Overview
This dataset contains 321,562 images from camera traps in Nkhotakota Wildlife Reserve in Malawi. The reserve operates under a partnership between African Parks Network and the Malawi Department of National Parks and Wildlife (DNPW). The wildlife monitoring program is implemented through technical cooperation with USDA Forest Service.>
Labels are provided for 46 taxa along with counts of animals per image, and a subset of images (33,813) also have manually drawn bounding box annotations. Labels are provided at the image level and not by sequence. The dataset includes 164,139 empty photos from camera false triggers, containing no animals in them. Images of humans have been removed from this dataset but their entries are still present in the metadata.
Classes comprise a mix of species and higher taxonomic groups. For example, small antelopes were labeled to species when possible (e.g., common duiker, red duiker, klipspringer, Sharpe’s grysbok) but as “small antelope” otherwise. Species-level identification was not attempted for some taxa (galagos, genets, mongooses, squirrels). The most common labels are yellow baboon (54,119 images), cape bushbuck (17,062 images), bushpig (14,733 images), small antelope (13,289), and elephant (10,664 images).
These images were used to train and evaluate a YOLOv4 multiclass detector model. Splits (train, validation, and test) are noted in the metadata.
Citation, license, and contact information
If you use this dataset, please cite:
Appel CL, Subramanian A, Koning JS, Ngosi M, Sullivan CM, Levi T, Lesmeister DB. Developing custom computer vision models with Njobvu-AI: A collaborative, user-friendly platform for ecological research. Ecological Applications, 2025.
For questions about this data set, contact Cara Appel.
This data set is released under the Community Data License Agreement (permissive variant).
Data format
Annotations (including species tags and unique location identifiers) are provided in COCO Camera Traps format.
For information about mapping this dataset’s categories to a common taxonomy, see this page.
Downloading the data
Metadata is available here.