Authors

Abstract

In order to meet the demand for mechanisation of sunflower segmented harvesting and tray insertion for drying, an intelligent tray insertion dryer was designed and developed. The machine integrates the functions of disc picking, disc flipping, bar clipping, lifting and inserting, and uses SolidWorks for accurate mechanism design and simulation, incorporating the YOLOv5 model for efficient recognition of flower discs and the LeGO-LOAM algorithm for accurate navigation and map building. In the experiment, 81 sunflower samples were collected to analyse data on disc diameter, plant height, rod diameter and stalk diameter, and to verify the recognition accuracy of the YOLOv5 model in different directions. The results showed that the precision of disc recognition was 95.54%, accuracy was 89.94%, recall was 95.54% and F1 value was 0.89. Using the LeGO-LOAM algorithm tested at different path lengths, the root-mean-square error of the navigational build trajectory was 0.15 m, with a standard deviation of 0.10 m. This technological integration improves the operational efficiency and supports the mechanisation of sunflower insertion tray drying.

Abstract in Chinese

为了满足向日葵分段收获和插盘晾晒的机械化需求，我们设计并开发了一款智能插盘晾晒机。该机集成了摘盘、翻盘、剪杆、升降和插盘等功能，采用SolidWorks进行精确的机构设计和模拟，融合了YOLOv5模型的高效识别花盘和LeGO-LOAM算法的精确导航建图。实验中，采集了81株向日葵样本，分析花盘直径、株高、杆径和梗径数据，并验证YOLOv5模型在不同方向上的识别精度。结果显示，花盘识别精度达95.54%，准确度为89.94%，召回率为95.54%，F1值为0.89。利用LeGO-LOAM算法在不同路径长度下测试，导航建图轨迹的均方根误差为0.15m，标准差为0.10m。这种技术整合提高了作业效率，为向日葵插盘晾晒械化提供了支持。