稻纵卷叶螟危害后水稻叶片的光谱特征

【目的】阐明水稻受稻纵卷叶螟危害后不同受害程度的叶片、卷叶的分布形式及卷叶率对稻叶光谱特征的影响,获取诊断水稻受害程度的模型,以便为稻纵卷叶螟的遥感监测提供理论指标与方法。【方法】试验以不同受害等级的虫害叶及健康叶为材料,在室内恒定条件下采用ASD光谱仪分别测定不同受害程度、受害叶片的不同分布形式、及不同卷叶率下稻叶的光谱反射率,并采用直线回归法,建立基于光谱参数的水稻受害程度诊断模型。【结果】水稻虫害叶光谱反射率均随受害等级的增加,在绿光区(530—570nm)和近红外区(700—1050nm)降低,而在红光区(610—700nm)增加。能反演叶片受害程度的敏感波段为530—564nm、614—695nm和706—1050nm。建立了5个反演叶片受害程度的模型,诊断准确率在80%—90%之间,并且以741nm处的反射率对叶片受害程度的诊断效果最好。在卷叶率恒定的条件下,卷叶的分布位置对光谱反射率影响较小;而卷叶率对光谱反射率的影响较大,表现为随卷叶率的增大,450—500nm和610—700nm处的反射率增大,530—570nm和700—1050nm处反射率降低。差值植被指数(Rnir-Rred)、黄边面积(SDy)及红边面积与蓝边面积的差值(SDr-SDb)等指标均能将6个不同等级的卷叶率(0、10%、30%、50%、70%和90%)区分开,并且利用黄边面积(SDy)指标诊断卷叶率的准确率达86%。【结论】水稻受稻纵卷叶螟为害后,在叶片光谱反射率上有明显的表现,可以利用光谱特征来监测稻叶的受害程度及卷叶率大小。

Spectral Characteristics of Rice Leaves Damaged by Rice Leaf Roller

【Objective】 The effects of the damage degree, spatial pattern, and rate of damaged leaves by rice leaf roller (RLR) on spectral characteristics of leaves were studied in this paper in order to construct models for diagnosing the damage of rice, and to attain the parameters and methodology for automatic monitoring of the rice leaf roller. 【Method】 The spectral reflectances of different damage degrees, spatial patterns and damage rates of rice leaves were measured by ASD Hand-held Spectroradiometers, and the diagnosing models based on the spectral parameters were established using the linear regression method. 【Result】 The results indicated that the spectral reflectance of rice leaf decreased in the green (530-570 nm) and near-infrared (700-1 050 nm) wave-length regions with the increase of the damage degree whereas it increased in the red wave-length region (610-700 nm). The regions of 530-564 nm, 614-695 nm and 706-1 050 nm were the sensitive wave-length bands which could reflect the impaired level of rice leaf by the RLR. Five models were established for diagnosing the damage degree of rice leaf, and their diagnosing accuracies were 80%-90%. The maximum diagnostic accuracy belonged to the model based on the reflectivity at 741 nm. The spectral reflectance of rice leaves was less impacted by the spatial pattern of the damaged leaves when the rate of the damaged leaves was constant. However, the spectral reflectivity increased significantly at 450-500 nm and 610-700 nm, but decreased at 530-570 nm and 700- 1 050 nm as the rate of damaged leaves increased. The spectral indexes including vegetation index (Rnir-Rred), the areas of the yellow-edge (SDy), and the difference of the areas of the red-edge and blue-edge (SDr-SDb) could significantly distinguish the six grades of leaf roll rates (0%, 10%, 30%, 50%, 70%, and 90%). The area of yellow edge (SDy) index could diagnose the leaf roll rate and the accuracy of diagnoses was up to 86%. 【Conclusion】 The damage of rice leaves by RLR could be monitored by the spectral reflectance. Spectral indexes could reflect the harm degree of rice leaves and the rate of infested leaves.