[A-CG36] Satellite Earth Environment Observation

[ACG36-02] WebGIS, Web Map Service (WMS), and Web Processing Service (WPS) for Detecting Land Cover Change using Remotely Sensed Data

*Joel Calupas Bandibas¹, Shinji Takarada¹ (1.Geological Survey of Japan, AIST)

Keywords:Land Cover Change, Artificial Neural Networks, Remotely Sensing, Satellite Images, WebGIS

Land cover change detection is very important for environmental change monitoring and identifying areas affected by natural disasters. This paper presents a web-based platform and tools for detecting land cover change by identifying changes in the spatial signatures of land covers using multi-temporal remotely sensed data. Spatial signature change detection involves pixel window value changes identification to detect land cover change. This paper introduces two methods for change detection: The Mean Squared Distance and Artificial Neural Networks (ANN) computing. WebGIS system is developed for the implementation of these methods. The system is used to formulate Web Map Service (WMS) and Web Processing Service (WPS) to render and process remotely sensed data, respectively. Image processing is implemented on the server side using C++, GDALDataset C++ API, and PostGIS. ANN training is done on the WebGIS interface for more user control. Change detection computation results are displayed through the formulation of WMS on the fly for the generated land cover change map. This feature is very important for quick dissemination of land cover change information.
The developed system provides an interface for comparing the accuracy of different methods using different remotely sensed data. It can process remotely sensed data from different sensors like Landsat, ASTER, and Sentinel. The figure shows the interface for training ANN using Sentinel 1 SAR data to detect land cover change caused by the January 1, 2024 earthquake in Ishikawa prefecture, Japan.