Non-Invasive Environmental Monitoring Using Sensor Data and Machine Learning Techniques

Abstract

Automatic monitoring of the variations of animal populations over time can provide indicators of environmental degradation, given that animal species are sensitive to their environmental conditions. This monitoring can be conducted with the support of Machine Learning algorithms trained with data provided by non-invasive sensors, which generate data that show the animals in their daily lives without interfering with their natural behavior. This paper describes works developed in the Institute of Computing at the Federal University of Amazonas on this topic.