[DF Scanner New Feature] AI-based broadleaf tree crown segmentation function (beta version) now available

DeepForest Technologies Co., Ltd. (hereinafter referred to as DeepForest) has released a new “AI Crown Segmentation (beta version)” function for its forest analysis software DF Scanner. This function is available from DF Scanner Pro v1.6.0, released on September 30, 2025.

 

What is the AI ​​Crown Segmentation?

Much research has been conducted worldwide on broadleaf tree crown segmentation, but due to the variation in shape and size of broadleaf trees, automatic segmentation has been an extremely difficult challenge.
Now, by utilizing AI (deep learning) and training a wide range of training data, DeepForest has successfully developed a function that automatically separates broadleaf forest crowns.
This function will be included in the Pro version of DF Scanner as the “AI Crown Segmentation”.

No specialized knowledge is required to use it; AI Crown Segmentation can be performed simply by specifying an orthomosaic taken by a drone. By combining it with DF Scanner’s tree species identification and size estimation functions, it becomes possible to perform efficient forest analysis with simple operations, even for broadleaf forests, which have previously been difficult to analyze automatically.


* This function is not available with DF Scanner Lite.
* The orthomosaic used must conform to our recommended specifications with a resolution of approximately 2.5 cm.
* It can also be used for coniferous and mixed forests, but accuracy may be lower.
* To use this function, an additional AI model must be installed.
* We recommend using a computer with a GPU. Analysis is possible on computers without a GPU, but processing will take longer.

DF Scanner offers a two-week free trial license.
You can also try out the AI ​​Crown Segmentation function here.

Free trial available here

Download the software here

 

As of September 30, 2025, this feature will be in beta, and we will continue to update the model to improve accuracy. Currently, we are creating training data mainly for broad-leaved trees, but in the future we plan to add data for conifers as well, so that it can handle a wider range of data.
If you have any questions about this feature or would like to share your feedback after using it, please feel free to contact us using the inquiry form.