Abstract: In order to improve the visual detection ability of the conductor and ground wire sag of UHV transmission line, and realize the real-time measurement and alarm of the cross span distance of the conductor and ground wire, a method based on binocular vision technology for UHV transmission line sag detection is proposed. A visual imaging model is constructed, fusion filtering and information enhancement technology are used for image enhancement, artificial intelligence visual imaging features are analyzed, based on binocular vision technology, visual reconstruction through optimal control of image fusion is realized. Using fusion clustering analysis method, through ambiguity identification and parameter fusion method, combined with the scene reconstruction technology, the three-dimensional reconstruction of the ground wire sag of UHV transmission line is carried out, and AR technology is used to carry out the ground wire sag of UHV transmission line visual information processing, real-time measurement and optimization processing results of the crossing distance of the ground lead. The simulation results show that the method has better visual reconstruction ability, higher measurement accuracy and shorter detection time.