Drone-Based Sheep Recognition and Counting: A Comparative Study of Color-Shape, and YOLOv5 Algorithms

Ticona J.R.C.; Chipana G.J.F.; Cabrera G.H.C.; Chambilla T.R.B.; Ponce G.R.V. · Smart Innovation, Systems and Technologies (2025) · DOI: 10.1007/978-3-031-92651-8_23

Resumen

This study presents sheep detection and quantification methods for video frames captured by UAVs and examinate their advantages over traditional UAV video recording. The implementation of two digital image processing approaches for pattern recognition is based on color-shape analysis and the YOLOv5 classifier to detect sheep in a field from a video source. Both methods are implemented in Python 3 using OpenCV tools, where morphological operations and feature classification algorithms are applied. These methods perform multiple operations on the video source and have proven to be effective for detecting sheep images under different lighting conditions, as well as at varying distances between the UAV and the sheep. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Instituciones: Universidad Nacional del Altiplano