自动目标识别系统的有效性评价方法

针对自动目标识别（automatic target recognition， ATR）评估中存在“性能－代价”在扩展工作条件中相互制约的情况，提出了一种新的ATR有效性评价方法。该方法从评估系统效率角度出发，提出并定义了ATR有效性，使得评估实践中无需对ATR系统的性能或代价进行严格限定。基于数据包络分析(data envelopment analysis, DEA)方法并结合ATR评估问题的实际特点，讨论了带有AHP约束锥的C2GS2模型用于ATR有效性评估的合理性及若干细节问题。最后，通过一个应用实例说明如何运用该评价方法解决实际的ATR评估问题。

Approach to effectiveness evaluation of automatic target recognition system

This paper investigates the automatic target recognition (ATR) evaluation problem,where the ATR performance and system cost are restricted by each other in the extended operation condition (EOC).Based on the concept of system efficiency,a new evaluation approach is proposed.This new approach gives the ATR efficiency definition,so there is no need to limit the performance or cost of ATR system rigidly in practice.According to the requirements of ATR evaluation,the paper discusses the rationality to use a C~2GS~2 model with AHP restraint cone and some related details.Finally,an application example is given to illustrate how to use this new approach to sovle some real ATR evaluation problems.