Precise 3D measurement of internode elongation just below shoot apex for growth diagnosis of greenhouse cucumber plants using SfM/MVS method

Abstract

Measurement of internode elongation just below the shoot apex or growing point of the main stem is important for assessing plant growth. However, it is difficult to directly measure internode elongation on climbing plants with many leaves, such as cucumber plants. It is also difficult to measure the stem length of tall leafy plants in the field, and is prone to measurement errors. In addition, touching plants to measure them can stress them. Here, we measured internodal growth just below the shoot apex by using a 3D point cloud model reconstructed using Structure from Motion and Multi-View Stereo (SfM/MVS) methods under greenhouse conditions. The SfM/MVS method could nondestructively measure the internode elongation of multiple plants accurately and simultaneously with a root mean square error of 3.1 mm. Elongation was most active in the top two internodes and ceased in older internodes. Average elongation lengths of internodes 1 and 2 as counted from the top (6.7-7.7 mm day-1) were significantly greater than that of internode 3 (3.37 mm day-1), which was significantly greater than those of internodes 4 to 6 (0.0-0.5 mm day-1). These growth rates of two top internodes are the indicator of plant growth, which can be used for plant diagnosis. Our quantitative method for assessing internode elongation can be used under normal greenhouse conditions. Traditional 2D measurements face challenges due to occlusion, which this 3D method overcomes by digitally removing leaves for clear node visibility. 3D measurements enable time series analysis of internode elongation, which is difficult to measure in situ. The 3D data can be stored for later reanalysis.

Plant phenotyping relevance

SfM/MVSによるキュウリの節間伸長を非破壊・3D・同時測定する方法を開発し、精度（RMSE）を検証している。植物形質の取得法が研究の中心である。

Code and data availability

The article describes SfM/MVS 3D reconstruction of cucumber plants and internode elongation measurements, but contains no data availability statement, no public repository deposit, and no author URL for images, point clouds, phenotypic data, or analysis code. All URLs in the text are citations to prior work. Software (

No evidence-backed public reproduction asset is currently recorded.