DOI：10.3969/j.issn.1672-3872.2025.13.003
基金项目：大学生创新创业训练计划项目“基于APF-FMT算法的叶菜类蔬菜‘种养采’一体机器设计”（202410566013）；广东海洋大学深蓝智能机电产品创新团队（CXTD2023009）
作 者：黎科亮 1，3 ，余 江 1，3 ，陈瀚琳 2，3 ，刘术杰 1，3 ，林钰贵 2，3 ，刘盈锋 1，3
（1. 广东海洋大学机械工程学院，广东 湛江 524088；2. 广东海洋大学电子与信息工程学院，广东 湛江 524088；3. 广东海洋大学深蓝智能机电产品创新团队，广东 湛江 524088）
摘 要：【目的】实现叶菜类蔬菜的全程机械化作业，提高蔬菜生产机械化水平，使我国叶菜类蔬菜的机械化种植达到既省工又省力的效果。【方法】基于YOLOv5s算法设计了一种适用于多种叶菜类蔬菜（如生菜、菠菜等）收获的多功能综合收获机。 整机采用履带底盘，以蓄电池作为直流电源，集成螺旋喂入式采收机构、等间距有序播种浇水机构、转盘式全幅施肥机构和起垄机构等。【结果】1）该机器能够实现从收获到再次种植的全程自动化作业。2）机器识别系统平均精确率（Precision）达到85.3%，平均召回率（Recall）达到93.1%，综合评估指标平均F1分数（F1-Score）达到0.892。显著优于传统图像处理方法约80%的精确率水平，性能提升明显。【结论】此设计有效提升了叶菜从播种到收割全过程的作业效率，增强了作业的连贯性和协调性。相较于传统的人工采收方式，可以节约大量成本、提高经济效益、降低农民的劳动强度，有助于推动农业机械化水平的提升，促进农业现代化发展。
关键词：叶菜类蔬菜；综合收获机；YOLOv5s；农业现代化；自动化
Multifunctional Leafy Vegetables Integrated Harvester Based on YOLOv5s Algorithm
Author: Li Keliang1,3, Yu Jiang1,3, Chen Hanlin2,3, Liu Shujie1,3, Lin Yugui2,3, Liu Yingfeng1,3
(1.School of Mechanical Engineering, Guangdong Ocean University, Zhanjiang 524088, China; 2.School of Electronic and Information Engineering, Guangdong Ocean University, Zhanjiang 524088, China; 3.Deep Blue Intelligent Electromechanical Product Innovation Team, Guangdong Ocean University, Zhanjiang 524088, China)
Abstract: [Objective] To achieve full mechanization of leafy vegetables, improve the mechanization level of vegetable production, and achieve a labor-saving and labor-saving effect in the mechanized planting of leafy vegetables in China. [Method] A multifunctional integrated harvester suitable for harvesting various leafy vegetables (such as lettuce, spinach, etc.) was designed based on the YOLOv5s algorithm. The whole machine adopts a tracked chassis, with a battery as the DC power source, integrated with a spiral feeding harvesting mechanism, equidistant orderly sowing and watering mechanism, rotary full width fertilization mechanism, and ridge forming mechanism. [Result] 1) The machine is capable of achieving fully automated operations from harvesting to replanting. 2) The average precision of the machine recognition system reaches 85.3%, the average recall rate reaches 93.1%, and the average F1-score of the comprehensive evaluation index reaches 0.892. Significantly better than traditional image processing methods with an accuracy level of about 80%, with a significant improvement in performance. [Conclusion] This design effectively improves the operational efficiency of leafy vegetables from sowing to harvesting, enhancing the coherence and coordination of operations. Compared with traditional manual harvesting methods, it can save a lot of costs, improve economic efficiency, reduce farmers’ labor intensity, and help promote the improvement of agricultural mechanization level and the development of agricultural modernization.
Keywords: leafy vegetables; integrated harvester; YOLOv5s; agricultural modernization; automation
引文信息：[1]黎科亮，余江，陈瀚琳，等.基于YOLOv5s算法的多功能叶菜类蔬菜综合收获机[J].南方农机，2025，56(13)：14-17+31.
查看全文请下载PDF文件↓
