KEY TECHNOLOGIES FOR DENSE MUSHROOM GROUP PICKING BASED ON IMPROVED DBSCAN CLUSTERING ALGORITHM

• Published in: Scientific Bulletin. Series C, Electrical Engineering and Computer Science no. 3; p. 147
• Main Authors: Li, Yafei, Zhu, Xuanzhang
• Format: Journal Article
• Language: English
• Published: Bucharest University Polytechnica of Bucharest 01.01.2025
• Subjects: Algorithms; Clustering; Harvesting; Mushrooms; Picking; Run time (computers)
• ISSN: 2286-3540

Traditional mushroom harvesting techniques are inadequate to meet the increasing demands of modern agriculture. This study proposes a dense mushroom cluster harvesting planning technique based on an improved Density-Based Spatial Clustering of Applications with Noise (DBSCAN). The proposed method combines clustering and harvesting planning by optimizing the DBSCAN algorithm. Key metrics such as clustering accuracy, intra-cluster point omission probability, and running time were analyzed and compared to existing algorithms. The optimized DBSCAN showed superior performance with clustering accuracy of 94.6%, intra-cluster omission probability of 2.5%, and a reduced running time of 0.25s. The system achieved a recognition accuracy of 93.8% and 95.8% for mushroom clusters, with picking success rates of 93.2% and 94.7%, respectively. This study introduces a novel harvesting planning approach that improves the efficiency and success rate of mushroom harvesting, reducing damage and meeting the required harvesting efficiency.