摘要
针对全社会用电量需求的增大所引起的电力输送和用电安全与人力维护的成本较高等问题,借助图像增强技术与深度学习技术实现对电力设备图像的识别与分析。首先,利用图像增强技术完成对电力设备图像集的扩充和增强;其次,搭建MobileNet卷积神经网络完成对电力设备图像进行数据训练与识别。最终实验表明,MobileNet的运行速度为0.02 s一张图片,检测率可达到96%,该方法适用于图像特征较明显、大量重复检测的电力设备检测场景。
In view of the increasing demand for power consumption in the whole society,the high cost of power transmission,power safety and human maintenance caused by it,this paper uses image enhancement technology and deep learning technology to realize the recognition and analysis of power equipment images.Firstly,image enhancement technology is used to expand and enhance the power equipment image set.Secondly,a MobileNet convolutional neural network is built to complete data training and recognition of power equipment images.The final experiment shows that MobileNet runs at a speed of 0.02 seconds per image,and the detection rate can reach 96%.This method is suitable for the detection scene of electrical equipment with obvious image features and a large number of repeated detections.
出处
《工业控制计算机》
2024年第3期157-158,166,共3页
Industrial Control Computer
