Crop Status Sensing System by Multi-Spectral Imaging Sensor (Part 2)

Wheat Growth Estimation and GIS Map Generation

Abstract

The main objective of this research is to develop a remote sensing system that can determine crop status using a multi-spectral imaging sensor (MSIS). To detect crop information, high image quality is required under various ambient light conditions. Useful data was obtained by adjusting CCD gain before sensing and controlling exposure time for each capture. The unplanted ridges in the image makes the error of the vegetation cover rate. Therefore, the algorithm which can automatically segment the unplanted ridge area was developed. In this research, the images were obtained at flag leaves stage and flowering stage. The crop statuses were estimated by the multiple-linear regression of the sensing data and ground-truth data (SPAD value, crop height, stem number, dry weight, nitrogen contents and yield). As a result of the analysis, the coefficients of determination at the flag leaves stage were 0.72 (SPAD value), 0.59 (crop height), 0.58 (stem umber), 0.59 (dry weight), 0.62 (nitrogen contents) and 0.61 (yield). The coefficients of determination of SPAD value, nitrogen contents and yield at the flowering stage were 0.72, 0.62 and 0.54. However, the other parameters didn't show any correlation. Finally, the GIS maps regarding crop status were generated for the experimental field.