Spectral Characteristics Analysis and Water Content Detection of Potato Plants Leaves

Abstract

The water content of potato leaves in different leaf positions was studied using hyperspectral technique in this research. According to leaf position, the leaves were divided into upper, middle and lower 3 groups. The experiment was conducted, in which the fresh leaves were collected and dried, and the reflectance spectra and water content were measured. The relationship between the spectral characteristics of different leaf positions and water content distribution of potato plants was analyzed. It was found that the reflectance and water content of fresh leaves in different positions were different. With the leaf position from top to bottom, the reflectance increased successively (R upper <R middle<R lower) in 862.9-1311.9 nm; the reflectance reduced in 1311.9-1403.6 nm; the reflectance at the middle and upper leaves was significantly higher than lower leaves (R upper >R middle >R lower) in 1403.6-1704.2 nm and the average value of water content increased in turn. After drying, with the leaf position from top to bottom, the spectral reflectance of leaves decreased successively (R upper<R middle<R lower) in 862.9-1704.2 nm and the average water content increased in turn. It also showed that the reflection spectra of 862.9-1311.9 nm was sensitive to the difference of leaf age and tissue structure for dry processing; the correlation coefficient between the reflectance and water content was greater than 0.93 in 1400.3-1500.7 nm. Combining the correlation analysis and PCA, 866.4 nm and 1406.8 nm were selected to establish the multiple linear regression (MLR) model for the water content detection of potato leaves, the model calibration coefficient (Rc2) was 0.9507 and the validation coefficient (Rv2) was 0.8189.The validation coefficients in the upper, middle, and lower leaf position were 0.9481, 0.7900 and 0.7673.