基于视频监控参考量的异常行为检测研究

为了解决传统算法难以检测一般动态场景情形下人体运动目标的问题，文中提出了一种新的人体运动异常行为的检测方法，该方法组合利用视频监控各个的参考量。文中针对视频序列中人的行为进行分析，目的是检测出人的异常行为，具体涉及：人体运动目标的检测、跟踪与提取，异常行为检测等。文中阐述了异常行为检测的相关概念，介绍了视频监控参考量各个参数的计算方法，探讨了异常行为检测与分类技术的关系。结合异常行为检测与分类的相似性，提出了基于视频监控参考量的算法的异常行为检测方法，给出了其计算方法，并确定了检测的过程，分析该方法的特点和优势。

Abnormal Behavior Detection Research Based on Reference Amount of Video Surveillance

In order to solve the traditional algorithm is difficult to detect a human moving target in the case of dynamic scenes , this paper proposes a new human motion abnormal behavior detection method. The method combines the use of video monitoring each of the reference amount , analyzes the video sequence into the behavior , for the purpose of the abnormal behavior to detect discrepancies , specific- ally related to ： human moving target detection, tracking and extraction, abnormal behavior detection. It elaborated the concept of abnor- mal behavior detection, and introduced the method of calculating the various parameters referenced to the amount of video surveillance. The algorithm combines the advantages of detection and classification of abnormal behavior and the similarity, and made reference to the amount of video-based surveillance of aberrant behavior detection algorithm, which is calculated is given, and to determine the detection process, analysising of the characteristics and advantages of this method.