基于Faster-RCNN算法的玉米叶面病害识别系统

Corn Leaf Disease Recognition System Based on Faster R-CNN Algorithm

针对玉米种植面积广,但易受病害影响导致产量下降,农民对玉米叶面病害识别困难的问题,以玉米叶面健康、大斑病、小斑病和锈病4种叶面种类为研究对象,采用Faster-RCNN建立识别模型,并在此基础上开发智能玉米叶面识别系统。首先对5998张图片的数据集采用LabelImg工具进行分类标注,训练集和验证集比例为9:1;然后使用ResNet50神经网络架构对标注好的数据集进行训练,得到最优权重的PTH文件;最后将算法通过API接口部署到使用Django搭建的后端框架中,在前端调用算法进行玉米叶面检测识别,结合Neo4j知识图谱,将玉米叶面病害种类的详细信息以及解决方案,通过知识图谱进行展示。玉米叶面病害识别结果表明:平均识别很高,达到96%。

In response to the widespread cultivation of maize, which is susceptible to diseases leading to decreased yields, this study developed an intelligent maize leaf disease recognition system. Focusing on four types of leaf diseases—healthy, common rust, gray leaf spot, and northern leaf blight—a Faster R-CNN model was trained using a dataset of 5998 annotated images categorized through the LabelImg tool, with a training-validation split ratio of 9:1. Employing the ResNet50 neural network architecture, optimal weights were obtained in a .pth file. Subsequently, the algorithm was deployed through an API integrated into a backend framework built with Django, facilitating frontend maize leaf detection. Leveraging a Neo4j knowledge graph, detailed information on maize leaf disease types and solutions were presented. Results demonstrated an average recognition rate of 96%.