Research On Anti-External Force Destruction Technology Of Transmission Line Based On Image Analysis

With the rapid development of China’s modernization construction,the safe operation of electric power equipment plays an inestimable role in the development of national economy and social stability at all times.The traditional transmission line security protection work is mainly carried out by manual means,but there are some safety problems in this way,for instance,it is difficult to find safety hazards in time and there is no way to monitor various areas around the clock.This thesis identifies the type of external force damage in the transmission line monitoring image by using deep learning technology,and predicts the possibility of external force damage on the transmission line,thereby avoiding the occurrence of safety accidents,ensuring the safe operation of the power grid and personal safety of the people.Based on the Faster RCNN deep learning network,this thesis identifies and predicts the transmission line monitoring image.The traditional Faster RCNN network is optimized which makes images recognized faster and more accurately,and realizes the automation of the external line damage type detection of the transmission line.Firstly,according to the migration learning principle,the feature extraction part is optimized by combining the pre-training data with the transmission line image data.The convolutional layer parameters in the convolutional network remain unchanged during the training process,and the data is fine-tuned to achieve a better fit between the model and the data in this paper.Finally the cross-validation method is used to verify the network output,and the accuracy of 92.3% is obtained on the data set,which proves that the improved network model has higher accuracy in processing the transmission line monitoring images that is more complicated and quality is general.Based on the previous network model design,an intelligent early warning system for transmission line monitoring image is realized,which is tested to have the function of real-time query and early warning processing on the transmission line monitoring image.