头戴式摄像机变电站防误操作图像识别

摘要

为了满足变电站智能防误的需求，提出了一种基于图像识别方法的变电站防误操作智能识别系统。利用头戴式摄像机获取变电站现场的实时视频图像，基于尺度不变特征变换对标志牌以及操作目标进行智能识别，实现对操作人员定位，防止走错间隔。采用隐马尔科夫模型对采集到的操作人员手部轨迹进行匹配识别，结合操作票判断操作人员是否有误操作行为发生。本文详细介绍该系统的组成和实现方法。通过变电站内的现场测试验证了所提出方法的有效性。%A substation anti-mal-operation intelligent identification system based on image recognition method is proposed in order to meet the needs of substation intelligent anti-mal-operation. The real-time video image of substation site is captured by head-camera, the scale invariant feature transform is used to recognize the signboard image and the operation target, and it can determine whether the working staff has accessed to the proper bay. The movement trajectory of the hand is matched and recognized by building the Hidden Markov Model, and it can distinguish whether there exist anti-mal-operation by referring to the operation order. The composition and realization method of the system is introduced in this paper. The results by the site tests in substation were presen￣ted to verify the effectiveness of the proposed method.