Australian Plant Phenomics Network · ANU Node

PhenoFusion3D

Integrated 3D–Hyperspectral Plant Analytics

Extract structural and spectral traits from plant scans. Visualise plants in 3D, derive key hyperspectral indices, and correlate traits with phenotypes to accelerate plant science and breeding decisions.

Explore Features→View Pipeline
PhenoFusion3D — Plant Scan Analysis
Plant Scans
Wheat_Scan_001
Barley_Phenotype_07
Canola_Trial_B
Sorghum_Row_12
NDVI
0.72
Chlorophyll
42.3
LAI
3.8
Explore
About the Project

Accelerating Plant Science with Precision Analytics

PhenoFusion 3D integrates cutting-edge 3D scanning with hyperspectral imaging to provide an unprecedented view into plant health, growth, and genetic traits. Developed at the ANU node of the Australian Plant Phenomics Network (APPN), it enables researchers to measure plant performance efficiently, objectively, and non-destructively over time.

By correlating structural geometry with spectral data, PhenoFusion3D accelerates the development of improved crops for a changing climate — supporting food security and more productive, sustainable farming systems.

Food SecuritySustainable FarmingCrop DevelopmentClimate Resilience
400+
Spectral Bands Processed
3D
Point Cloud Visualisation
10x
Faster Trait Analysis
100%
Non-Destructive Scanning
🌿
Partner Organisation
Australian Plant Phenomics Network (APPN)
Providing national plant phenotyping solutions, backed by specialist technical and data expertise. ANU Node · The Australian National University
Visit APPN →
Core Capabilities

Everything You Need for Plant Phenomics

A complete desktop analytics toolkit built for plant scientists and breeders.

Interactive 3D Visualisation

Explore plant geometry with full 3D point cloud rendering. Overlay spectral bands and indices directly onto plant structure for intuitive spatial analysis.

Hyperspectral Band Mapping

Process 400+ spectral bands with automated white/dark calibration. Compute NDVI, chlorophyll, moisture, and custom spectral indices from raw scan data.

Automated Trait Extraction

Run end-to-end trait extraction pipelines — from raw scan ingestion through background removal to structured, quantitative phenotypic output.

Phenotype Correlation Engine

Correlate extracted spectral and structural traits with agronomic features or breeding targets. Identify meaningful genotype–environment interactions.

Calibration Workflows

Built-in white reference and dark current calibration pipelines ensure measurement accuracy and reproducibility across different scanning conditions.

Exportable Research Outputs

Generate publication-ready tables, plots, and reports. Export processed data in standard formats compatible with common research and breeding workflows.

Analysis Pipeline

From Raw Scan to Actionable Insights

A reproducible, end-to-end pipeline for plant phenomics research.

STEP 01

Load Plant Scans

Import raw hyperspectral and 3D lidar/structured-light scan data from supported sensor formats. Supports batch loading of multiple plant specimens.

STEP 02

Calibrate & Preprocess

Apply white reference and dark current calibration workflows. Automated background removal isolates plant structure for clean analysis.

STEP 03

3D Reconstruction

Generate dense 3D point clouds representing plant geometry. Align and fuse multi-view scans into a unified plant model.

STEP 04

Spectral Mapping

Project hyperspectral data onto the 3D model. Compute spectral indices (NDVI, chlorophyll, water content) per point in the cloud.

STEP 05

Trait Extraction

Run automated trait extraction pipelines to quantify structural features (height, branching, leaf angle) and spectral phenotypes.

STEP 06

Export & Report

Generate structured output tables, visualisation plots, and reproducible pipeline reports ready for downstream analysis and publication.

phenofusion3d — pipeline output
$ phenofusion3d run --input scans/wheat_001/ --calibrate --extract-traits
[INFO] Loading hyperspectral scan: wheat_001_hyperspectral.hdr
[INFO] Applying white reference calibration... ✓
[INFO] Background removal (Otsu threshold)... ✓
[INFO] 3D point cloud reconstruction (156,892 pts)... ✓
[INFO] Spectral mapping: 421 bands projected... ✓
[INFO] Trait extraction pipeline running...
→ NDVI mean: 0.724 | Chlorophyll: 42.3 µg/cm² | LAI: 3.81
→ Plant Height: 48.2cm | Branch Count: 7 | Leaf Area: 234cm²
[SUCCESS] Report exported: wheat_001_report.csv, wheat_001_plots.pdf
Interactive Visualisation

See Plants in a New Dimension

Fuse 3D geometry with spectral imaging to reveal invisible plant traits.

