期刊论文

Rapeseed root tumor grading method based on improvedResNet34 network and transfer learning

中文

武汉轻工大学学报

2095-7386

2023

42

03

108-114+119

10.3969/j.issn.2095-7386.2023.03.016

2021年湖北省教育厅教学研究项目(2021351)。

本校为第一机构

针对传统油菜根肿瘤分级方法精准性差、效率不高等问题,提出了一种基于改进的ResNet34网络和迁移学习的油菜根肿瘤分级算法。利用数据增强技术对采集的油菜根肿瘤图像进行样本数据的扩充,在ResNet34模型的基础上,选择不...展开更多 针对传统油菜根肿瘤分级方法精准性差、效率不高等问题,提出了一种基于改进的ResNet34网络和迁移学习的油菜根肿瘤分级算法。利用数据增强技术对采集的油菜根肿瘤图像进行样本数据的扩充,在ResNet34模型的基础上,选择不同的位置增加SE(Squeeze-and-Excitation)注意力分支,强化ResNet34对油菜根部病害区域的特征识别能力,提高网络识别油菜根肿瘤的精度,最后利用迁移学...

Aiming at the problems of poor accuracy and low efficiency of traditional rapeseed root tumor grading methods,a rapeseed root tumor grading algorithm based on improved ResNet34 network and transfer learning was proposed.To expand the sample data of collected rapeseed ro...MORE Aiming at the problems of poor accuracy and low efficiency of traditional rapeseed root tumor grading methods,a rapeseed root tumor grading algorithm based on improved ResNet34 network and transfer learning was proposed.To expand the sample data of collected rapeseed root tumor images,data enhancement technology was util...