Abstract: The “hot spot” phenomenon of photovoltaic modules is an important factor affecting the photovoltaic power generation system. In order to solve the problems of high labor intensity and low efficiency of traditional manual inspection, this paper proposes an intelligent inspection method of photovoltaic power station hot spot based on Unmanned Aerial Vehicle (UAV) infrared vision, which effectively improves the intelligent level of photovoltaic power station health status assessment. Firstly, the state characteristics of photovoltaic hot spot were analyzed, and the coupling relationship model of hot spot infrared image, temperature and severity was established. Then, in order to improve the accuracy of hot spot detection, a target detection method based on hybrid intensive (MI-YOLO) is proposed to enhance the expression ability of shallow feature information. Then, an intelligent inspection scheme of UAV was designed to obtain the multidimensional information of hot spots in real time. Finally, taking recognition speed, recall rate and accuracy rate as evaluation indicators, multiple sets of simulation experiments were carried out to compare the influence of different activation functions, image size, image brightness and false detection factors of "fake" targets. Experimental results show that MI-YOLO can effectively detect the location, quantity, category, and severity of PV module hot spots, which can meet the real-time requirements of power plant operation and maintenance, improve power generation efficiency, and reduce operation and maintenance costs.