SCOUT: Non-intrusive wildlife monitoring system based on computer vision

Tracking and monitoring wildlife outdoors required manually reviewing large volumes of images, many of them false positives caused by the mere movement of vegetation, in addition to frequent trips to the field to maintain the deployed equipment.

The project's objectives were to:

• Build a non-intrusive monitoring system for wildlife surveillance in open spaces.
• Develop a field image capture device supported by sensor and communication technologies, computer vision, and machine learning.
• Centralize data from the different monitored areas on a platform that would make it possible to monitor the deployed devices and raise alerts about any irregularity, both in the natural environment and in the equipment itself.

SCOUT was structured around the following areas:

• Smart capture device: field equipment that applied computer vision and machine learning to select relevant images, reducing false positives caused by the movement of vegetation.
• Remote monitoring platform: a centralized system that periodically received data from the different areas, generated alerts in case of anomalies, and enabled remote tracking of the devices' status.
• Battery life optimization: improvement of the capacity and power consumption of the detection device to reduce maintenance trips to the field.

Biosfera Consultoría Medioambiental had the collaboration of CTIC in developing the project. The system made it possible to reduce the time needed to analyze wildlife flows in a given area and to expand the scientific knowledge available on protected animal species, with the resulting socioeconomic impact in the territories where it was applied.