农机化研究2024，Vol.46Issue(4)：58-62,5.
深度学习算法下的采摘机器人系统优化研究
Study on System Optimization for Picking Robot Based on Deep Learning Algorithm
摘要
Abstract
In order to achieve the automation for apple picking,this optimization for picking robot was designed based on image depth learning algorithm.Firstly,the filter processing was done for image collected by the CCD camera,based on gaussian filtering.Secondly,regional recommendation learning network was added to VGG16 deep learning network,so that identification sample size was reduced,and recognition rate was improved.Thirdly,coordinate transformation be-tween image coordinates and actual spatial coordinates was achieved by monocular vision system and laser rangefinder,picking manipulator control was achieved by double feedback system.In the end,the system was tested,the result showed that this system had accuracy target recognition and manipulator control accuracy.关键词
采摘机器人/改进型深度学习网络/高斯滤波/视觉定位Key words
picking robot/improved deep learning network/gaussian filtering/visual positioning分类
农业科技引用本文复制引用
张军凯,李欣,韩俊先,赵娟,程龙雪..深度学习算法下的采摘机器人系统优化研究[J].农机化研究,2024,46(4):58-62,5.基金项目
河北省高等学校科学技术研究项目(ZC2021222) （ZC2021222）
