摘要
随着水果产业链自动化的进步,水果采摘机器人也在不断地发展,对目标水果实现精准识别是水果采摘机器人最重要的组成部分之一。本文将对水果目标识别的各类算法应用状况进行对比,同时阐述各类算法在水果采摘机器人上的应用现状。其中,现有的水果目标识别方法主要有3部分,分别是传统图像处理技术、各种机器学习算法和使用各类卷积神经网络的深度学习算法。此外,论述了水果图像数据集常用的处理方法和算法模型的改进方法,通过对图像数据进行处理增强和算法的改进来提高对目标的识别准确率。最后,分析各类算法在水果目标识别的未来研究趋势。
With the development of fruit industry chain automation,fruit picking robot was also being developed and studied.Accurate recognition of target fruit was one of the most important components of fruit picking robot.This paper compares the application status of various algorithms for fruit target recognition,and expounds the application status of various algorithms in fruit picking robot.Among them,the existing fruit target recognition methods mainly have three parts:traditional image processing technology,various machine learning algorithms and deep learning algorithms using various convolutional neural networks.In addition,the common processing methods of fruit image data set and the improvement method of algorithm model are discussed.Through the processing and enhancement of image data and the improvement of algorithm,the accuracy of target recognition was improved.Finally,the future research trend of various algorithms in fruit target recognition was analyzed.
作者
陈品岚
朱立学
张世昂
Chen Pinlan;Zhu Lixue;Zhang Shiang(College of Mechanical and Electrical Engineering,Zhongkai University of Agriculture and Engineering,Guangzhou 510225,China;College of Automation,Zhongkai University of Agriculture and Engineering,Guangzhou 510225,China)
出处
《现代农业装备》
2022年第2期8-13,42,共7页
Modern Agricultural Equipment
基金
广东省重点领域科技研发计划项目(2019B20223003)
广东省现代农业产业技术体系创新团队项目(粤农函[2019]1019号)
南亚热带水果关键技术研究与装备实验项目(KB2010008)。
关键词
采摘机器人
水果识别
图像处理
机器学习
深度学习
picking robot
fruit identification
image processing
machine learning
deep learning
