摘要
夹切一体的菠萝(Ananascomosus(L.)Merr.)采摘器在进行田间采摘作业时,需要自主确定果茎切割点位置,而菠萝果茎处容易被植株叶片和苞叶遮挡,采用单一图像处理的方法难以准确识别到果茎切割点位置,为此提出一种多传感器信息融合的菠萝果茎切割点位置检测方法。将深度相机和多组光电传感器结合,利用改进的YOLOv5目标检测算法融合RGB-D深度信息,实现对菠萝冠芽顶部至果实底部长度测量,再利用光电传感器信号变化判断菠萝采摘器是否到达冠芽顶部位置,并将冠芽顶部作为起始位置,控制采摘器下降速度和时间,从而保证采摘器底部安装的切割刀准确抵达果茎切割点位置。台架试验结果表明,该方法对真实菠萝果茎切割点检测成功率达到85%,满足菠萝采摘机器人作业过程中果茎切割点检测准确性要求。
The pineapple(Ananas comosus(L.)Merr.)picker with integrated clip needs to indepen⁃dently determine the cutting points in fruit stem of pineapple when picking in the field.Pineapple stems are easily obstructed by plant leaves and bracts,making it difficult to accurately identify the cutting points in fruit stem with a single method of image processing.A method for detecting the cutting points in fruit stem of pineapple based on the fusion of multi-sensor information was proposed.The length from the top of the pineapple crown bud to the bottom of the fruit was measured by combining a depth camera with multiple sets of photoelectric sensors and utilizing an improved YOLOv5 object detection algorithm to fuse RGB-D depth information.The changes in signal of photoelectric sensor were used to determine whether the pineap⁃ple picker has reached the top position of the crown bud.The top of the crown bud was used as the starting position to control the descent speed and time of the picker to ensure that the cutting blade installed at the bottom of the picker accurately reaches the cutting point of the fruit stem.The results of bench test showed that this method had a success rate of 85%in detecting the cutting points in real fruit stem of pineapple,meeting the accuracy requirements of detecting the cutting points in fruit stem of pineapple during the opera⁃tion of pineapple picking robots.It will be of great significance for realizing the intelligent pineapple pick⁃ing,and will lay a foundation for the subsequent development of pineapple picking robots in the field.
作者
焦锐
马瑞峻
陈瑜
伍恩慧
杨金鹏
温国政
潘雄
JIAO Rui;MA Ruijun;CHEN Yu;WU Enhui;YNAG Jinpeng;WEN Guozheng;PAN Xiong(College of Engineering,South China Agricultural University,Guangzhou 510642,China)
出处
《华中农业大学学报》
CAS
CSCD
北大核心
2024年第5期21-30,共10页
Journal of Huazhong Agricultural University
基金
广东省科技计划项目(2021B1212040009)。
关键词
菠萝
采摘器
果茎切割点
位置检测
多传感器信息融合
YOLOv5
pineapple(Ananas comosus(L.)Merr.)
picker
cutting points in fruit stem
position detection
fusion of multi sensor information
YOLOv5