A Framework of image processing and machine learning utilization for flood disaster management

Abstract

Flood is one of the annual disasters in many places. It has not been well-managed yet both pre-disaster and post-disaster. Image processing and machine learning are commonly utilized for disaster management systems such as forecasting any potential flood by monitoring the water level in rivers and dams. However, it has a limited framework to be implemented as a strategic plan in flood management. Thus, this study aims to develop a framework for image processing and machine learning utilization for flood management. This study involves Padang, West Sumatera, Indonesia as a sample. It was conducted in three stages; 1) categorize the strategic plans and policies; 2) gather relevant literature; 3) analyze data. As findings, this study proposes a framework consisting of enhanced disaster preparedness, improved coping capacity, and completion of post-disaster reconstruction and rehabilitation. Involvement of the government, researchers and industry are mandatory. Government and researchers should collaborate to establish policies and regulations. Researchers should conduct studies with financial support from the industry. Meanwhile, the industry should be a public-private partnership with the government. In addition, the involvement of the private sector and the government are important factors that must exist to support research in this field.