基于机器视觉的变电站一体化安防管控系统设计

Design of substation integrated security control system based on machine vision

摘要:

为解决变电站安防设备简单、智能化程度低问题，设计出基于机器视觉的一体化安防管控系统，安装现场视觉控制器作为系统硬件前端信息采集中心，利用以太网实现各设备通信连接一体化，门电路和变电站周界环境终端电子相连接，检测激光电子围栏危险入侵检测，序列巡检围栏编号，确定危险入侵围栏位置，通过线性二次型调节器算法对机器人期望位姿和实际位姿之间的耦合关系实行解耦处理，将处理结果作为神经网络的输入，输出机器人移动轨迹跟踪控制结果，完成视觉识别定位，实现动态物体监测、火苗、温度等多种危险因素的自动识别和预警。经过实验分析证明：系统通过机器视觉准确监测到了高温设备的位置和温度，并且响应性能良好，安防信号传输环境稳定，危险情况预判能力及预警能力较强，鲁棒性好。

Abstract:

In order to solve the problem of simple and low intelligence of substation security equipment, an integrated security control system based on machine vision is designed. The field visual controller is installed as the front end information acquisition center of the system hardware. The Ethernet is used to realize the integration of communication and connection of all devices, and the gate circuit is connected with the terminal electronics of the substation perimeter environment. The laser electronic fence is detected for dangerous intrusion detection, and the fence number is inspected in sequence to determine the location of the dangerous intrusion fence. Through linear quadratic regulator algorithm, the coupling relationship between the desired pose and the actual pose of the robot was decouped, and the processing results were taken as the input of the neural network to output the tracking control results of the robot's moving trajectory, complete the visual identification and positioning, and realize the automatic identification and early warning of dynamic object monitoring, fire, temperature and other dangerous factors. The experimental analysis shows that the system can accurately monitor the position and temperature of high-temperature equipment through machine vision, and has good response performance, stable security signal transmission environment, strong ability to predict and warn dangerous situations, and good robustness.