| 注册
首页|期刊导航|农机化研究|基于机器学习的叶菜收获机割台仿形控制系统研究

基于机器学习的叶菜收获机割台仿形控制系统研究

杨云聪 姚立健 刘宇 王守卫 姚黎文

农机化研究2025,Vol.47Issue(11):28-37,10.
农机化研究2025,Vol.47Issue(11):28-37,10.DOI:10.13427/j.issn.1003-188X.2025.11.004

基于机器学习的叶菜收获机割台仿形控制系统研究

Contour Control System of Leaf Vegetable Harvester Based on Machine Learning

杨云聪 1姚立健 1刘宇 1王守卫 1姚黎文1

作者信息

  • 1. 浙江农林大学 光机电工程学院,杭州 311300
  • 折叠

摘要

Abstract

A machine learning based cutting table profiling control system was designed to address the issues of contact sensors being easily affected by contact forces and incomplete ridge surface information detected in the profiling leafy veg-etable harvester.Firstly,preprocess the data from the 2D LiDAR sensor to obtain the ranging information of the leafy vegetable canopy surface in the area to be harvested.Then,a point cloud acquisition method was proposed that integrates 2D LiDAR ranging information and inertial measurement unit perception velocity information.Further proposed a machine learning model for planar fitting of point cloud data,which took into account the impact of longitudinal and transverse un-dulations on the system.Finally,the control system adjusted the extension and contraction of the push rods on both sides based on the cutting plane fitted during each sampling period to complete the profiling.The results of the inter ridge har-vesting experiment showed that after installing the cutting table imitation control system,the cutting table changed with the fluctuation of the ridge surface,resulting in a decrease in crushing rate and leakage rate.Compared with the manual controlled harvesting experiment,the average crushing rate decreased by 6 percentage points,the average leakage rate decreased by 9 percentage points,and the average crushing rate and leakage rate were 23%and 22%,respectively.The proposed method for determining the cutting plane can effectively fit the plane,run quickly,and had strong robustness.The designed cutting table profiling control system improved the reliability of ridge detection and achieved low crushing rate and low leakage rate profiling harvesting operations for leafy vegetable harvesters.

关键词

叶菜收获机/点云/机器学习/拟合/仿形

Key words

leaf vegetable harvester/point cloud/machine learning/fitting/profiling

分类

农业科技

引用本文复制引用

杨云聪,姚立健,刘宇,王守卫,姚黎文..基于机器学习的叶菜收获机割台仿形控制系统研究[J].农机化研究,2025,47(11):28-37,10.

基金项目

浙江省尖兵领雁攻关项目(2022C02042) (2022C02042)

农机化研究

OA北大核心

1003-188X

访问量0
|
下载量0
段落导航相关论文