基于弱监督学习的小样本早期苹果叶片病害检测  

Early detection of apple leaf disease with few-shot based on weakly supervised learning

作　　者：王勇[1] 周强 吴凯 Wang Yong;Zhou Qiang;Wu Kai(Modern Education Technology Center,Kaifeng Vocational College of Culture and Arts,Kaifeng,475000,China;School of Electronic Information,Xi an Polytechnic University,Xi an,710048,China;Chengdu Institute of Computer Application,Chinese Academy of Sciences,Chengdu,610041,China)

机构地区：[1]开封文化艺术职业学院现代教育技术中心,河南开封475000 [2]西安工程大学电子信息学院,西安市710048 [3]中国科学院成都计算机应用研究所,成都市610041

出　　处：《中国农机化学报》2024年第12期193-199,共7页Journal of Chinese Agricultural Mechanization

基　　金：国家自然科学基金(62101021)。

摘　　要：针对现有苹果叶片病害检测方法的性能过度依赖标注数据集的问题,提出一种基于弱监督学习的小样本早期苹果叶片病害检测算法。首先,利用一组共享权重的主干网络将病害叶片映射到高维特征空间;其次,利用多层注意力机制建立双分支特征语义关联模块,并在关联语义特征图上生成指导查询图片中新病害类型分类的原型集;再次,利用无参数的匹配方法计算原型集与查询图片中新病害叶片特征间的相似度,根据相似度值定位与识别病害区域;最后,利用虚线框标注建立弱监督学习机制,并借助标签平滑交叉损失端到端优化模型。通过在开源的Plant Village数据集和自建的早期苹果叶片病害数据集上进行试验,所提出方法分别实现96.39%、94.81%的精准率,96.71%、94.67%的召回率和97.24%、95.20%的F1值,优于当前经典的目标识别算法。In order to address the problem of excessive reliance on annotated datasets in existing apple leaf disease detection methods,a few-shot early apple leaf disease detection algorithm based on weakly supervised learning was proposed.Firstly,a set of backbone networks with shared weights was used to map the diseased leaves into a high-dimensional feature space.Secondly,a dual-branch feature semantic correlation module was established by using a multi-layer attention mechanism,and prototype sets for classifying new disease types in query images were generated on the correlation semantic feature map.Thirdly,the similarity between the prototype sets and the features of diseased leaf in the query image was calculated by using a non-parametric matching method,and the disease regions were located and recognized based on the similarity values.Finally,a weakly supervised learning mechanism was established by using dashed rectangle annotations,and the model was optimized end-to-end with the help of label smooth cross-entropy loss.Experimental results on the open-source Plant Village dataset and a self-built early apple leaf disease dataset demonstrated that the proposed method achieved precision rates of 96.39%、94.81%,recall rates of 96.71%、94.67%,and F1 scores of 97.24%、95.20%.

关 键 词：苹果叶片病害检测 小样本学习 弱监督学习 多层注意力机制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] S436.611[自动化与计算机技术—计算机科学与技术]