Research On Hot Spot Detection And Location Technology Of Photovoltaic Power Plant Based On Infrared Thermal Imaging And Visible Light Image

In recent years,the photovoltaic power generation industry has developed rapidly.The hot spot effect not only seriously affects the power generation efficiency of photovoltaic power plants,but also seriously affects the safety of photovoltaic power plants.The traditional hot spot detection and positioning methods are difficult to meet the actual needs of large-scale photovoltaic power plants.Therefore,this paper selects the infrared thermal imaging image and visible light image of photovoltaic power station in real operation as the research object,and proposes a method for hot spot detection and localization of photovoltaic array based on infrared image and visible light image,which realizes automatic detection and localization of hot spot,details as follows:(1)A photovoltaic array detection method based on improved Adaboost algorithm is proposed.Based on the prior knowledge of the regular geometry of the PV array,the Adaboost algorithm with improved weight update method is used to train the cascaded classifier that can identify the PV array to realize the preliminary detection of the complete PV array,and then remove the background area based on this.The photovoltaic power generation component is detected by finding the connected area and the average length and height of the photovoltaic module according to the result of removing the background.(2)Two hot spot detection and discrimination methods based on the complete detection frame of the complete PV array and the detection frame of the PV module were proposed.In order to improve the accuracy of hot spot detection,the hot spot is firstly screened by using the highest temperature and average temperature of the detection area,and then the temperature data in the sliding window is used for secondary determination of the hot spot,and finally the suspicious hot spot obtained by the two detections is performed.The final result is obtained with the operation and position judgment.(3)A hot spot imaging method for PV array infrared thermal imaging images and visible light image registration based on slope and length constraints is proposed.Firstly,the improved Harris algorithm is used to extract the feature points in the two images,and then the feature points are coarsely matched based on the minimum Euclidean distance criterion.Finally,based on this,the slope and length are screened to obtain the final fine registration result,so as to achieve hot spot localization.The experimental results show that the proposed algorithm not only realizes the automatic detection and localization of hot spots of photovoltaic arrays,but also has high operating efficiency and accuracy,which meets the actual needs of large-scale photovoltaic power plants.