Comparison Between Boundary Color Method and Haar Cascade Classifier Applied on Sheep Detection and Quantification on a Live Video Streaming for a Quadcopter UAV

This paper compares two methods applied to a system for detecting and quantifying sheep in live video frames captured by unmanned aerial vehicles (UAVs). The proposed system consists of a UAV equipped with a wireless streaming video system that transmits video frames to a base station. These video frames are then processed using two digital image processing approaches: the boundary color method and the Haar cascade classifier. Both methods are implemented in Python 3 with the OpenCV library and employ morphological operations and feature-classifying algorithms to detect patterns and objects in the video source. Both methods have proven effective in detecting and quantifying sheep images under various lighting conditions and at different distances between the UAV and the sheep. However, the boundary color method demonstrates greater robustness in outdoor environments.