Infrared imaging and machine learning techniques for plant root location and depth prediction

Abstract

A plant’s root system absorbs water and necessary nutrients, and synthesizes organic matter, which is essential for plant growth and regeneration. Therefore, investigating root system architecture (RSA) can potentially provide deep understanding and useful information about plant growth. Current approaches involve soil-coring and use of mini-rhizotrons, which can damage the root or be time consuming. Groundpenetrating radar has been employed but is not suitable for small plants because of the resolution needed. Nuclear magnetic resonance could provide valuable information of tiny roots, but the equipment is costly. In this study, infrared imaging—a-non-destructive method—was used to reveal the shape and position of small root systems, such as sugar beet roots. The finite element analysis (FEA) methodology was implemented toA plant’s root system absorbs water and necessary nutrients, and synthesizes organic matter, which is essential for plant growth and regeneration. Therefore, investigating root system architecture (RSA) can potentially provide deep understanding and useful information about plant growth. Current approaches involve soil-coring and use of mini-rhizotrons, which can damage the root or be time consuming. Ground-penetrating radar has been employed but is not suitable for small plants because of the resolution needed. Nuclear magnetic resonance could provide valuable information of tiny roots, but the equipment is costly. In this study, infrared imaging, a-non-destructive method, was used to reveal the shape and position of small root systems, such as sugar beet roots. The finite element analysis (FEA) methodology was implemented to validate the practicality of applying infrared imaging to detect roots. Artificial neural network (ANN) methods were used to determine the existence of a root system. Support vector machine (SVM) and ANN were employed to predict root depth and statistical tests were used to compare the results. The results of these experiments suggest that infrared imaging can be used to predict the presence and depth of roots. validate the practicality of applying infrared imaging to detect roots. Artificial neural network (ANN) methods were used to determine the existence of a root system. Support vector machine (SVM) and ANN were employed to predict root depth and statistical tests were used to compare the results. The results of these experiments suggest that infrared imaging can be used to predict the presence and depth of roots.

Plant phenotyping relevance

赤外線画像と機械学習を用いて植物根系の存在、形状、位置、深さを推定する手法を開発・検証しており、植物フェノタイピング手法が研究の中心である。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.