资 源 简 介
We present a geometrical method for detecting water stress levels among Mungbean (Vigna radiata) plant samples grown under a controlled environment. Thirty plant samples, divided into two setups (setup A for dried plants and setup B for overwatered), were subjected to manual stressing and were processed. Using an image of a sample, a series of image processing operations - binarization, thresholding (manual and automatic), erosion, dilatation, blob filtering, contour tracing, and skeleton approximation were used to approximate the stem and the leaf angles of the Mungbean plant. Using the computed plant angles (leaves and stem), three key plant features (stem angle, average of the angle of the leaves, and the average difference of the leaf angle from the stem) were then correlated with the soil moisture content of the sample. Results showed that there is a linear association between the key features and the moisture content of its soil for both setups. The stress coefficient, a singl