Thinning is a critical component of radiata pine forest management, improving stand growth, stability, and overall forest health. As the forestry sector continues to seek safer, more efficient, and cost‑effective approaches to thinning — particularly on steep or difficult terrain — chemical thinning (eThinning) has gained increasing attention.
Recent research and operational trials have highlighted eThinning as a potential alternative to manual thinning, reducing exposure to safety risks while maintaining silvicultural outcomes. However, evaluating the success of chemical thinning at scale presents its own challenges — particularly when it comes to rapid, objective stocking assessment.
What’s New in TreeTools Virtual Plot V2.18
With the latest release of TreeTools.ai Virtual Plot, Interpine has introduced an ALPHA release of our Post‑Chemical Thinning AI detection model, designed to support rapid stocking assessment using drone imagery.
This new model focuses on:
- Detecting remaining crop trees
- Differentiating crop from chemically treated trees showing signs of dieback or stress
- Quantifying post‑treatment stocking to help evaluate thinning outcomes
By using drone‑acquired imagery, the model enables assessment at scale while minimising time on the ground.
How the eThinning AI Model Works
Following a chemical thinning operation, treated stems typically exhibit signs of stress or dieback over time. The ALPHA eThinning model leverages these differences in tree condition to:
- Identify and count surviving crop trees
- Exclude treated stems from stocking totals, based on a visual assessment of the canopy condition. This being based on the discolouration of the tree canopy, from the increasing stress trees post treatment.
- Provide a clear, repeatable indication of thinning effectiveness and consistency.
This approach allows forest managers and advisers to assess success earlier and more efficiently, supporting informed follow‑up decisions.
ALPHA Release – What This Means
This model is being released as an ALPHA feature within Virtual Plot. That means:
- It is functional and available to users
- Performance will vary by forest type, treatment timing, and imagery quality
- Results should be reviewed alongside professional judgement
Virtual Plot has in inbuilt refine tool, which enables users to refine AI results by editing the outcomes and ensuring a final precise and accurate inventory, even as we continue to development the AI implementation of this new model.
Importantly, this release is intended to support collaborative development, enabling the model to improve as more real‑world data is tested and validated.
A Collaborative Effort with Industry and Research
The development of this capability forms part of Interpine’s ongoing work with NZ Forest Growers Research, through the Precision Silviculture research programme, which is funded through the MBIE programme grant No C04X2101.
Recent studies by Bioeconomy Science Institute (Main, R. et al 2025) indicate clear discolouration developing from Day 25 post thinning using normal RGB drone-based imagery, such as that collected using the DJI Matrice 4 Enterprise. This showing that reliable detection is possible within the first few weeks after treatment, creating opportunities to use UAV imagery for early validation of chemical thinning, particularly in steep, remote and hazardous terrain where ground checks are often costly and difficult.
We also acknowledge the many forest owners who have already contributed data to support the development of this initial ALPHA release of the Chemical Thinning (eThinning) AI model. Their early involvement has been critical in helping ground the model in real‑world forestry conditions, which was a key next step and finding from the Technical Paper above.
We continue to inviting users to trial the eThinning AI model and provide feedback and new imagery, with this we can:
- Improve detection accuracy across different sites and conditions
- Refine interpretation of stress and dieback signatures
- Ensure the model delivers practical value for operational forestry
This collaborative approach ensures innovation is grounded in real forests, not just technical theory.
Get Involved, Share Feedback, Insights, and Example Datasets
We encourage Virtual Plot users actively involved in chemical thinning operations to:
- Trial the eThinning AI model within Virtual Plot
- Compare AI‑derived stocking results with field observations
- Share feedback, insights, and example datasets
Your involvement directly contributes to improving the model and advancing precision silviculture practices across the sector.
Some foresters, as part of early eThinning adoption, have established mixed trial areas combining traditional thin‑to‑waste and chemical thinning within the same stand. The TreeTools Virtual Plot Chemical Thinning ALPHA model is designed to work across both thinning methods, enabling consistent stocking assessment and comparison within these mixed‑treatment environments.
Get Started, Download and Start Using TreeTools Virtual Plot
The introduction of chemical thinning analytics is another step toward integrated, data‑driven forest management, where aerial data, AI, and silvicultural expertise come together to support better decisions.
As the model matures beyond ALPHA, we’ll continue to share updates, improvements, and guidance with our users.
More info on background research projects and studies:
Main, R. et al 2025. Early Detection of Herbicide-Induced Tree Stress Using UAV-Based Multi-spectral and Hyperspectral Imagery. Forests, 16 (8), and Main R, et al. 2025. Remote Sensing of Chemical Thinning, Technical Paper No PSP-T036 FGR.