《Proceedings of the Second International Symposium on Intelligent Information Technology in Agriculture (ISIITA)》2003年
THE VERTICAL NITROGEN DISTRIBUTION IN CANOPY AND ITS SPECTRAL RESPONSE IN WINTER WHEAT
【摘要】：正 The precision of quantificational remote sensing (RS) can be improved by tapping the potential of sensor, transferring model and exactly describing of the object. We think exactly describe the object is the base of improving inversion accuracy of RS. For this purpose, the vertical distribution of leaf nitrogen (N) concentration and Chl(a+b) content in canopy and their spectral response in winter wheat were investigated. Apparent descending trend of leaf N from the upper layer to the lower layer were found. At early stage,there were stable N gradient between different layers. Greater N gradient emerged at middle and latter growth stage and the gradient sharpened with the N application amount. The vertical distribution of Chl(a+b) was similar to that of N. But there were greater gradient of Chl(a+b) concentration than that of N content between upper and middle layer and less between middle and lower layer. The amount of N application decreased the Chl (a+b) gradient between middle and lower layer, which differ from that of leaf N gradient. Under lower N condition, there existed significant spectral reflectance difference among different layers at the red wavebands, from 1400 nm to 1800 nm and from 1950 nm to 2300 nm. And, the spectral reflectance of lower layer was significantly higher than that of upper and middle layer. But the amount of N application didn't affect the characteristics of spectral reflectance for leaf in different layers. Besides, the correlative coefficients between canopy spectral reflectance and foliar biochemical contents of different layers were analyzed. The spectral reflectance was significantly correlated with foliar Chl (a+b) content of middle and lower layers. We thought the inversion accuracy of RS could be improved by using layer information. Future work should involve giving more attention on the method for multi-angle analyzing and establishing inversion model.