Frontveg

A plugin for foreground vegetation segmentation, tailored for trellised vegetation row images. It uses RGB images to perform inference and allows users to manually refine the generated mask.

---

The method was developped by Herearii Metuarea, PHENET PhD at LARIS (French laboratory located in Angers, France) and Abdoul Djalil Ousseini Hamza, AgroEcoPhen Engineer at IRHS (French Institute located in INRAe Angers, France) in Imhorphen team (bioimaging research group lead) under the supervision of Eric Duchêne (Research Engineer), Morgane Roth (Research Engineer) and David Rousseau (Full professor). This plugin was written by Herearii Metuarea and was designed in the context of the european project PHENET.

---

This napari plugin was generated with copier using the napari-plugin-template.

Installation¶

You can install frontveg via pip:

pip install frontveg

To install latest development version :

pip install git+https://github.com/hereariim/frontveg.git

GPU is mandatory for time processing and models running (especially Grounding-DINO). Please visit the official PyTorch website to get the appropriate installation command: 👉 https://pytorch.org/get-started/locally

Exemple : GPU (CUDA 12.1)

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

Description¶

This plugin is a tool to perform image inference. This plugin contained two steps of image processing. First, from RGB image, a depth map is estimated and then thresholded based on the estimated depth histogram modes to detect foreground and background regions in image. Second, a Grounding DINO model detects foliage in the foreground. The output is a binary mask where white colour are associated to foliage in the foreground.

The plugin is applicable to images of trellised plants; in this configuration, it has been applied to images of pome fruit trees (apple), stone fruit trees (apricot) and climbing plants (grapevine).

Contact¶

Imhorphen team, bioimaging research group

42 rue George Morel, Angers, France

• Pr David Rousseau, david.rousseau@univ-angers.fr
• Abdoul Djalil Ousseini Hamza, abdoul-djalil.ousseini-hamza@inrae.fr
• Herearii Metuarea, herearii.metuarea@univ-angers.fr

Contributing¶

Contributions are very welcome. Tests can be run with tox, please ensure the coverage at least stays the same before you submit a pull request.

License¶

Distributed under the terms of the BSD-3 license, "frontveg" is free and open source software

Issues¶

If you encounter any problems, please file an issue along with a detailed description.