Urbanization advances at unprecedented rates, resulting in negative effects on the environment and human well-being. Remote sensing has the potential to mitigate these effects by supporting sustainable development strategies with accurate information on urban growth. Deep learning-based methods have achieved promising urban change detection results from satellite images. However, these algorithms often necessitate powerful cloud-based hardware, which introduces challenges such as latency, bandwidth usage, and data privacy concerns. In the present work, we introduce an end-to-end system for detecting urban change from satellite images, utilizing Edge AI techniques to process data close to their sources and mitigate the challenges associated with cloud usage. The results of the experiments conducted during this study have been promising. The use of active learning to overcome the challenge of annotating satellite images has halved the number of samples needed and the model training time, while achieving performance similar to training on the entire annotated dataset. Furthermore, the compression technique adopted has reduced the model size by more than 50%, thereby minimizing the storage space required at the edge without significant loss of performance. The change detection achieved an IoU of 87% on the challenging Spacenet7 dataset, highlighting the effectiveness of the implemented approach.