A PyTorch domain library providing datasets, samplers, transforms, and pre-trained models for geospatial and remote sensing data.
4.2K+8 this week607PythonMITWell maintainedUpdated 2 days ago
weecology/DeepForest
Python package for detecting ecological objects such as tree crowns and birds in airborne RGB imagery, built on torchvision.
- Stars
- 788
- Forks
- 275
- Open issues
- 107
- Last commit
- 9 days ago
What is it?
DeepForest is a Python package for training and running object detection models on airborne imagery, aimed at ecological monitoring. It ships with a pretrained tree crown detection model and a bird detection model, both single-class and designed to be extended to species classification with new data.
The package uses deep learning object detection networks from torchvision to predict bounding boxes around individual trees in RGB imagery. It is meant to simplify retraining detection models for different forests, sensors and spatial resolutions, so users can annotate their own data and train custom models rather than starting from scratch.
The project is MIT licensed and is documented on Read the Docs. It is published on PyPI and requires Python 3.10 or newer.
Key features
- 01Pretrained tree crown detection model for RGB imagery.
- 02Pretrained bird detection model for high-resolution airborne imagery.
- 03Supports training custom detection models from user-annotated data.
- 04Built on torchvision's object detection modules.
- 05Installable from PyPI.
Use cases
- Delineating individual tree crowns in forest survey imagery.
- Detecting birds in high-resolution airborne photos for wildlife surveys.
- Retraining a detection model for a new forest type or sensor.
In collections
Topics
FAQ
- What is DeepForest?
- Python package for detecting ecological objects such as tree crowns and birds in airborne RGB imagery, built on torchvision. DeepForest is a Python package for training and running object detection models on airborne imagery, aimed at ecological monitoring. It ships with a pretrained tree crown detection model and a bird detection model, both single-class and designed to be extended to species classification with new data.
- Is DeepForest free to use? What license does it have?
- DeepForest is open source under the MIT license. Check the license terms before using it in commercial products.
- Is DeepForest still maintained?
- Actively maintained. The last commit to DeepForest was on Oct 28, 2026.
- What can I use DeepForest for?
- Delineating individual tree crowns in forest survey imagery. Detecting birds in high-resolution airborne photos for wildlife surveys. Retraining a detection model for a new forest type or sensor.
- Where can I get DeepForest?
- The source code and installation instructions are on GitHub: https://github.com/weecology/DeepForest
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Ultralytics YOLO is a Python library for computer vision covering detection, segmentation, classification, pose estimation, tracking and more.
63.2K+188 this week12KPythonAGPL-3.0Updated 7 hours ago
Ultralytics YOLOv5, a PyTorch-based computer vision toolkit for object detection, instance segmentation and image classification with training and export.
58.3K+10 this week17.4KPythonAGPL-3.0Well maintainedUpdated 3 days ago
A PyTorch-based toolkit from Apple for training small and large deep neural networks, including CLIP, LLMs, classification, detection and segmentation models.
7K+1 this week542Jupyter NotebookOtherUpdated last month
A Python library for sliced inference that runs object detection models over tiles of large images to find small objects.
AI & Machine Learning#computer-vision#object-detection#small-object-detection#instance-segmentation#tiling
5.6K+11 this week794PythonMITWell maintainedUpdated 6 days ago
Python library for image augmentation in computer vision, covering pixel, spatial, medical and 3D transforms with consistent handling of masks, boxes and keypoints.
AI & Machine Learning#computer-vision#data-augmentation#image-processing#deep-learning#object-detection
591+9 this week45PythonAGPL-3.0Well maintainedUpdated 12 hours ago