C2A-DC: A context-aware adaptive data cube framework for environmental monitoring and climate change crisis management

Abstract

A context-aware adaptive data cube (C2A-DC) framework based on Earth Observation (EO) data for environmental monitoring to mitigate Climate Change (CC) effects is proposed. It has the property of combining DC formation, calculation of Remote Sensing (RS) operations, deep and machine learning algorithms for classification and data harmonization by updating the layers of the DC using information obtained from the previous operations, all applied to EO data originating from a satellite selected by a stakeholder. Moreover, the proposed framework is context-aware adaptive in the sense that it allows for environmental monitoring tasks according to a stakeholder’s needs, which is in contrast to existing works being constrained to a single type of environmental monitoring. Furthermore, with the proposed framework end-users do not have to make any additional processing action to handle the data as they are given to them as a set of analysis-ready data, harmonically aligned under a unique DC. To showcase the C2A-DC framework’s effectiveness and applicability in environmental monitoring, crisis management, and post-disaster assessment, it is applied to two real-world cases. The first one corresponds to cloud and shadow detection in a flood event, with several Convolutional Neural Network (CNN) architectures being trained on the C2S-MS floods dataset for classification, while k-means clustering and RS product calculations are used for monitoring scenery changes. The second one corresponds to deforestation monitoring from fires, evaluated using RS products.