A Prevention and Control Method of Bird Harm Based on Bird's Nest Refined Detection and Bird Harm Level Assessment

In the current construction of smart grids, the problem of bird damage pose a serious threat to the safe operation of the transmission line. To this end, this article proposes a method of preventing and controlling bird harm based on the precision testing of the bird's nest and the rating of bird harm. First, the Mask2Former model is used for image segmentation. This model can accurately identify the bird's nest in the complex background based on the Transformer architecture. After that, the lightweight deep convolutional neural network was planned to extract the visual characteristics, and the two-way encoder representation method was used to analyze the text characteristics to study the deep fusion scheme of the image and text information. On this basis, this article is designed with the bird harm level evaluation model of time and space semantic information. This model intends to use the time and space diagram neural network to predict the trend of the shape of the bird's nest form, and combine the random forest algorithm to evaluate the bird's harm level. The experimental results show that the method proposed in this article has achieved remarkable results in the division of the bird's nest, with an average accuracy rate of 99.37%. Finally, this article designs the visual interactive interface of bird damage prevention and control, and verifies the effectiveness of the method of preventing the prevention and control of birds. Through this study, it provides strong technical support for the prevention and control of the transmission line, which helps ensure the stable operation of the power grid.