摘要：Before measuring the spectral reflectance of a crop canopy, it is necessary to determine the distance from spectrometer to canopy or measure the distance in combination with field calibration after fixing the distance from canopy to spectrometer. However, the actual distance from canopy to spectrometer is hard to obtain as a canopy’s reflecting surface is not a planar surface, increasing the difficulty of collecting data in agricultural fields and measurement errors. This paper describes a method of automatically measuring the distance from spectrometer to canopy using a binocular vision system. This method involved the use of binocular vision to obtain the three-dimensional (3D) point cloud data of crop canopy as well as a statistical analysis and weighting of the data obtained to automatically detect the distance from spectrometer to canopy. Reduced measurement errors compared to manual distance measurement and simplified reflectance measurement process can be achieved using this method. In this study, a set of miniature binocular cameras and an independently designed spectrometer were assembled and used to conduct a field measurement of the reflectance of wheat canopy. In the measurement experiment, the distance from spectrometer to canopy top was between 60 and 100 cm. The 3D point cloud density of the wheat canopy in the vertical direction followed a normal distribution; the values of canopy reflectance calculated by weighting the 3D point cloud data were stable, with a maximum relative deviation of 5.92%, an average relative deviation of 3.91%, and a relative standard deviation of 3.39%. The experimental results suggested that the proposed method enables automatic and stable determination of the distance from the spectrometer to canopy and avoids the need to consider the distance from canopy to spectrometer during reflectance measurement. Therefore, this simplified method can facilitate field measurement of spectral reflectance using a hand-held or rack-mounted spectrometer.
