Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey on 3D Reconstruction Techniques in Plant Phenotyping: From Classical Methods to Neural Radiance Fields (NeRF), 3D Gaussian Splatting (3DGS), and Beyond

Jiajia Li, Xinda Qi, Seyed Hamidreza Nabaei, Meiqi Liu, Dong Chen, Xin Zhang, Xunyuan Yin, Zhaojian Li

TL;DR

The paper addresses the need for scalable, accurate 3D reconstruction in plant phenotyping to enable automated high-throughput trait analysis. It surveys classical active and passive reconstruction methods alongside cutting-edge neural approaches like NeRF and 3DGS, detailing their methodologies, applications, and performance in diverse crops. It discusses the strengths and limitations of NeRF (photorealism and depth capture versus computational cost) and 3DGS (real-time rendering and scalability with occlusion challenges), and surveys evaluation metrics across pixel-level, geometry-level, and trait-level tasks. It outlines future directions including multi-modal data fusion, hyperspectral 3D reconstruction, VR/AR visualization, and integration with downstream phenotyping applications to enable robust, scalable plant analysis.

Abstract

Plant phenotyping plays a pivotal role in understanding plant traits and their interactions with the environment, making it crucial for advancing precision agriculture and crop improvement. 3D reconstruction technologies have emerged as powerful tools for capturing detailed plant morphology and structure, offering significant potential for accurate and automated phenotyping. This paper provides a comprehensive review of the 3D reconstruction techniques for plant phenotyping, covering classical reconstruction methods, emerging Neural Radiance Fields (NeRF), and the novel 3D Gaussian Splatting (3DGS) approach. Classical methods, which often rely on high-resolution sensors, are widely adopted due to their simplicity and flexibility in representing plant structures. However, they face challenges such as data density, noise, and scalability. NeRF, a recent advancement, enables high-quality, photorealistic 3D reconstructions from sparse viewpoints, but its computational cost and applicability in outdoor environments remain areas of active research. The emerging 3DGS technique introduces a new paradigm in reconstructing plant structures by representing geometry through Gaussian primitives, offering potential benefits in both efficiency and scalability. We review the methodologies, applications, and performance of these approaches in plant phenotyping and discuss their respective strengths, limitations, and future prospects (https://github.com/JiajiaLi04/3D-Reconstruction-Plants). Through this review, we aim to provide insights into how these diverse 3D reconstruction techniques can be effectively leveraged for automated and high-throughput plant phenotyping, contributing to the next generation of agricultural technology.

A Survey on 3D Reconstruction Techniques in Plant Phenotyping: From Classical Methods to Neural Radiance Fields (NeRF), 3D Gaussian Splatting (3DGS), and Beyond

TL;DR

The paper addresses the need for scalable, accurate 3D reconstruction in plant phenotyping to enable automated high-throughput trait analysis. It surveys classical active and passive reconstruction methods alongside cutting-edge neural approaches like NeRF and 3DGS, detailing their methodologies, applications, and performance in diverse crops. It discusses the strengths and limitations of NeRF (photorealism and depth capture versus computational cost) and 3DGS (real-time rendering and scalability with occlusion challenges), and surveys evaluation metrics across pixel-level, geometry-level, and trait-level tasks. It outlines future directions including multi-modal data fusion, hyperspectral 3D reconstruction, VR/AR visualization, and integration with downstream phenotyping applications to enable robust, scalable plant analysis.

Abstract

Plant phenotyping plays a pivotal role in understanding plant traits and their interactions with the environment, making it crucial for advancing precision agriculture and crop improvement. 3D reconstruction technologies have emerged as powerful tools for capturing detailed plant morphology and structure, offering significant potential for accurate and automated phenotyping. This paper provides a comprehensive review of the 3D reconstruction techniques for plant phenotyping, covering classical reconstruction methods, emerging Neural Radiance Fields (NeRF), and the novel 3D Gaussian Splatting (3DGS) approach. Classical methods, which often rely on high-resolution sensors, are widely adopted due to their simplicity and flexibility in representing plant structures. However, they face challenges such as data density, noise, and scalability. NeRF, a recent advancement, enables high-quality, photorealistic 3D reconstructions from sparse viewpoints, but its computational cost and applicability in outdoor environments remain areas of active research. The emerging 3DGS technique introduces a new paradigm in reconstructing plant structures by representing geometry through Gaussian primitives, offering potential benefits in both efficiency and scalability. We review the methodologies, applications, and performance of these approaches in plant phenotyping and discuss their respective strengths, limitations, and future prospects (https://github.com/JiajiaLi04/3D-Reconstruction-Plants). Through this review, we aim to provide insights into how these diverse 3D reconstruction techniques can be effectively leveraged for automated and high-throughput plant phenotyping, contributing to the next generation of agricultural technology.
Paper Structure (23 sections, 21 equations, 7 figures, 4 tables)

This paper contains 23 sections, 21 equations, 7 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Preliminaries of 3D Reconstruction
  3. Classical methods
  4. Active reconstruction methods
  5. Passive reconstruction methods
  6. Neural radiance fields (NeRF)
  7. 3D Gaussian splatting (3DGS)
  8. Metrics for evaluating 3D reconstruction
  9. Pixel-level metrics
  10. Geometry-level metrics
  11. Trait-specific metrics
  12. Current Progress of 3D Reconstruction in Plant Phenotyping
  13. Classical methods
  14. NeRF in plant phenotyping
  15. 3DGS in plant phenotyping
  16. ...and 8 more sections

Figures (7)

  • Figure 1: Comparison of different 3D reconstruction techniques and their corresponding reconstruction processes: (a) LiDAR, (b) Structured Light, (c) Structure from Motion (SfM), (d) Neural Radiance Fields (NeRF), and (e) 3D Gaussian Splatting (3DGS).
  • Figure 2: Architecture of AgriNeRF and downstream fruit detection. Adapted from chopra2024agrinerf.
  • Figure 3: Workflow comparison of NeRF-based and traditional 3D reconstruction methods for corn plants. Adapted from arshad2024evaluating.
  • Figure 4: Multi-view renderings of the NeRF model (a) Frontviews (b) Side views (c) Top views (d) Elevation views zhao2024exploring.
  • Figure 5: Framework for 3D phenotyping for bell pepper using NeRF-based reconstruction and 3D scanning zhao2024exploring. The process integrates action camera and 3D scanner data, followed by NeRF reconstruction, scale restoration, segmentation, and phenotypic measurements.
  • ...and 2 more figures