基于无人机遥感的大面积农作物叶片病虫害识别研究  

Identification of Crop leaf diseases and insect pests based on UAV remote sensing

作　　者：贺园园[1] HE Yuanyuan(Xi’an Traffic Engineering Institute,Xi’an 710300,China)

机构地区：[1]西安交通工程学院,西安710300

出　　处：《自动化与仪器仪表》2023年第1期216-220,共5页Automation & Instrumentation

基　　金：校级中青年基金项目《基于地铁的城市物流配送系统研究》(21KY-32)。

摘　　要：病虫害会影响农作物的产量,在农业生产过程中,需要及时发现并准确识别出病虫害,才能够保证农业的种植产量,研究基于无人机遥感的大面积农作物叶片病虫害识别方法。配置八旋翼无人机和多光谱相机,搭建无人机遥感识别平台;以棉花叶螨为病虫害识别对象,选择农作物生长数据,划分健康叶片和叶螨病害范围。通过无人机遥感技术设计识别方法,利用叶螨的自身反射曲线,对应病虫害光谱指数特征,分辨不同等级下叶螨的反射光谱情况;定义遥感影像中的像素阈值,建立遥感影像识别模型,对图像前后景熵值进行分割;利用几何校正原理,对应划分区域采样点的大地坐标系,多项式转换图像位置坐标,对叶螨进行识别。实验结果表明:在叶螨发生的高峰期进行识别,能够直接对健康棉田和螨害区域进行划分,并且可以在空间和时间序列上,分析叶螨生长特点,具有实际应用 效果。Diseases and insect pests will affect the output of crops.In the process of agricultural production,it is necessary to timely discover and accurately identify diseases and insect pests to ensure the output of agricultural planting.In this paper,the identification method of large area crop leaf diseases and insect pests based on UAV remote sensing is studied.Equipped with eight-rotor UAV and light sensing sensor,uav remote sensing recognition platform is built.Using cotton tetranychus mites as pest identification objects,crop growth data were selected to divide the range of healthy leaves and tetranychus mites.The uav remote sensing technology was used to design the identification method,and the reflection spectrum of tetranychus mites at different levels was distinguished by using the reflection curve of tetranychus mites to correspond to the spectral index characteristics of pests and diseases.The pixel threshold of remote sensing image is defined,the recognition model of remote sensing image is established,and the entropy value of foreground and background image is segmented.According to the principle of geometric correction,the geodetic coordinate system of the sampling points in the region was used to transform the image position coordinates by polynomial to identify the tetranychus mites.The experimental results showed that the identification of tetranychus mites at the peak of occurrence could directly divide the healthy cotton field and mite infestation area,and analyze the growth characteristics of tetranychus mites in space and time series,which had practical application effect.

关 键 词：无人机遥感 大面积农作物 大病虫害 识别方法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]