Automated classification of stems and leaves of potted plants based on point cloud data

Abstract

The accurate classification of plant organs is a key step in monitoring the growing status and physiology of plants. A classification method was proposed to classify the leaves and stems automatically based on the point cloud data of the potted plants. Leaf samples and stem samples were selected automatically by using the three-dimensional (3D) convex hull algorithm and the two-dimensional (2D) projection grid density respectively, and were used to construct the leaf and stem training sets. Then, the point cloud data were classified into leaf points and stem points by using the support vector machine (SVM) algorithm. The point cloud data of three potted plants were used in the experiment. The proposed method was compared with the standard classification, the random selection method and the manual selection method. Among these methods, the proposed method is automated and time-saving. The results show that the proposed method had a good overall performance on accuracy and running time. The proposed method is efficient and effective on the leaf and stem classification of the plant point cloud data.

Plant phenotyping relevance

植物の点群データから葉と茎を自動分類する手法の開発と比較検証が研究の中心であり、器官構造の表現型抽出に該当する。

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.