Development of Non-destructive Discrimination Model for Water Stress in Chinese Cabbage Seedlings using Hyperspectral Imaging

The emphasis and importance of non-destructive evaluation technique for field crop has been on the increase recently, because the effects of abnormal climates such as high temperature, drought, and salinity, has caused growth disorders in crops, leading to quality deterioration and reduced yields. In this study, a model was developed to evaluate moisture stress in Chinese cabbage using short wave infrared hyperspectral imaging (SWIR HSI) technique. To develop the optimal model, various preprocessing methods were employed with the Partial Least Squares (PLS) and Least Squares Support Vector Machines (LS-SVM) techniques. In addition, the discrimination of water stress between normal and water-stressed Chinese cabbage seedlings was evaluated using the value of volumetric water content (VWC) in the soil and morphological changes of SWIR HSI images. Consequently, the normal and water-stressed groups were discriminated using the morphological change and VWC. In addition, the discrimination accuracy of PLS and LS-SVM models was approximately 89 % and 91 %, respectively. Therefore, the developed models demonstrated a high potential for early detection of water stress in for Chinese cabbage.