在英国诺丁汉大学Sutton Bonington校区(52.8°N, 1.2°W)实测感染锈病与普通花叶病大豆的单叶光谱数据, 利用连续统去除法对原始光谱数据进行处理, 筛选对病害及锈病严重度敏感的波段, 构建植被指数对感染锈病与普通花叶病及不同严重度锈病的大豆进行识别研究。 研究发现普通花叶病胁迫下的大豆光谱反射率在可见光区域均大于健康大豆的, 而锈病胁迫的大豆光谱反射率在绿光区随病情严重度增加而减小, 在红光区随病情严重增强而增大。 根据大豆光谱变化特征设计了一个植被指数R500×R550/R680对大豆病害进行识别, 通过计算不同病害及不同严重度之间的J-M距离对指数识别病害能力进行检验, 结果表明指数R500×R550/R680能够较好的识别出大豆锈病与普通花叶病, 且该指数在识别大豆锈病严重度方面也有较强的能力。 研究结果对农作物病害遥感监测与防治具有重要的理论价值与实际应用意义。

The objective of this paper is to identify disease and its severity of soybean by using single leaf spectral data in the field. The soybean spectral were measured in the Sutton Bonington Campus of University of Nottingham( 52.8°N, 1.2°W), which infected rust disease(RD) and common mosaic disease (CMD), respectively, and continuum removal method was used to process the original spectral data, and sensitive bands were selected for disease and disease severity, and vegetation index was designed for identifying RD and CMD of soybean. The result showed spectral reflectance of soybean under CMD stressed is more than that of health in the visible region. However, spectral reflectance of soybean under RD stressed will decrease in the green region and that will increase in the red region with disease severity increasing. According to the spectral changing features, a new index R500×R550/R680 was designed for identifying the disease of soybean. In order to test the index identifying disease ability, the J-M distances were calculated among health, RD and CMD. The result indicated index R500×R550/R680 can better identify RD and CMD, at the same time, the index has good ability for discriminating the disease severity of soybean. The research results of this paper has important theoretical value for crops disease monitoring and prevention and practical application meanings.