Affiliation:
1. Maulana Azad National Institute of Technology, India
Abstract
This chapter focusses on adaptive convolutional neural network (ACNN) classifier, an innovative approach that tackles the challenges of constrained time, cost, and lacking accuracy faced by the conventional classification methods. Preservation of disappearing ponds and streams is crucial for sustainable water management, where remote-sensing analysis plays a pivotal role. Deep learning (DL) emerges as a potent solution, addressing these limitations effectively. ACNN based on DL is validated using the GaoFen image dataset (GID) and employs adaptive normalization for precise classification of small water bodies among larger ones. ACNN ensures rapid training without compromising the accuracy. The computation and assessment of water quality index (WQI) guides specific actions, enabling the water resource management to care for acceptable water areas, fostering biodiversity and fisheries. Conversely, areas falling outside these standards could be treated by Regional Municipal Corporation, facilitating ecological balance, pollution control, flood prevention, and ensuring public water supply.