基于无线视频传感网的高压塔架监控系统研究

Research for Electronic Tower Supervising System Based on Wireless Video Sensor Networks

针对电力系统高压塔架监控的通信和视频检测两个难点,提出了一种以DSP为核心处理器,结合无线传感器网络的高压塔架监控系统。系统对高压塔架进行实时地监控,当检测到异物侵入监控区域时,通过无线传感器网络将现场图片传递给监控中心。系统采用无线网的方式,有效解决了有线网在监控范围、能源获取等方面的局限性。系统设计中,异物检测是系统实现的关键及难点,也是本文主要的研究内容。本文设计了一种快速检测异物侵入的算法,首先通过对称差分的背景更新算法得到背景,再利用当前帧与背景图像的差分得到二值图像,最后进行形态学腐蚀膨胀和阴影消除得到异物运动区域。对实际场景检测结果表明,该系统能在不同的光照条件下,快速建立和更新背景,并能实时准确地检测出侵入物体。

An electronic tower supervising system is proposed to deal with the communication and the video detection problems in the electronic tower supervising of electric power system. The proposed supervising system adopts the wireless sensor networks and uses DSP as the main processor. The system supervises the electronic tower in real-time fashion, when the invading body in the supervise area is detected, the scene image is then sent to the supervisor center through wireless sensor networks. As the system adopts the wireless networks that the limits of the wired networks such as supervising extension and energy sources are well resolved. How to detect the invading body is the key problem in the system design, and it is the primary content of this research. In this paper, a fast detecting algorithm is presented to detect the invading body. First, the background of the image is obtained by applying the symmetrical difference background updating algorithm. Then we make the difference for the present frame and the background to get the binary image. At last, the binary image is processed by morphologic erode and dilate arithmetic to get the moving area of invading body. The experiment of real scene detecting shows that the proposed system can quickly build and update the background in different lighting conditions and exactly detect the invading body in real-time fashion.