摘要
以自然环境下的枇杷为研究对象,基于EasyDL定制化训练和服务平台,建立针对成熟枇杷和未成熟枇杷的识别训练模型,并分别用100张未在训练数据集内的成熟枇杷和未成熟枇杷图片素材对训练模型进行验证,识别结果大部分满足要求。通过对比分析成熟枇杷的识别结果数据,总结得到其识别结果受到环境光强度、拍照视野和图像质量以及枇杷的品种等多方因素的影响,导致识别结果变化幅度大。该研究弥补了对于自然环境中成熟枇杷识别的空白,也为今后枇杷采摘机器人的实现打下研究基础。
This project takes loquat in the natural environment as the research object,establishs recognition training models for mature and immature loquats based on EasyDL customized training and service platform,and verifies the training model with 40 mature loquat and immature loquat pictures that are not in the training data set,and most of the recognition results meet the requirements.By comparing and analyzing the recognition results of mature loquat,it is concluded that the recognition results are affected by many factors,such as ambient light intensity、camera field of vision、image quality and loquat varieties,resulting in large changes in the recognition results.The proposal of this project makes up for the blank of mature loquat recognition in the natural environment,and also lays a research foundation for the realization of loquat picking robot.
作者
徐航煌
钱佳宁
马伟平
XU Hanghuang;QIAN Jianing;MAWeiping(New Engineering Industry College,Putian University,Putian,Fujian,351100,China;Electromechanical and Information Engineering College,Putian University,Putian,Fujian,351100,China)
出处
《福建农机》
2024年第1期29-33,38,共6页
Fujian Agricultural Machinery
基金
莆田学院科研项目(2019124)
福建省中青年教师教育科研项目(科技类)(JAT220295)。
