基于知识推理的机场目标识别  

Airport target recognition based on knowledge inference

作　　者：陈韶斌[1] 蔡超[1] 丁明跃[1] 周成平[1] 

机构地区：[1]华中科技大学图像识别与人工智能研究所多谱信息处理技术国家级重点实验室,湖北武汉430074

出　　处：《红外与激光工程》2011年第3期548-552,共5页Infrared and Laser Engineering

基　　金：国家863计划探索类项目(2007AA12Z166)

摘　　要：提出了一种新的基于知识的机场目标的识别方法。首先建立机场目标知识库,然后运用知识推理,通过相关知识确定机场各组成部分,最后通过验证各部分组成机场的各种可能性,确定是否存在机场,从而完成对机场目标的识别。实验结果表明:该方法在识别机场的同时,还确定了机场各组成部分的位置,使得机场识别结果更加可靠,验证了该方法的可行性。同时,该方法还具有较大的灵活性,通过更新知识库的知识以及采取更多合适的图像特征,也适用于其他类型的目标识别。A novel knowledge inference based method for airport target recognition was proposed.This method was composed of the following steps： first,a knowledge base of airport targets was set up;then,knowledge inferring was performed and airport components were ascertained according to relevant knowledge of the components;finally,the existence of airport was determined by verifying all possible combinations of airport components and thus the airport target was recognized.Experimental results show that the method not only can recognize the airport target,but also determine the locations of all components of the airport target,which remarkably increase the reliability of the airport target recognition.It verifies the feasibility of the proposed method.Furthermore,the method is rather flexible in the sense that it can also be conveniently applied to the recognition of other kinds of targets by updating knowledge in the base and using more appropriate image features.

关 键 词：机场目标识别 基于知识的系统 基于部件的推理 C语言集成产生式系统(CLIPS) 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]