建立大田早稻农学光谱估产模式研究初报

大田试验结果表明,不同氮素水平的水稻冠层光谱之反射率存在着极显著差异。其中630～690nm,760～900nm和520～600nm三个波段可作为检测水稻氮素水平的敏感波段。利用测定光谱反射,通过光谱变量计算以估测LAI和Chl等农学参数是可行的。单一生育期光谱建模估产的精度较低,而且应用时生育期也难统一。RVI复合估产模式和LAI-RVI综合估产模式的精度很高,但仍只是“样板模式”,实际应用于大田估产时,仍需进行多点试验,进一步验证模式并建立一套修正方法,以适应不同熟制和不同产量水平的需要。

Study on the establishment of agronomic spectral yield estimation model for early rice in the field.

The field experiment was conducted to establish the agronomic spectral yield estimation model for the early rice. The results showed that the rice canopy reflecting spectra of various nitrogen levels had significant differences,and the bands of 630-690,760-900 and 520-600 nm could be used as the sensitive bands for the determination of the nitrogen contents. It is feasible to estimate the LAI,ch1 and the other agronomic parameters by spectral reflectance determination through the vegetation index calculations. The accuracy of yield estimation models established by one stage spectral data was not satisfied. But the accuracies of comprehensive yield es ima ion models and comprehensive LAI-RVI yield estimation models, which were established by two or three stages' data,were quite high. Multiple field experiments of different sites have to be done for the model modification to adapt the different production condition,especially for various crop rotations and rice species.