CAJ | 학술논문

针对视频监控系统人工监视容易产生疏漏的问题,研究在铁路场景下基于智能视频监控图像序列处理和分析的识别物体入侵行为检测技术,给出铁路入侵物体检测、定位与跟踪的适用算法,实现基于轨迹点行为模型的人侵行为分析与理解.设计和开发了基于OpenCV的铁路入侵检测实验平台,在此平台上:实现背景建模、运动检测与跟踪等项功能,初步实现了对入侵行为的识别、分析以及对危险行为的报警.
침대시빈감공계통인공감시용역산생소루적문제,연구재철로장경하기우지능시빈감공도상서렬처리화분석적식별물체입침행위검측기술,급출철로입침물체검측、정위여근종적괄용산법,실현기우궤적점행위모형적인침행위분석여리해.설계화개발료기우OpenCV적철로입침검측실험평태,재차평태상:실현배경건모、운동검측여근종등항공능,초보실현료대입침행위적식별、분석이급대위험행위적보경.
Focusing on the problem easily leading to the oversights in the manual video surveillance system, this paper studies the invasion detection technology in the railway scene based on the image sequence processing and analysis, which gives the algorithms of the invasion object detecting, locating and tracking, and the approaches for behavior analysis and interpretation with the point trajectory behavior model. With the OpenCV tools, an experiment platform for the railway invasion detection is designed and developed. This platform has accomplished the background modeling, motion detecting and tracking etc.. And it can initially realize the identification and analysis of the invasion behavior, and the warning for the dangerous behavior.