Options
Semantic Segmentation of PolSAR Images for Various Land Cover Features
Journal
International Geoscience and Remote Sensing Symposium (IGARSS)
Date Issued
2021-01-01
Author(s)
Kotru, Rahul
Shaikh, Musab
Turkar, Varsha
Simu, Shreyas
Banerjee, Satyaswarup
Singh, Gulab
Abstract
Land-cover classification is one of the core applications in the field of remote sensing. It is a valuable resource for city planners to achieve sustainable development. Many metropolitan cities are experiencing disorganized growth with a high intensity of urban sprawl due to the economic pull and better standards of living offered in metropolitan cities when compared to the surrounding rural areas. If this pattern of growth continues, it will lead to unsustainable development. This leads to an increase in pressure on urban infrastructure and the ecosystem. The traditional methods which are used for urban mapping are time consuming. Instead, Microwave Remote Sensing can be used to acquire geographical data which can be used to develop a decision support system to help urban settlement planners. This paper suggests the use of Semantic Segmentation to extract the various land cover features from Polarimetric Synthetic Aperture Radar (PolSAR) images using the Fully Convolutional Network (FCN) based modified UNet architecture, that will help in the analysis of land-cover in areas prone to urban sprawl by utilizing the elements of coherency matrix.
Subjects