基于遥感图像的甘蔗收割机路径规划方法
Path planning method for sugarcane harvesters based on remote sensing images
潘明章 1曾玥 2向本杰 3阙文帅 2苟轩源 2官维2
作者信息
- 1. 广西大学机械工程学院,广西南宁 530004;亚热带农业生物资源保护与利用国家重点实验室,广西 南宁 530004
- 2. 广西大学机械工程学院,广西南宁 530004
- 3. 广西玉柴机器股份有限公司,广西玉林 537005
- 折叠
摘要
针对我国目前甘蔗收割机作业高度依赖人工操作、自主导航效率低的问题,本文提出了一种基于遥感图像特征提取技术的新颖路径规划方法.首先利用图像处理技术建立了甘蔗田遥感图像的环境模型,然后利用Floyd算法进行田间转场路径的搜索与长度计算.此外,通过将转场路径长度作为亲和度评价标准,对免疫算法进行改进,实现了多田块的遍历次序优化,进而解决了对不规则形状甘蔗田的全局路径规划问题.基于Floyd算法的改进免疫算法迭代进化最快,收敛时间最短,且能够更有效适应不同复杂程度的甘蔗田遥感图像.
Abstract
To solve the issues of high manual dependence and low autonomous navigation rates in sugarcane harvester operation in China,a path planning method based on remote sensing images was proposed.Firstly,an environmental model for remote sensing images of sugarcane fields was established through image processing techniques.Next,a search and length calculation for inter-field transition paths were implemented based on the Floyd algorithm.The length of transition paths was then used as an affinity evaluation index to improve the immune algorithm,which was used to optimise the traversal order of multiple fields,and solve the problen of global path planning for irregularly shaped sugarcane fields.The improved immune algorithm based on the Floyd algorithm has the fastest iterative evolution,the shortest algorithm convergence time,and it can best adapt to remote sensing images of sugarcane fields with varying complexities compared to other path planning algorithms.
关键词
甘蔗收割机/路径规划/图像特征提取/Floyd算法/免疫算法Key words
sugarcane harvester/path planning/image feature extraction/Floyd algorithm/immune algorithm引用本文复制引用
基金项目
国家自然科学基金项目(U23A20330)
出版年
2024
NETL
NSTL
万方数据