Geo Trax

--- license: cc-by-4.0 base_model: - Ultralytics/YOLOv8 base_model_relation: finetune pipeline_tag: object-detection library_name: ultralytics num_parameters: 11137922 github: https://github.com/rfonod/geo-trax language: - en tags: - ultralytics - yolov8 - object-detection - aerial-imagery - drone - vehicle-detection - birds-eye-view - geo-trax - trajectory - urban-traffic - tracking - arxiv:2411.02136 datasets: - rfonod/songdo-vision - Voxel51/VisDrone2019-DET - detection-datasets/coco model-index: - name: rfonod/geo-trax/geotrax_hbb_yolov8s_1920_v1 results: - task: type: object-detection dataset: type: rfonod/songdo-vision name: Songdo Vision (v1) revision: v1 split: test metrics: - type: precision # mAP@0.5 not available as a standard metric type on HF value: 0.951 name: mAP@0.5 - type: precision value: 0.711 name: mAP@0.5:0.95 - type: precision value: 0.911 name: Precision - type: recall value: 0.935 name: Recall --- # Geo-trax: YOLOv8s Vehicle Detector for Drone BEV Imagery [](https://github.com/rfonod/geo-trax) [](https://pypi.org/project/geo-trax/) [](https://huggingface.co/spaces/rfonod/geo-trax) [](LICENSE) [](https://doi.org/10.1016/j.trc.2025.105205) [](https://arxiv.org/abs/2411.02136) [](https://huggingface.co/datasets/rfonod/songdo-vision) [](https://huggingface.co/datasets/rfonod/songdo-traffic) [](https://www.real-lab.ch/geo-trax) [](https://youtu.be/gOGivL9FFLk) This is the default detection model for **[Geo-trax](https://github.com/rfonod/geo-trax)**, a comprehensive pipeline for extracting georeferenced vehicle trajectories from high-altitude drone (bird's-eye view) video footage....