Automated Leak Detection in Drip Irrigation Systems using RGB and Thermal Sensor Fusion

Water leaks are a common issue in surface drip irrigation systems. Visual inspection of irrigation pipelines by humans is the most prevalent method for leak detection. However, this approach is costly and labour-intensive due to the need for frequent on-site visits. This paper describes an AI based sensor fusion algorithm to automatically detect leaks along the drip lines using RGB and thermal images collected from a low-cost ground vision system. The proposed algorithm was tested using images collected from vineyard under various light conditions. Results indicated that proposed sensor fusion detection algorithm is accurate and efficient.