A Review of Optical-Based Three-Dimensional Reconstruction and Multi-Source Fusion for Plant Phenotyping.

Abstract

In the context of the booming development of precision agriculture and plant phenotyping, plant 3D reconstruction technology has become a research hotspot, with widespread applications in plant growth monitoring, pest and disease detection, and smart agricultural equipment. Given the complex geometric and textural characteristics of plants, traditional 2D image analysis methods are difficult to meet the modeling requirements, highlighting the growing importance of 3D reconstruction technology. This paper reviews active vision techniques (such as structured light, time-of-flight, and laser scanning methods), passive vision techniques (such as stereo vision and structure from motion), and deep learning-based 3D reconstruction methods (such as NeRF, CNN, and 3DGS). These technologies enhance crop analysis accuracy from multiple perspectives, provide strong support for agricultural production, and significantly promote the development of the field of plant research.

Plant phenotyping relevance

植物フェノタイピング向けの3次元再構築技術を体系的にレビューしており、画像ベースの形質取得手法が中心的です。

Code and data availability

This is a review article surveying 3D reconstruction techniques for plant phenotyping. The supplied blocks contain no authors' phenotype datasets, plant images, sensor data, analysis code, trained models, or supplements with such assets. All referenced works are cited prior studies, and no public repository or data- or

No evidence-backed public reproduction asset is currently recorded.