首页 | 本学科首页   官方微博 | 高级检索  
     

约翰逊准则的视频监控目标检测性能评估方法
引用本文:游瑞蓉,王新伟,任鹏道,何军,周燕.约翰逊准则的视频监控目标检测性能评估方法[J].红外与激光工程,2016,45(12):1217003-1217003(6).
作者姓名:游瑞蓉  王新伟  任鹏道  何军  周燕
作者单位:1.中国科学院大学,北京 100049;
基金项目:国家自然科学基金青年项目(61205019);国家自然科学基金(61475150)
摘    要:针对目前视频监控领域里目标发现、识别和辨认没有统一的国际标准和行业标准问题,对比分析了传统视频监控目标检测标准和红外热成像约翰逊准则,提出了基于约翰逊准则的视频监控目标检测性能评估方法,以等效条纹线对数代替传统模拟视频监控DRI标准中的目标-屏幕比,建立了CCTV Johnson-DRI标准。通过线对数绝对量化的评价标准更利于数字视频监控应用中目标发现、识别和辨认的性能评估。最后将提出的CCTV Johnson-DRI标准应用到激光夜视系统,进行了目标发现、识别和辨认实验,实验结果表明:所提出方法的有效性以及与传统视频监控DRI标准的等效性。该方法可用于白天彩色视频监控,LED/激光夜视等安防视频监控领域目标发现、识别和辨认相关性能的评估及系统指标设计。

关 键 词:视频监控    约翰逊准则    视频监控目标检测标准    目标发现    目标识别    目标辨认
收稿时间:2016-04-05

Target observation performance evaluation method for video surveillance based on Johnson criteria
Affiliation:1.University of Chinese Academy of Sciences,Beijing 100049,China;2.Optoelectronic System Laboratory,Institute of Semiconductors,Chinese Academy of Sciences,Beijing 100083,China
Abstract:Target detection, recognition and identification have no uniform international standards and industry standards in video surveillance currently. Comprehensively compared and analyzed the traditional CCTV DRI standards with Johnson criteria and proposed a target observation performance evaluation method with CCTV Johnson-DRI criteria for video surveillance. Applied equivalent number of line pairs instead of the target-screen ratio in traditional analog video surveillance DRI standards. The absolute quantitative evaluation criteria were more beneficial for evaluating the performance of digital video surveillance systems. Finally, the demonstrated DRI testing in laser night vision, and the experimental results show CCTV Johnson-DRI criteria were effective and equivalent to the traditional video surveillance DRI standards. It showed this method can be using for evaluating and designing system performance for target observation in daytime color video surveillance and LED/laser night vision systems.
Keywords:
点击此处可从《红外与激光工程》浏览原始摘要信息
点击此处可从《红外与激光工程》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号