In this paper, a video edge detection algorithm based on AI technology is presented to improve the accuracy of power operation safety monitoring. Firstly, the block and low rank tensor recovery video denoising algorithm is used to filter out internal noise in videos, which improves video clarity and makes video edge information more prominent. Secondly, the proximal support vector machine (PSVM) is used to detect the edges of denoised videos, thereby improving the effectiveness of video edge detection. The experimental results show that the proposed algorithm outperforms other algorithms in video edge detection, further ensuring the safety of power operations.