首页|行株间机械除草机横向纠偏系统设计与试验

行株间机械除草机横向纠偏系统设计与试验

扫码查看
为提高机械除草机田间精准作业能力,避免除草部件产生较大偏移伤害作物,设计了横向纠偏系统并进行相关试验分析.该系统由作物行识别与偏移量计算子系统、机械纠偏装置和控制执行子系统3部分组成.作物行识别与偏移量计算子系统由Intel RealSense D435深度相机、Jetson NX嵌入式计算设备等组成.通过改进的YOLOv5深度学习算法对作物行进行识别定位,再对得到的作物行中心点采用最小二乘法拟合直线计算出偏移量,发送给控制执行子系统.控制执行子系统接收作物行识别与偏移量计算子系统发送的偏移量,通过PID算法控制液压缸伸缩,实现机械纠偏装置的横移对行.试验结果表明:在拖拉机前进速度分别为0.63、1.56、2.25 km/h时,横向纠偏系统的平均对行准确率分别为91.8%、88.7%、87.4%,对行纠偏平均响应时间分别为1.4、0.8、0.5 s.该系统可以有效缓解农机具在大田除草作业过程中对行不准造成作物植株损伤的问题.
Design and Experiment of a Lateral Correction System for Inter/Intra-Row Mechanical Weeder
In order to improve the precision operation ability of the mechanical weeder in the field and avoid the large deviation of the weeding parts to damage the crops,a lateral correction system was designed and related experiments were analyzed.The system consists of three parts:crop row identification and offset calculation subsystem,mechanical deviation correction device and control execution subsystem.The crop row identification and offset calculation subsystem consists of Intel RealSense D435 depth camera,Jetson NX embedded computer,etc.The improved YOLOv5 deep learning algorithm was used to identify and locate the crop rows,and then the offset was calculated by fitting the straight line with the least squares method to the center point of the crop rows and sent to the control execution subsystem.The control execution subsystem received the offset from the crop row identification and offset calculation subsystem,and controlled the expansion and contraction of the hydraulic cylinder through the PID algorithm to realize the transverse alignment of the mechanical deviation correction device.The test results showed that when the forward speed of the tractor was 0.63,1.56 and 2.25 km/h,respectively,the average accuracy of the lateral correction system was 91.8%,88.7%and 87.4%,respectively,and the average response time of the correction was 1.4,0.8 and 0.5 s,respectively.The system can effectively alleviate the problem of crop damage caused by inaccurate alignment of agricultural machinery in the process of field weeding operations.

lateral correctiondeep learninghydraulic systemPID controltest

刘天佐、高原源、赵明宝、周婧、袁源、王爱臣

展开 >

江苏大学农业工程学院,江苏镇江 212013

横向纠偏 深度学习 液压系统 PID控制 试验

2025

西南大学学报(自然科学版)
西南大学学报编辑部

西南大学学报(自然科学版)

北大核心
影响因子:0.825
ISSN:1673-9868
年,卷(期):2025.47(1)