SDM-YOLO: A Lightweight Detection and Ranging Model for Unmanned Docking of Airport Cargo Loaders

Addressing challenges such as poor environmental adaptability, low detection precision, and deployment difficulties encountered by airport cargo loaders during unmanned docking, this paper proposes SDM-YOLO, which is a lightweight, high-precision object detection and ranging model. The core innovations designed in this paper are as follows: (1) An automatic docking detection network model was designed. A dedicated attention mechanism is embedded within the network’s backbone to enhance the model’s feature extraction capability for critical regions. Dual-path feature interaction is added to the neck network to optimize the efficiency of multi-scale feature fusion. The detection head is decoupled into separate classification and regression task architectures, and an Efficient IoU is introduced to optimize for aspect ratio changes caused by distortion. A novel structured channel pruning method is devised to substantially reduce model size while maintaining precision. (2) Based on the above detection model, an integrated ranging module was designed and implemented. An innovative single-camera vision ranging method based on a standard projection plane was proposed. Based on the image information of the cargo door, an image correction module, a contour recognition module, and a distance measurement module were designed to achieve a real-time ranging function between the loader and the aircraft cargo door. Experimental results conducted on the apron demonstrate that compared to the baseline model, the SDM-YOLO achieves a 3.1% improvement in mAP, a 0.821% reduction in model parameters, and a 1.36-fold increase in inference speed. Compared to rangefinders, target recognition error decreased by 0.74%, while ranging error was reduced by 63.3%.

Authors

Institutions

Publication Details

Journal
Aerospace
Published
2026-11-16
DOI
https://doi.org/10.3390/aerospace13111071
Primary Topic
Advanced Neural Network Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

SDM-YOLO: A Lightweight Detection and Ranging Model for Unmanned Docking of Airport Cargo Loaders

Zhiqi Yan, Shuimiao Kang, Hengjia Zhu, Yongsheng Shi et al.
6 links
Aerospace
Advanced Neural Network Applications
article

SDM-YOLO: A Lightweight Detection and Ranging Model for Unmanned Docking of Airport Cargo Loaders

Zhiqi Yan, Shuimiao Kang, Hengjia Zhu, Yongsheng Shi, Xiangyu Suo
article en

Abstract

Addressing challenges such as poor environmental adaptability, low detection precision, and deployment difficulties encountered by airport cargo loaders during unmanned docking, this paper proposes SDM-YOLO, which is a lightweight, high-precision object detection and ranging model. The core innovations designed in this paper are as follows: (1) An automatic docking detection network model was designed. A dedicated attention mechanism is embedded within the network’s backbone to enhance the model’s feature extraction capability for critical regions. Dual-path feature interaction is added to the neck network to optimize the efficiency of multi-scale feature fusion. The detection head is decoupled into separate classification and regression task architectures, and an Efficient IoU is introduced to optimize for aspect ratio changes caused by distortion. A novel structured channel pruning method is devised to substantially reduce model size while maintaining precision. (2) Based on the above detection model, an integrated ranging module was designed and implemented. An innovative single-camera vision ranging method based on a standard projection plane was proposed. Based on the image information of the cargo door, an image correction module, a contour recognition module, and a distance measurement module were designed to achieve a real-time ranging function between the loader and the aircraft cargo door. Experimental results conducted on the apron demonstrate that compared to the baseline model, the SDM-YOLO achieves a 3.1% improvement in mAP, a 0.821% reduction in model parameters, and a 1.36-fold increase in inference speed. Compared to rangefinders, target recognition error decreased by 0.74%, while ranging error was reduced by 63.3%.

AerospaceVol. 13(11)
Civil Aviation University of China (CN)
Industry, innovation and infrastructure
Advanced Neural Network Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon