论文部分内容阅读
近年来随着自动目标识别(ATR)技术的快速发展,对于目标识别系统性能科学评估方法的研究越来越迫切。在总结已有方法的基础之上,针对基于模糊综合评判的评估方法中指标权重的难以合理量化,从而导致评估结果对权重数值敏感而不稳健这一瓶颈问题,引入了云理论中的正向云模型,实现了权重由定性到定量的不确定性转换。通过采用外场实测数据的仿真,验证了改进后的方法比传统的模糊综合评判方法具有更好的稳健性。
With the rapid development of automatic target recognition (ATR) technology in recent years, the research on the method of scientific evaluation of target recognition system performance is more and more urgent. On the basis of summarizing the existing methods, aiming at the bottleneck problem that the evaluation result is sensitive and not robust to the weight value, it is difficult to reasonably quantify the weight of the indicator in the evaluation method based on fuzzy comprehensive evaluation. The positive Cloud model, to achieve the weight from qualitative to quantitative uncertainty conversion. Through the simulation of field measured data, it is verified that the improved method has better robustness than the traditional fuzzy comprehensive evaluation method.