A Low-Cost Labeling Service for Satellite Imagery Data

Abstract

Satellite imagery data can be used to detect areas of built-up land. Various satellite image data processing platforms have emerged in line with the sophistication of technological developments. However, from these many platforms, there are code, access, and storage limitations. So, to make a good dataset, correction, and collaboration between users are needed. A good dataset will produce a good model. Therefore, in this study, we present the software architecture of data labeling services. The proposed research method consists of data management using MongoDB, data labeling service backend using FastAPI, and frontend using Flask and LeafletJS. Furthermore, we present some empirical results of this study regarding usability, performance, cost, and comparative analysis of our proposed low-cost satellite imagery data labeling service.