Abstract
AbstractWater, as an important resource, is critical to sensible development. For the benefits of the climate and humanity, the scenario with water assets needs to be assessed precisely. However, it is miles exceptionally troublesome, tedious, and quite hard to assess a massive water area using conventional ground survey strategies. Remote sensing for quite a while has been applied in water asset evaluation. The present-day techniques depending on water indices, viz., the water ratio index (WRI), the normalized difference vegetation index, the normalized difference water index, the modified normalized difference water index are typically carried out for water resource assessment, which delineates the waterbody through utilizing predefined threshold value. Generally, threshold selection is difficult due to the presence of environmental noise, which includes vegetation, shadow, cloud, and built-up land, prompting misclassification. In this experimental study, a new technique for water body extraction using object-based image analysis (OBIA) and deep learning approach is put forward to lessen the errors. The trial outcomes display that for the proposed strategies, overall mapping accuracy is better than the ones made using the water indices.KeywordsWaterbody extractionIndex methodXGBoost
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have