A Comparative Investigation of Supervised Classification Approaches to Mapping Flooded areal extent in and around Puzhal Lake, Chennai, using image processing

Flooding is a natural occurrence that can result in detrimental effects on people’s mental health as well as their financial well-being. Flood inundation maps not only assist in the process of urban planning but also assist in the localization of regions that are likely to be flooded in the event of future floods. In the event of a catastrophe, when it is not possible to collect ground data using traditional methods, satellite photos are the method of choice for collecting real-time imagery. The technique and classification methods used for the development of flood maps in and around Puzhal reservoir, which is located close to Chennai, Tamil Nadu, in the Indian subcontinent, are discussed in this study. Puzhal reservoir is in Tamil Nadu. The floods that struck the southern part of the state in November and December 2015 were caused by century-high rainfall records, and the maps that have been generated are based on satellite data that was recovered both before and after the floods. Both the pre-flood and post-flood image data were processed with three supervised classification methods, namely the parallelepiped classifier, the minimum distance classifier, and the maximum likelihood classifier, and the error in the user selection of features was determined with the help of a confusion matrix. These methods were used to analyse the data obtained from the images taken before and after the flood. In addition, the data were checked for accuracy by utilising statistical data that was produced from the values of indices such as the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Water Index (NDWI). The maximum accuracy of the selected region was obtained using the parallelepiped classifier for the data gathered before the flood, while the greatest accuracy was obtained using the maximum likelihood classifier for the data gathered after the flood. It was discovered that there has been a considerable increase in the cover of green vegetation and the extent of water in the post-flood image in compared to the pre-flood complement. This was noticed while looking at the image in terms of areal extent.