Hyperspectral Estimation On Chlorophyll And Water Contents In Young Apple Leaves

Water or chlorophyll content is a good leading indicator of the growth of fruit trees,fruit tree leaves like the color, and quality and their morphological structures along withthe water and chlorophyll content sof how many in a series of changes, spectrumreflectivity will be corresponding change. Therefore, it can be done according to thereflective spectrum characteristics of blades for fruit trees moisture content of chlorophyll,real-time monitoring and diagnosis. This research took two-year cane Fuji as testmaterials, wintry spring shoots stop growing for apple leaf chlorophyll, water content andspectral reflectance, first order differential spectrum, spectral parameters such ascorrelation analysis, and at the same time picking adult fruit apple leaf simple comparativeanalysis. Based on different data processing methods of sensitive wavelength and band,using linear regression analysis, principal component analysis, stepwise regression analysisto construct high spectral estimation model of chlorophyll and water content, and accuracyof inspection, the following results:(1) Young apple leaves within the scope of visible light spectral reflectance slightlyhigher than the mature trees, it has a higher reflectivity than fruit of fruit trees by around3%in750-1350nm band range. But in1350-2500nm wavelength range, spectralreflectance of the high and low to distinguish is not obvious. Young trees leaf spectralcurve has a tendency to shift to the left on the whole spectrum from350nm to2500nm. Inred edge position such as location, spectral peaks, absorption valley have more obviousmanifestations with a difference of3-7nm.(2) The different spectral reflection characteristic of chlorophyll content of youngleaves, and in different wavelength range show different pattern: with the increase ofchlorophyll content and spectral reflectance in the green area around550nm reflectivity isreduced, increased reflectivity in red light district from700to780nm. Which in1350-2500nm wavelength range, the higher chlorophyll content, the higher the spectralreflectance. Young tree leaves reflectance peaks at around1650nm and2250nm in largerdifference under different moisture content conditions. (3)Chlorophyll content and spectral reflectance in young leaves more than one orderdifferential of the band has a good correlation, water content and lnR, R ’good correlation,the highest correlation coefficient reached.. And the study found that there is negativecorrelation between the spectra of the two bands420nm and555nm build parameters andchlorophyll, moisture content increase20%to30%.(4) With the first order differential spectra data sensitive wavelength and band, throughthe principal component analysis to build the model compared with other model couidbetter estimate of young leaf chlorophyll content:y=1.7539-157.1959*R′530-108.0764*R′693+2.6512*R′700+443.2963x4*R′587-23.0239*R′731+42.6015*R′741-1.8254*R693-712R~2=0.6562, RMSE=0.05742.(5) With the logarithmic transformation data by stepwise regression method was usedto build model compared with other model can better estimate of young leaf water content,y=0.4530+2.9029*ln R756-3.3445*ln R948-4.1786*ln R1572+3.6750*ln R1705+2.5089*ln R~2091-1.9856ln R~2302R~2=0.5699, RMSE=0.02860.