基于视觉信息的继电保护压板投退位置的自动识别研究

Image auto-recognition of relay pressed board in on/off states based on visual information

如何有效地进行压板投退管理对电力系统安全稳定运行具有重要作用,但是压板的投退状态的获取是个难点问题,一直没有好的解决办法。该文利用遥视系统上传的视频图像,从拍摄的保护屏的视频流中抽出特定的静态有效帧,利用计算机图像识别算法自动识别设定的有效区域内的压板的投退状态,并将压板状态信息作为继电保护管理系统输入信息进行管理。利用现场采集的压板图像信息对匹配与比较算法进行识别研究表明,所提算法具有较高的准确性,抗干扰能力和实时性也很好。

That how to manage the on/off states of relay pressed board is very important to the safety and stability of power systems, but it is difficult to get its on/off states, which has not been solved all through. With upsent images of remote vision system, a static frame is drawn out from video signal of recorded protection-screens, the auto-recognition of on/off states of relaying pressing board at the setting area is implemented by the arithmetic of computer image, and the recognition of on/off information is put to relay protection management system as the input data. The matching and comparing arithmetic research making use of local images indicates that this arithmetic has the capability of high veracity, anti-jamming and good real time.