基于改进CenterNet的柑橘叶片病害检测
Disease detection of citrus leaves based on improved CenterNet
李栋 1仲婷 1王笋 2李大华 1于晓1
作者信息
- 1. 天津理工大学电气工程与自动化学院天津市复杂系统控制理论与应用重点实验室,天津 300384
- 2. 天津同实智研科技有限公司,天津 300384
- 折叠
摘要
针对柑橘叶片病害表现大小不一,为解决检测过程中出现的漏检、误检、准确率不高的问题,提出了改进CenterNet模型.在特征提取网络RestNet 50的前两个残差层的一系列残差结构中引入特征增强改进的空洞卷积池化金字塔(improved atrous spatial pyramid pooling,IASPP)模块,扩大浅层感受野,获取更多小目标叶片病害的细节信息,增强浅层特征的显著性;引入双向加权特征融合模块(bi-directional feature pyramid network,BiFPN),有效融合特征提取网络浅层和深层叶片病害信息;为提高整体检测效果,引入多尺度通道注意力机制(multi-scale channel attention module,MS-CAM).训练后的模型对柑橘病害叶片进行检测,实验结果表明,相比于原模型CenterNet,所提模型的R值提高了8.32％,mAP提高了4.53％,AP 0.5∶0.95上升了27.3％,可实现柑橘种植中对叶片小目标、中目标、大目标病害的精准检测.
Abstract
Aiming at the different sizes of the disease manifestations of citrus leaves,the model of improved CenterNet is proposed to solve the problems of missing detection,false detection and low accuracy in the detection process.The feature enhancement improved atrous spatial pyramid pooling(IASPP)module was introduced into a series of residual structures of the first two residual layers of the feature extraction network RestNet50 to expand the shallow receptive field,obtain more detailed information of small target leaf diseases,and enhance the significance of shallow features.The bidirectional feature pyramid network(BiFPN)module was introduced to effectively integrate the shallow and deep leaf disease information of the feature extraction network.In order to improve the overall detection effect,the multi-scale channel attention module(MS-CAM)was introduced.The trained model was used to detect citrus disease leaves.The experimental results show that,compared with the original model CenterNet,the R-value of the proposed model is increased by 8.32％,mAP is increased by 4.53％,and AP 0.5∶0.95 is increased by 27.3％.It can achieve accurate detection of small target,medium target and large target leaf diseases in citrus planting.
关键词
柑橘叶片病害/CenterNet/多尺度特征融合/特征增强Key words
citrus leaf disease/CenterNet/multi-scale feature fusion/feature enhancement引用本文复制引用
基金项目
国家自然科学基金(61502340)
天津市自然科学基金(18JCQNJC01000)
天津市教委科研计划项目(2018KJ133)
天津市复杂系统控制理论与应用重点实验室开放基金项目(TJKL-CATCS-201907)
天津理工大学教学基金项目(YB20-05)
出版年
2024
NETL
NSTL
万方数据