Live Preview — Wheat_Scan_001
Spectral Bands
Active Band
NDVI
Vegetation index
X: 0.00 Y: 0.00 Z: 0.00
156,892 pts
Low
High
NDVI Index
360° Rotation
Freely rotate and zoom around the full 3D plant model
Per-Point Spectra
Click any point to inspect its full spectral signature
Time-Series Playback
Replay plant growth across multiple scan timepoints
Technology Stack

Built with Scientific Rigour

A modern scientific Python stack powering reproducible plant phenomics research.

Core Application
🐍
Python
Primary language for backend processing and pipeline orchestration
📐
NumPy / SciPy
Scientific computing and numerical processing of spectral arrays
👁️
OpenCV
Computer vision, image preprocessing, and background segmentation
3D Processing
🔷
Open3D
3D point cloud processing, visualisation, and mesh reconstruction
🌐
VTK / PyVista
High-performance 3D rendering and scientific visualisation
☁️
PCL (via Python)
Point cloud filtering, segmentation, and feature extraction
Hyperspectral Analysis
🌈
SPy (SpectralPy)
Hyperspectral image I/O, calibration, and band manipulation
🤖
scikit-learn
Machine learning for spectral classification and trait regression
📊
Matplotlib / Plotly
Scientific plotting and interactive spectral visualisations
Desktop Application
🪟
Qt / PyQt6
Cross-platform desktop GUI with native look and feel
🐼
Pandas
Structured data handling for trait tables and export pipelines
💾
HDF5 / Zarr
Efficient storage of large hyperspectral datasets
📖
Open Science Principles
PhenoFusion3D is built on open-source foundations with clear documentation and reproducible pipelines, designed for adoption and extension by the plant science community.
Organisation & Mission

Supporting Plant Science at Scale

About APPN

Australian Plant Phenomics Network

The APPN provides a national network of plant phenotyping solutions and sensors, backed by specialist technical and data expertise, to enable researchers to measure the performance of plants efficiently, objectively, and non-destructively over time.

At the ANU node, we specialise in controlled environment phenotyping and developing advanced software to support plant phenomics research at national scale.

Government
Academic Research
Start-Up / Entrepreneurial
Not-for-Profit
Project Goals

Intended Outcomes

  • ✓A usable desktop application for end-to-end plant scan processing
  • ✓Interactive 3D visualisation with hyperspectral overlays
  • ✓Exportable outputs (tables, plots, reports) for research workflows
  • ✓Clear documentation and reproducible pipelines for future adoption
🌾
Crop Breeding
Screen thousands of genotypes for stress tolerance, yield potential, and quality traits in controlled environments.
🧬
Genetic Research
Correlate spatial and spectral phenotypes with QTL markers to accelerate genotype-to-phenotype discovery.
🌡️
Climate Adaptation
Quantify plant responses to heat, drought, and elevated CO₂ for developing climate-resilient crops.
🏭
Agritech Integration
Exportable data pipelines that integrate with existing breeding management and decision support systems.
Meet the Team

Built by Researchers & Students

PhenoFusion3D is supported by APPN ANU and developed collaboratively with guidance from the project leadership team and contributions from student members at ANU.

Saswat Panda
⭐
Team Lead

Saswat Panda

Saswat.panda@anu.edu.au

Saswat Panda leads the PhenoFusion3D project at APPN ANU, guiding the overall project direction and supporting the team's technical and research goals.

Project Leadership
SS
Supriyo Shafkat Ahmed
Data Lead

Supriyo supports the project as Data Lead, contributing to data-related planning, organisation, and technical guidance for the team.

SUPRIYOSHAFKAT.AHMED@anu.edu.au
Student Team Members · ANU
AR
Adithya Rama
Student at ANU
Adithya.Rama@anu.edu.au
TS
Tanisha Sharma
Student at ANU
Tanisha.Sharma@anu.edu.au
HZ
Howard Zhang
Student at ANU
u7877905@anu.edu.au
TX
Tianyu Xu
Student at ANU
Tianyu.Xu@anu.edu.au
Get Involved

Advance Plant Science
Together

PhenoFusion3D is developed in collaboration with the APPN at ANU. Reach out to learn more about the project, dataset access, or collaboration opportunities.

Contact APPN ANU NodeEmail Project Lead
Location
ANU Node, Canberra
Research School of Biology
Project Lead
Saswat.panda@anu.edu.au
Direct contact for project enquiries
Open Source
Reproducible Pipelines
Documentation & codebase
PhenoFusion3D
© 2026 PhenoFusion3D · Developed at the ANU Node of the Australian Plant Phenomics Network
APPN WebsiteANU