The automatic docking of airport special vehicles is an inevitable requirement for the development of smart airports in the future;It is the key step for automatic docking is to accurately identify and locate aircraft door.Aiming at this problem,an air-craft door recognition and positioning method based on improved YOLOv5 and monocular vision is proposed.By adding a lightweight convolutional block attention module(CBAM)in the model,the algorithm improves its ability to extract features from aircraft doors;A spatial pyramid pooling cross stage partial connection(SPPCSPC)is introduced to solve the problem of repetitive feature extraction in YOLOv5,and improving the number of group convolution groups to 4 and the detection accuracy of the algorithm;The pixels of corner points in the candidate frame are obtained,and the spatial geometric relationships are utilized to achieve the accurate three-di-mensional positioning of the aircraft door.The experimental results show that the mean average precision(mAP)of improved YOLOv5 algorithm reaches 96.5％,the mAP of improved algorithm is 5.6％higher than that of the original algorithm.The real-time maximum positioning errors of 19 m and 1 m in front of the aircraft door are 0.15 m and 0.01 m,respectively,which can meet the re-quirements of maintaining a safe distance of 5～10 cm from the aircraft door after the docking of special vehicles.
identification and positioning of aircraft doorairport special vehiclesautomatic docking of airportYOLOv5 algo-rithmthree-dimensional positioning
维普
万方数据
