A deep learning approach to anthropogenic load assessment for sustainable coastal tourism

Humans' activities are playing a significant role in affecting our planet conditions, accelerating climate change, and wasting our natural resources. Among the actions that can be done to support sustainability and people's awareness about this issue, monitoring environmental conditions and collecting data can be performed thanks to the availability and the adoption of digital and technological solutions, such as Internet of Things infrastructures and deep learning approaches. This paper proposes an approach to support sustainable coastal tourism, which is based on the Social Internet of Things paradigm and on a deep learning-based classifier, with the aim of monitoring and counting people's presence on beaches and coasts. The paper describes a system prototype architecture and a classifier model we have defined, discussing the results obtained in the testing phase we have conducted. The defined model shows interesting and promising results in terms of precision and accuracy.