基于全卷积网络的配电线下鱼塘垂钓识别方法

Fish pond fishing identification method under distribution line based on Full Convolution Network

随着电网建设规模的不断扩大,穿越鱼塘的输配电线路日益增多,为解决输配电线路下垂钓者发生触电风险隐患的问题,文中提出了一种基于全卷积网络的配电线下鱼塘垂钓识别方法。通过安装在输配电线路下方鱼塘旁的摄像头采集鱼塘视频图像,采用全卷积网络图像识别算法对现场的图片进行分析,判断是否有垂钓者及鱼竿物体,当垂钓者携带鱼竿进入鱼塘时,通过扬声器和警示灯进行告警。在某县鱼塘进行验证,其准确率为99.2%,应用结果验证了文中所提方法的有效性。

With the continuous expansion of power grid construction scale,there are more and more transmission and distribution lines passing through fish ponds.In order to solve the hidden danger of electric shock of fishermen in sagging transmission and distribution lines,a fishing identification method of fish ponds under distribution lines based on Full Convolution Network is proposed in this paper.The video images of fish ponds are collected by the camera installed next to the fish ponds under the transmission and distribution lines.The Full Convolution Network image recognition algorithm is used to analyze the pictures on the scene to judge whether there are anglers and fishing rod objects.When the angler enters the fish pond with the fishing rod,the alarm is given through the speaker and warning light.It is verified in a fish pond in a county,and the accuracy is 99.2%.The application results verify the effectiveness of the proposed method.