Research on Automatic Landing and Positioning of Indoor Substation Inspection Unmanned Aerial Vehicle Based on Machine Vision Assisted Technology

If the indoor substation inspection drone deviates from the predetermined position during landing, it may collide with important and expensive equipment such as distribution cabinets and transformers in the substation room, causing damage and affecting the stability of power supply. To this end, a method for automatic landing and positioning of indoor substation inspection unmanned aerial vehicles based on machine vision assisted technology is studied. Capture and transmit landing information of indoor substation inspection unmanned aerial vehicles through multispectral cameras and self-organizing network communication technology; Gaussian filtering, wavelet decomposition, and bilateral filtering algorithms based on machine vision assisted technology are used to enhance drone landing images; Constructing a reference coordinate system for automatic landing and positioning of indoor substation inspection drones; The center position of the landing target is determined using a kernel correlation filter, and the landing image is segmented. The residual learning network is used to optimize the automatic landing positioning of the indoor substation inspection drone. The experimental results show that after enhancing the drone image, the brightness, contrast, and clarity of the image are significantly improved; The landing position coordinates obtained by this method are relatively close to the actual position coordinates, with small average center offset and average positioning error, which improves the accuracy of automatic landing positioning for indoor substation inspection drones.