Water Quality Sensor Noise Classification based on Drone Movement using Machine Learning

Water pollution caused by human activities and environmental changes has become a critical issue. The challenges of directly analyzing water quality in remote or inaccessible aquatic environments highlight the importance of combining drones and water quality sensors. However, external factors, such as wind-induced drone movements, can introduce noise into sensor readings, compromising the accuracy of data collection. This study focuses on addressing the patterns of drone movements and their impact on sensor noise to enhance the reliability of water quality monitoring. In this research, experimental results demonstrate that machine learning models effectively analyzed the movement patterns, with detailed performance metrics for each algorithm. XGBoost (extreme gradient boosting), Random Forest, Decision Tree, and Logistic Regression were among the methods that were assessed. The Random Forest model outperformed the others, including XGBoost, Decision Tree, and Logistic Regression, with an accuracy of 97%, compared to 96% and 85%, respectively. The findings identify Random Forest as the most effective algorithm for analyzing movement patterns and mitigating noise in sensor data, thereby improving the overall reliability of water quality assessments.