作　　者： ; ; ;

机构地区： 华南农业大学工程学院南方农业机械与装备关键技术省部共建教育部重点实验室

出　　处： 《现代科学仪器》 2010年第2期10-14,共5页

摘　　要： 为了研究作物病虫害发生后植株冠层多光谱图像特征值的变化规律,设计了一套室内多光谱视觉检测系统试验平台。文章介绍了该试验平台在稻飞虱发生早期水稻冠层多光谱图像特性研究中的应用。在实验室条件下对盆栽稻株进行接虫,然后分12个时段(接虫后14h、16h、20h、22h、24h、38h、44h、48h、62h、69h、73h)对受稻飞虱侵害后的稻株冠层进行多光谱图像采集,共采集到NIR、R、G、B四个通道及其组合通道(CIR、RGB)的稻株冠层图像561张,然后在Matlab中进行图像处理和数据分析,结果表明:稻飞虱迁入后,NIR通道的冠层叶片灰度值算术平均值在时间维度上光谱图像特征值变化非常显著,其次是G、R通道。因此,在稻飞虱迁入后,应着重观察NIR、G、R通道冠层叶片灰度值算术平均值在时间维度上的变化规律,从中挖掘用于稻飞虱入侵检测和危害程度分级的相关模型。 In order to study the crop stress induced by diseases or insect pests,a Multi-Spectral Vision Detection System（MSVDS）was developed.Used the MSVDS,a pilot study was conducted to determine the feasibility using multispectral image processing techniques to detect stress in rice cased by Rice Plant-Hopper（RPH）.Rice seeds were planted in plastic flats.After insect pests were infested with RPHs（mixed with Nilaparvata Lugens and Sogatella Furcifera）at a density of 13 per plant,canopy multi-spectral images,including the channels of NIR,R,G,and B,were acquired at 12 time points（14h,16h,20h,22h,24h,38h,44h,48h,62h,69h,and 73h after infested respectively）.Images and derived vegetation indices were processed and analyzed using Matlab.Results showed that there are significantly changes of the leaf gray value of NIR channel images during the early stages of RPH infestation.It indicated the potential for detecting RPH-induced stress in rice using a MSVDS.

关 键 词： 稻飞虱 多光谱图像 时间维度 胁迫检测 遥感

领　　域： [机械工程] [机械工程] [机械工程]