The Study On Hyperspectral Remote Sensing Monitoring Of Hydro-regime In Cotton Based On Fluorescence Parameters

[Research object]:By setting of a field exper iment with two cotton cultivars among four level of irrigating water treatments,the paper studied the characteristics of the reflectance and the fluorescence parameters in different critical growth stages of cotton,and established the correlation relationship between the chlorophyll fluorescence parameters and spectral parameters,it could provide a theory for real-time and rapid monitoring the water stress by using hyperspectral remote sensing.[Research methods]:The Filedspec Pro FR 2500 backhanging ground hyperspectral radiometer（ASD,US A）was used for the determination of two cotton cultivars with severe water shortage irrigation,I1(1050m³·hm^-2),water shortage irrigation I2(2160 m³·hm^-2),appropriate irrigation I3(4320 m³·hm^-2)and full irrigation(6300 m³·hm^-2).A series of hyperspectral variables were analyzed and calculated by the canopy reflectance spectra of the critical growth period.At the same time,the PAM-2100 portable chlorophyll fluorescence analyzer（Walz,Effeltrich,Germany）was used to synchronize the leaf fluorescence parameters of two cotton cultivars.The yield and fiber quality information under different water treatments were obtained at the harvest period.The correlations between spectral variables and each index were analyzed and the remote sensing models for monitoring the water regime of cotton were established.[Research results]:（1）The yield of two cotton cultivars increased with the increasing of irrigation amount,that is,I4>I3>I2>I1.The yield of Xinluzao NO.45 was significantly higher under I3 and I4 treatments than that of drought treatments（I1 and I2）,while the seed cotton yield of Xinluzao NO.62 had a very signif icant difference between four water treatments.The upper-half-mean length,uniformity index,breaking tenacity and elongation percentage of cotton fiber increased with the increasing of water irrigation amount,while the micronaire value increased with the severity of drought stress.And the upper-half-mean length,uniformity index,breaking tenacity,micronaire value and elongation percentage of cotton fiber among the different water treatments reached significantly different（p<0.05）or highly dominant differences（p<0.01）.This study showed that the seed cotton yield and fiber quality index of two cotton cultivars had signif icant or extremely significant differences among different water gradients.It was proved that the water gradient was set effectively,so the hyperspectral monitoring of the drought condition of cotton could be further studied on this basis.（2）Under different moisture treatments,the canopy reflectance spectra of two cotton cultivars were similar,and the D_λred（red edge slope）and Sred（red edge area）of the two cultivars increased signif icantly with the increasing of water irrigation.This study show ed that red edge parameters can effectively monitor the water status of cotton canopy.The ten vegetation indices of each cotton variety at full bloom stage and full bolling stage were significantly positively or negatively related to lint yield（p<0.01）.And the vegetation indices with the greatest correlation of two cotton cultivars appeared at the full bolling stage,corresponding to RVI and MSAVI,respectively(r_45-RVI=0.899**,r_62-MSAVI=0.991**,n=20,P<0.01).（3）The diurnal variation trend of the fluorescence parameters Fm’and Yield of two cotton cultivars was"V"type,and they were reduced to the minimum at 14:00 p.m..The diurnal variation trend of 1-qP and qN was the same,showing the trend of first increasing and then decreasing,most of which reached the maximum at 14:00 p.m..The variance analysis of the fluorescence parameters at the full bolling stage showed that qP was decreased by drought stress,while qN increased.The fluorescence parameters of the two cotton cultivars showed significant or extremely significant correlation with the vegetation index.During the full-bolling stage,the Fv/Fm of single leaf of Xinluzao NO.62 was positively correlated with canopy reflectance with the bands of 350-514 nm,612-692 nm and 1945-2076 nm,but was negatively correlated with the bands of 708-1361 nm and 1621-1740 nm.The vegetation index,NDVI and RVI,were constructed by using sensitive bands,and there was a higher correlation between RVI and Fv/Fm(r_RVI-Fv/Fm=-0.7215^**,n=20,p<0.01),so established a correlated model b etween Fv/Fm and RVI,the Fv/Fm value was positively correlated between the estimated value and tested value(r估测_Fv/Fm-实测_Fv/Fm-45=0.7678^**,RMSE=2.591×10^-2,r估测_Fv/Fm-实测_Fv/Fm-62=0.7215^**,RMSE=1.186×10^-2,n=20).It showed that based on the correlation between Fv/Fm and RVI,the water stress of two cotton cultivars during the full-bolling stage could be monitored by using RVI.The Fv/Fm of single leaf of two cotton cultivars at five critical growth stages was signif icantly negatively correlated with canopy reflectance with the bands of 714-1362nm and 1599-1748 nm bands,and the highest correlation coefficient appeared in 1207nm(r_ρ1207=-0.5510^**,n=40,p<0.01).Established a correlated model between Fv/Fm andρ₁₂₀₇,the Fv/Fm value was positively correlated between the estimated value and tested value(r_{restimated Fv/Fm-measured Fv/Fm}=0.551^**,n=40,RMSE=1.182×10^-2).It was proved that based on the correlation between Fv/Fm andρ₁₂₀₇,the water stress condition of cotton during the whole growth period can be monitored by usingρ₁₂₀₇ quickly and undestructively.