Camera Trap Data Analysis for Grevy's Zebra Monitoring

Camera trapping is the most widely adopted method of monitoring wildlife since it is a non-invasive and cost effective method of monitoring wildlife remotely. By analysing camera trap images, conservationists can better understand wildlife behaviour and species and population dynamics. Due to the vast amount of data generated by camera traps, manual selection of images with species of interest proves tedious. This study leverages object detection and image classification models to automate the processing of camera trap images to monitor the endangered Grevy's zebra, first distinguishing zebras from other animal species and then differentiating between two closely related zebra species: the endangered Grevy's zebra and the more common Plains zebra. The presented results demonstrate that AI-assisted camera trap analysis can significantly enhance the efficiency of endangered species monitoring, allowing conservationists to focus on higher-level ecological analysis and decision-making. The potential implications of this approach for large-scale species monitoring are discussed and future improvements to AI-based wildlife classification systems suggested.