Orbital Area Obstacle Detection Based on GM-APD LiDAR
Author:
Affiliation:
1.44th Research Institute,China Electric Technology Group Corporation;2.National Key Laboratory of Laser Spatial Information, Harbin Institute of Technology
To address the problem of obstacle detection in complex rail environments, this paper proposes a detection method based on Geiger-mode avalanche photodiode (GM-APD) LiDAR. GM-APD LiDAR, with its high sensitivity and single-photon detection capability, provides technical support for detecting weak signals in complex scenarios. The study constructs a multi-level processing framework, filtering foreground noise through target distance estimation, reconstructing target depth and intensity information using a weighted first-order Gaussian matched filter algorithm, and achieving feature fusion classification and structural completion of tracks and obstacles by combining principal component analysis (PCA) with unidirectional template matching techniques. Experimental results show that the proposed method can effectively suppress interference from strong background light and weeds, improve image reconstruction quality in complex environments, and accurately separate low-reflectivity tracks from obstacles, providing a reliable solution for rail transit safety monitoring. This research will expand the application of GM-APD LiDAR in the field of rail transit and offer technical support for scenarios such as power inspection and intelligent operation and maintenance.