Automatic Detection of Spatiotemporal Urban Expansion Patterns by Fusing OSM and Landsat Data in Kathmandu

Abstract

During the last few decades, a large number of people have migrated to Kathmandu city from all parts of Nepal, resulting in rapid expansion of the city. The unplanned and accelerated growth is causing many environmental and population management issues. To manage urban growth efficiently, the city authorities need a means to be able to monitor urban expansion regularly. In this study, we introduced a novel approach to automatically detect urban expansion by leveraging state-of-the-art cloud computing technologies using the Google Earth Engine (GEE) platform. We proposed a new index named Normalized Difference and Distance Built-up Index (NDDBI) for identifying built-up areas by combining the LandSat-derived vegetation index with distances from the nearest roads and buildings analysed from OpenStreetMap (OSM). We also focused on logical consistencies of land-cover change to remove unreasonable transitions supported by the repeat photography. Our analysis of the historical urban growth patterns between 2000 and 2018 shows that the settlement areas were increased from 63.68 sq km in 2000 to 148.53 sq km in 2018. The overall accuracy of mapping the newly-built areas of urban expansion was 94.33%. We have demonstrated that the methodology and data generated in the study can be replicated to easily map built-up areas and support quicker and more efficient land management and land-use planning in rapidly growing cities worldwide.

Finding Coastal Megaclast Deposits: A Virtual Perspective

“…The development of the Google Earth Engine has permitted to pose the question of the invention of advanced digital tools for automatic detection of particular features on satellite images. For instance, such tools are helpful in urban studies [111] and archaeology [112]. In geomorphological studies, two successfully tested approaches are notable.…”

An Improved Boundary-Aware Perceptual Loss for Building Extraction from VHR Images

“…For memory efficiency, we set the input batch size to 4, and the popular poly learning rate schedule, as shown in Equation 10, is also applied to adjust the learning weight. lr = lr init 1 − iter max_iter power (10) where iter and max_iter represent the current and total epoch, respectively, and power is set to 0.95 in our experiments. The pipeline network UNet-Res34 reaches 88.94% and 98.71% on IoU and Acc.…”

“…Similar to the semantic segmentation task, building extraction is also a low-level pixel-wise labelling task aiming to classify each pixel into a building/no building class. It is the foundation for high-level tasks such as city planning [4,9] and population evaluation [10]. For pixel-wise labelling tasks, the fully convolutional network (FCN) [11] is the most popular and classical deep learning model; however, the boundary areas of the results predicted by the FCN model are always inaccurate and blurred.…”

Spatiotemporal Modeling of Urban Growth Using Machine Learning

“…When diverse and skilled human capital interacts, cities can boost ecosystems of research, development, innovation, and entrepreneurship [15][16][17]. It is this duality between the challenges and benefits of cities that makes the understanding of their growth so important for us.To predict how cities are going to look like in the future, urban planners rely on historical data and urban growth models [10,[18][19][20] . These models can have multiple outputs, but they commonly provide estimations of binary urban footprints (i.e., maps with urban or non-urban categories) or population distribution within a region of study over time [21].…”

“…To predict how cities are going to look like in the future, urban planners rely on historical data and urban growth models [10,[18][19][20] . These models can have multiple outputs, but they commonly provide estimations of binary urban footprints (i.e., maps with urban or non-urban categories) or population distribution within a region of study over time [21].…”