areas selected by available matching suitability indexes may change.In order to solve this problem,two novel indexes,phase congruen-cy correlation length and efficient contour density,are introduced at first.A matching suitability analysis method is proposed based on fusion of multi-indexes by evidence theory.The single index is not sufficient to characterize information content,stability and repetition rate simultaneously.So evidence theory is used to combine the basic belief assignments constructed by the indexes.The matching suitability area is determined by the combination results.Simulation results show that the base image and sensed image can still be well matched in the matching area obtained by the proposed method,even if the sensed images are quite noisy.%景象匹配适配性分析中，针对现有适配性指标较少考虑成像条件不同、实测图受干扰等导致基准图适配概率发生变化的问题，改进并提出了相位相关长度和有效轮廓密度两个新指标。然后针对单一指标难以同时有效兼顾信息量、稳定性以及重复率等适配标准的问题，给出不同指标置信指派构造方法，采用证据推理对多指标进行融合，提出一种基于证据推理的多指标融合适配性分析方法。仿真实验验证了在采用该方法分析得到的适配区域上进行景象匹配，当基准图和实测图差异较大时，正确匹配概率仍然较高。