A Nondestructive Method to Estimate Plant Height, Stem Diameter and Biomass of Rice under Field Conditions Using Digital Image Analysis

Abstract

Plant phenotyping intends measuring complex plant traits, and is important in agricultural research for enhancing yield improvement. Manual plant phenotyping is laborious and destructive, and hence a less-laborious and non-destructive method is required. Here, we proposed a nondestructive method to estimate continuous data of plant traits such as height, stem diameter and biomass using a low cost time-lapse camera. The camera was installed at a rice field in Japan, and captured images for four target plants every three hour. The plant height and stem diameter were determined from the images by referencing scale bars that were placed next to the target plants and above the ground surface. Both the height and the diameter were compared to directly measured ones, and the relationships between those were in good agreement. Plant volumes were estimated from the height and stem diameter assuming a shape of rice plant is cylindrical. Above ground biomass without panicles was determined by rice plants sampled from the field. The determined biomass increased in proportion to the plant volume, and its relationship used to produce continuous data of the rice biomass. The results suggest that the proposed method can be considered as a useful tool of the plant phenotyping.

Evaluation of RGB and Multispectral Unmanned Aerial Vehicle (UAV) Imagery for High-Throughput Phenotyping and Yield Prediction in Barley Breeding

“…The additional information about the canopy architecture can then be integrated into the modelling of yield formation and biomass and has the potential to improve the models [28,48]. On the one hand, many authors reported that RGB indices outperform spectroradiometric VIs in terms of characterizing plant growth [49][50][51], separating vegetation from bare soil [52], assessing nitrogen [27,53] and detecting foliar diseases [54,55]. This could be in part due to the fact that spectroradiometric indices exhibit longer wavelengths in the NIR range than RGB indices and are thus more susceptible to canopy architecture, which affects the reflective properties of plants, and soil mixing pixels, which occurs especially at low spatial resolution [56].…”