Abstract
Density of soybean canopy was measured by the image analysis of color and infrared pictures taken by 35mm cameras and its applicability to the estimation of biomass of soybean canopy was examined. The measurements of the density were based on the gray level of pixels on digital image data in which the original information of the density of the canopy was non-linearly biased through several signal conversions such as photographing, videotizing, decoding of composit video signal, A/D conversion and etc. In this study, in order to compensate the bias, a multi-step gray-scale was photographed in the identical pictures with those to be analyzed for measuring the density of the soybean canopy. The gray level of the digital image was standardized so that the gray level of the steps of the gray scale in the digital image was proportional to the actual density of the steps. Then, the density of the objective canopy was calculated from the average gray level of the canopy in the standardized digital image. For an objective canopy, digital images in three different spectral bands were obtained from a color picture and a digital image in a near-infrared band was obtained from an infrared picture. Consequently density values in 4 different bands corresponding to those digital images were measured.
The soybean canopies for 3 types of varieties, 3 kinds of planting density and 4 different growth stages were examined and the exponentially transformed values of the density or the functions of those showed somewhat linear relations with the soybean canopy biomass such as LAI, fresh weight, dry weight and dry leaf weight. The results were similar to those given before by the studies with spectral-meters and indicated that the procedure in this study was applicable to estimate the soybean canopy biomass.
