Quantification of canopy heterogeneity and light interception difference within greenhouse cucumbers based on terrestrial laser scanning

Abstract

Reconstructing 3D architecture of cucumber populations for multi-scale phenotypic analysis poses significant challenges in greenhouse crop research. Cucumber canopy architecture directly impacts light interception and the plant growth conditions. Terrestrial Laser Scanning (TLS) was employed to capture the 3D point cloud of cucumber plants at various growth stages, named as real plant canopy (RPC). A novel method, CP-FEC-RG, combining Fast Euclidean Clustering with Region Growing algorithm, was developed to segment cucumber plants and extract phenotypic traits both at plant and leaf scales. The virtual plant canopies (VPCs), namely VPC-H, VPC-M and VPC-L were constructed representing high, medium, and low growth potentials based on the data collected via TLS. A radiative transfer model was adopted to compare the radiation interception capabilities of both RPC and VPCs. An average recall rate of 92.2% was achieved for leaf segmentation. Growth differences were observed among the segmented individual plants and leaves, with coefficients of variations for phenotypic traits ranging from 0.13 to 0.48 for individual plants and from 0.21 to 0.54 for leaves. For daily cumulative light interception, VPC-L showed a reduction of 17.1% compared to RPC, whereas VPC-M and VPC-H exhibited increases of 18.2% and 30.1%, respectively. These findings highlight the importance of using the RPC for the accurate calculations of light interception and provide a solid foundation for applying TLS in the 3D phenotypic analysis of crops in solar greenhouses.

Plant phenotyping relevance

TLSによる3D形状取得、植物・葉の分割、表現型形質抽出手法を開発・評価しており、フェノタイピング手法が研究の中心である。

Code and data availability

公開状態または取得可能な本文経路を確認できませんでした。

No evidence-backed public reproduction asset is currently recorded.