| 基于高光谱数据的棉田虫害鉴别研究 |
| |
| 引用本文： | 王小龙,邓继忠,黄华盛,邓宇森,蒋统统,钟兆基,张亚莉,文晟.基于高光谱数据的棉田虫害鉴别研究[J].华南农业大学学报,2019,40(3):97-103. |
| |
| 作者姓名： | 王小龙 邓继忠 黄华盛 邓宇森 蒋统统 钟兆基 张亚莉 文晟 |
| |
| 作者单位： | 国家精准农业航空施药技术国际联合研究中心,广东广州510642;华南农业大学工程学院,广东广州510642;国家精准农业航空施药技术国际联合研究中心,广东广州510642;华南农业大学工程基础教学与训练中心,广东广州510642 |
| |
| 基金项目： | 广东省科技计划项目（2017A020208046，2017B010117010）；国家重点研发计划项目（2016YFD0200700）；国家发展和改革委2014年北斗卫星导航产业重大应用示范发展专项项目（20142564）；广东省教育厅重点平台及科研项目（2015KGJHZ007）；广州市科技计划项目（201707010047） |
| |
| 摘 要： | 【目的】快速、准确、无损伤地鉴别棉花虫害类别,以便针对性制定植保施药方案。【方法】对棉花叶片高光谱数据进行采集和分析。采用波段范围为350～2 500 nm的FieldSpec?3便携式光谱分析仪,分别获取受蚜虫和红蜘蛛危害的棉花叶片以及正常棉花叶片的高光谱数据。采用K-近邻和SVM算法区分受红蜘蛛和蚜虫侵害的叶片以及正常叶片。为进一步优化虫害识别模型、提高识别精度,利用主成分分析方法 (PCA)进行特征降维,并利用网格搜索法进行参数寻优。【结果】使用K-近邻算法和SVM算法构建了虫害识别模型,2种模型的识别率分别为86.08%和89.29%;引入PCA进行特征降维并使用网格搜索进行参数寻优后,可以提高虫害识别率,K-近邻算法和SVM算法的识别精度分别达到88.24%和92.16%。【结论】利用高光谱数据可以区分受蚜虫和红蜘蛛侵害以及正常的棉花叶片;结合PCA降维和网格搜索法,能够提高识别率且不需要获得具体的特征波段;对于受蚜虫和红蜘蛛侵害以及正常的叶片识别,基于径向基核函数的SVM算法优于K-近邻算法。
|
| 关 键 词： | 棉花虫害 K-近邻 支持向量机 高光谱数据 无损鉴别 |
| 收稿时间： | 2018/7/25 0:00:00 |
Identification of pests in cotton field based on hyperspectral data |
| |
| WANG Xiaolong,DENG Jizhong,HUANG Huasheng,DENG Yusen,JIANG Tongtong,ZHONG Zhaoji,ZHANG Yali and WEN Cheng.Identification of pests in cotton field based on hyperspectral data[J].Journal of South China Agricultural University,2019,40(3):97-103. |
| |
| Authors: | WANG Xiaolong DENG Jizhong HUANG Huasheng DENG Yusen JIANG Tongtong ZHONG Zhaoji ZHANG Yali and WEN Cheng |
| |
| Institution: | National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China and National Center for International Collaboration Research on Precision Agricultural Aviation Pesticides Spraying Technology, Guangzhou 510642, China;Engineering Fundamental Teaching and Training Center, South China Agricultural University, Guangzhou 510642, China |
| |
| Abstract: | |
| |
| Keywords: | cotton pest K-nearest neighbor support vector machine hyperspectral data non-destructive identification |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《华南农业大学学报》浏览原始摘要信息 |
| 点击此处可从《华南农业大学学报》下载免费的PDF全文 |
|
