Roadside LiDAR Vehicle Detection and Tracking Using Range and Intensity Background Subtraction

Abstract

In this study, we developed the solution of roadside LiDAR object detection using a combination of two unsupervised learning algorithms. The 3D point clouds are firstly converted into spherical coordinates and filled into the elevation-azimuth matrix using a hash function. After that, the raw LiDAR data were rearranged into the new data structure to store the information of range, azimuth, and intensity. Then, the dynamic mode decomposition method is applied to decompose the LiDAR data into low-rank backgrounds and sparse foregrounds based on intensity channel pattern recognition. The coarse-fine triangle algorithm (CFTA) automatically finds the dividing value to separate the moving targets from static background according to range information. After intensity and range background subtraction, the foreground moving objects will be detected using a density-based detector and encoded into the state-space model for tracking. The output of the proposed solution includes vehicle trajectories that can enable many mobility and safety applications. The method was validated at both path and point levels and outperformed the state of the art. In contrast to the previous methods that process directly on the scattered and discrete point clouds, the dynamic classification method can establish the less sophisticated linear relationship of the 3D measurement data, which captures the spatial-temporal structure that we often desire.

Semi-Automatic Annotation of 3D Radar and Camera for Smart Infrastructure-Based Perception

“…To the best of the author's knowledge, no work has been found regarding 3D radar point cloud-based background subtraction in the literature, and only [22] has described it, but for roadside 3D lidar sensor. The work proposed here is partially inspired by [22] to develop a suitable algorithm for 3D radarbased background subtraction.…”

“…Hence, a 3D radar background subtraction method is proposed in this work that filters out background clutter in a static setup to a large extent. This method is inspired by the roadside 3D lidar-based background subtraction technique described in [22]. These two solutions are also part of the proposed semi-automatic annotation methodology.…”

InfraDet3D: Multi-Modal 3D Object Detection based on Roadside Infrastructure Camera and LiDAR Sensors

“…This first step removes 69.9 % of points on average. Second, the detector finds points belonging to the ground by considering the Euclidean distance to a predefined plane model together with a threshold of 0.2 m. Third, the detector filters background artifacts within the region of interest based on the coarse-fine triangle algorithm [26]. The fourth step is radius outlier removal (n = 15, r = 0.8), which refines the extraction of the foreground point cloud.…”

Vehicle Speed Detection using Haar Cascade Classifier and Correlation Tracking

“…Researchers in the field of computer vision and image processing have proposed numerous approaches that contribute to the task of speed detection of vehicles. (3) proposed a solution for roadside LIDAR (Light Detection and Ranging) object detection using Range and Intensity Background Subtraction. The method retrieves a mobile object from a definite image and the retrieved object is resulted as the threshold of image differencing.…”