Obstacle detection and tracking for intelligent agricultural machinery

作者：Jiang, Wuhua; Chen, Wuwei^*; Song, Chuanzheng; Yan, Yajie; Zhang, Yuexin; Wang, Shicai

来源：COMPUTERS & ELECTRICAL ENGINEERING, 2023, 108: 108670.

DOI：10.1016/j.compeleceng.2023.108670

• Obstacle detection
• Tracking
• Dynamic clustering
• Association
• Kalman filter

The existing 3D lidar-based obstacle detection and tracking methods are inaccurate. Furthermore, as the number of obstacles increases, particularly in the agricultural machinery working scene with many machineries, ridges, and pedestrians, these methods become difficult to track. To address the aforementioned issues, a new obstacle detection and tracking method based on grid map is proposed. Firstly, a dynamic clustering algorithm that combines region-growing and density-clustering is developed. Secondly, geometric features and point cloud density are used to associates obstacles. Thirdly, a Kalman filter is designed to estimate the motion states of obstacles. Finally, a platform for agricultural machinery has been used to conduct comparative experiments. The proposed method improves detection accuracy by 4.86% and reduces average time by 47.8% when compared to Density Based Spatial Clustering of Applications with Noise (DBSCAN). And the tracking error of moving obstacles is less than 5%.