Surface Water Pollution Detection using Internet of Things

• Main Authors: Shafi, Uferah, Mumtaz, Rafia, Anwar, Hirra, Qamar, Ali Mustafa, Khurshid, Hamza
• Format: Conference Proceeding
• Language: English
• Subjects: Deep Learning; embedded systems; Internet of Things; Water quality monitoring
• ISSN: 1949-4106
• DOI: 10.1109/HONET.2018.8551341

Water is one of the primary requisites and crucial for sustaining the quality of life. In Pakistan its significance is more than ordinary due to the agrarian nature of the economy. Owing to increasing trend in urbanization and industrialization, the quality of water is continuously declining. For this purpose, we propose an Internet of Things (IoT) based water quality system capable of measuring the quality of water in near real time. The proposed solution is based on World Health Organization (WHO) defined water quality metrics. For this purpose, a real time embedded prototype has been developed to record the water quality parameters from the water samples collected from various sources across the study area. The hardware solution sends data to cloud for real time storage and processing. The processed data can be remotely monitored and water flow can be controlled using our developed software solution comprising of mobile app and a dashboard. In addition to water quality monitoring and control system, the predictive analysis of the collected data has been performed. For training purposes a dataset has been obtained from Pakistan Council of Research in Water Resources (PCRWR). Machine learning algorithms have been applied for classification of water quality and the experimental results indicate that deep neural network outperforms all other algorithms with an accuracy of 93%. The preliminary results have shown a high potential of scaling up this concept to an advanced level.