基于多模态传感器融合识别的四臂采摘机器人研究  

Research on the four-arm harvesting robot based on multi-modal sensor fusion recognition

作　　者：王利 郑子培 刘子谦 肖瑩 张金珠 梁小蝶 WANG Li;ZHENG Zipei;LIU Ziqian;XIAO Ying;ZHANG Jinzhu;LIANG Xiaodie(Guangdong Technology College,Zhaoqing 526100,Guangdong,China)

机构地区：[1]广东理工学院,广东肇庆526100

出　　处：《农业装备与车辆工程》2025年第3期17-21,共5页Agricultural Equipment & Vehicle Engineering

基　　金：广东理工学院2024年大学生创新训练计划省级项目“基于3D打印技术的自适应水果采摘机器人结构设计”(S202413720017);广东理工学院2024年度校级课堂教学改革项目“机器人学基础”(KTJXGG2024008);广东理工学院2022年科技项目(2022GKJZK011);广东理工学院2021年高等教育教学改革项目(JXGG202108)。

摘　　要：为提高农业生产采摘效率和降低人工成本,对基于多模态传感器融合识别的四臂采摘机器人开展研究。通过优化视觉识别系统和融合多种传感器信息,可实现对水果的精准识别和定位,结合贝叶斯估计和边缘检测算法,系统在复杂环境下对水果位置和形状的识别准确率得到提高。研究结果表明,多模态传感器融合识别的四臂采摘机器人可以协同工作,进一步提高采摘效率。为农业生产提供了一种新型的高效率采摘解决方案,具有较高的应用价值和推广前景。In order to improve the picking efficiency of agricultural production and reduce labor cost,a four-arm picking robot based on multi-modal sensor fusion recognition was studied.By optimizing the visual recognition system and integrating a variety of sensor information,accurate recognition and positioning of fruits can be achieved.Combined with Bayesian estimation and edge detection algorithms,the system could improve the recognition accuracy of fruit position and shape in complex environments.The research results showed that the four-arm picking robot identified by multi-modal sensor fusion could work together,further improving the picking efficiency.It provides a new high efficiency picking solution for agricultural production,and has high application value and promotion prospect.

关 键 词：多模态 传感器融合 采摘机器人 结构设计 控制系统 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]