| Thermal power generation is the main way of producing electric energy in my country.The flue gas containing sulfur nitrides produced in the process of thermal power generation will corrode the inner wall of the chimney.At present,the detection of the corrosion of the inner wall of the chimney mainly adopts the method of manual sampling and detection,and there is no mature solution for non-contact detection of the inner wall of the chimney.In view of the above problems,this paper proposes a UAV chimney hovering system,which lays the foundation for the subsequent implementation of the overall UAV chimney inner wall detection system.The main contents of this article are as follows:(1)A UAV chimney hovering system is designed.The main process of the UAV chimney hovering system is that the UAV detects the position of the chimney in the image through the image,and converts the pixel deviation into the actual position deviation.The actual position deviation is then sent to the drone,so that the drone finally hovers in the center of the chimney.(2)A chimney image detection algorithm based on YOLO and traditional image processing is proposed.A chimney dataset is established,and different network models of YOLO are trained.By reducing the width of the YOLOv5 n network,a fast chimney target detection algorithm is obtained.The chimney detection algorithm combines YOLO and traditional image processing.First,the trained YOLO model is used to detect the position range of the chimney in the image,and then traditional image processing is performed on the image of the position range to obtain the precise position of the chimney in the image.(3)Simulate the UAV flight control system,and realize the UAV speed control based on ADRC.The vision-based UAV position control is realized through the conversion of image pixel deviation and actual distance deviation.The overall drone chimney hovering system was tested to confirm the feasibility and reliability of the overall solution. |
PDF Full Text Request