无人机智能巡检技术在电力设备异常状态预测中的应用OA

This article delves into the application of UAV inspection technology and image processing technology in the field of power equipment,and mainly discusses the application of UAV inspection technology in visible light wavelength image recogni-tion.A region generation network-convolutional neural network(RPN-CNN)model based on image segmentation pre-weigh-ting,namely the weighted RPN-CNN model,is proposed,and the process that CNN extracts image features through convolu-tional layers,and the optimization process of the RPN-CNN model in power equipment detection are illustrated.Performance test results indicate that the weighted RPN-CNN model performs optimally in terms of loss rate,demonstrating the significant impact of image segmentation technology on improving the accuracy of abnormal state prediction.Among different initialization methods,the model using pre-training with weighted initialization shows the lowest loss rate.With an increase in the number of training samples,the adaptability and generalization capability of the weighted RPN-CNN model improve,making it suitable for power equipment inspection tasks of varying scales and complexities.The weighted RPN-CNN model achieves a prediction accuracy of 95.8％,showcasing its significant advantages in detecting abnormal states in power equipment.