I’ve recently been testing an AI segmentation plugin developed by TerraLab for QGIS, and I have to say — it’s seriously impressive.

It’s essentially a one-click segmentation tool for drone and satellite imagery. You load your orthophoto (and it also works with DEMs, DSMs and DTMs), run the tool, and within seconds your dataset is segmented and ready for further analysis. No lengthy training process, no complicated setup — just clean, usable outputs that you can immediately start classifying or refining.

For those of us working in conservation and environmental monitoring, especially with high-resolution drone data, segmentation is often the bottleneck. This speeds that step up dramatically.

I’m curious how others here are handling segmentation workflows in QGIS — are you using traditional OBIA approaches, machine learning classifiers, or something similar?

I’ll include a short YouTube demo below for anyone who wants to see it in action.

Would be great to hear thoughts from others who’ve tested it, especially in conservation applications (alien vegetation, habitat mapping, species detection, etc.).

Long form video showing how this tool can be used in tandem with machine learning for large scale vegetation surveying see here: