Image recognition technology is widely used in power systems.In order to solve the problems of helmet wearing mo-nitoring and pedestrian intrusion warning in power construction scenes,this paper proposes a target detection technology based on improved self-attention mechanism and image recognition technology.The paper proposes a channel self-attention mecha-nism,and realizes effective monitoring of power construction scenes in complex environments through feature extraction of multi-scale attention.Early warning can effectively ensure the safety of the power construction process.Finally,a series of comparative experiments are carried out to verify the method.The experimental results show that its recognition accuracy rea-ches 93.3％,which is at least 1.7％ higher than that of the comparative methods,which fully proves that the real-time monito-ring method of power scene based on the improved self-attention mechanism is effective.
关键词
目标检测/注意力机制/卷积神经网络/深度学习/电力场景监测
Key words
object detection/attention mechanism/convolutional neural network/deep learning/power scene monitoring
维普
万方数据