Cloud-Free Mosaic Method for Worldview -3 Images Using Multitemporal Image Analysis and Deep Learning
Heri Y. Sulyantara, Randy P. Brahmantara, Kurnia Ulfa, Danang S. Candra, Ferman S. Nugroho, Kuncoro A. Pradono, Marendra E. Budiono, Kiki W. Veronica, Orbita Roswintiarti, Sastra K. Wijaya, Abdul Haris
Abstract:
This study presents a novel methodology for generating high-quality mosaics from WorldView-3 satellite imagery, focusing on minimizing cloud cover. The study site, which is located in East Jakarta and its surrounding areas, was chosen due to its diverse land cover types. The proposed methodology involves developing a database to store and manage metadata from multiple WorldView-3 scenes, allowing for an efficient selection of cloud-free image tiles. The algorithm identifies optimal tiles based on cloud cover and acquisition date by dividing the image into several smaller tiles and using a sliding window approach. The selected tiles are then stitched together to create a seamless mosaic. The results demonstrate the effectiveness of the proposed method in generating high-quality mosaics with minimal cloud contamination, providing a valuable resource for applications such as urban planning, environmental monitoring, and disaster management. This study contributes to the advancement of remote sensing image processing techniques, particularly in the context of high-resolution satellite imagery.
Keywords:
Worldview-3, mosaic image, CNN.