Segmentation of image and video streams to detect unhealthy trees from drone footage

This project focuses on the detection of unhealthy trees in drone footage through image and video segmentation. The dataset consists of keyframes extracted from drone videos and annotated for unhealthy trees based on verbal description. Two models were trained: Logistic Regression and Convolutional Neural Network(CNN). The models demonstrate their efficiency in predicting and segmenting, contributing to the automated detection of tree health issues in environmental applications.