This dataset supports individual tree-crown object detection from high-resolution geospatial image chips in agroforestry and agricultural settings in India. Each release includes GeoTIFF imagery, clean COCO-format annotations for a single category (tree), image-level metadata and manifests, and multiple train/evaluation split definitions—including a fixed train / in-distribution test partition and a non–distribution-shift protocol for geographic robustness. The split metadata includes state, region, and zone fields to support geographic evaluation beyond random image splits.
This dataset supports individual tree-crown object detection from high-resolution geospatial image chips. It includes GeoTIFF imagery, COCO-format annotations, metadata, split files, and inspection samples.
The dataset is intended for research on individual tree detection, geographic generalization, agroforestry monitoring, and evaluation of remote-sensing models.
geotiffs/: image chips.annotations/: COCO-format tree-crown annotations and annotation summaries.metadata/: image-level metadata and release manifests.splits/: benchmark split definitions.samples/: small inspection sample.Performance may vary across regions, acquisition conditions, sensors, tree species, crop backgrounds, and crown sizes. Split metadata is provided so users can evaluate geographic robustness rather than relying only on random image-level performance.
Final public license text and imagery-provider redistribution terms should be confirmed before public release.