Raster Big Data Processing Using Spark with GeoTrellis

The accessibility of the Internet and Geographic Information Systems (GIS) has significantly affected how people live their daily lives. These systems use various GIS data, which can be categorized into two types: raster data and vector data. When working with continuous data types such as elevation and satellite imagery, raster data proves to be more useful. However, existing methods for processing such data involve converting raster data to vector data, resulting in a loss of details, decreased accuracy, and alterations to the original data. This research focuses on finding ways to improve the processing and analysis of large raster datasets without compromising precision or causing significant loss. In the proposed solution, the capabilities of GeoTrellis and Docker are utilized. GeoTrellis, with its support for various raster data storage formats and ability to process data without conversion, is well-suited for handling large-scale data processing. By using Docker to distribute computation across a cluster, the solution is able to efficiently process complex tasks within a reasonable timeframe. Additionally, a collection of algorithms is offered within the proposed system architecture to handle data. The effectiveness of processing raster data is demonstrated through a use case put into practice in this paper.