Polarimetric AI-enhanced infrared imaging sensors for autonomous anomaly detection

Purpose and goal:

The aim of the project is to evaluate the ability to detect man-made (flat) objects from land, from air and at sea, using polarization-sensitive (polarimetric) infrared (IR) cameras and to enable autonomous detection of deviant objects using artificial intelligence (AI). The goal of the project is also to develop polarimetric IR sensors with high resolution (HD) with maintained polarimetric contrast, to further increase the usability of such IR cameras.

Expected results and effects:

The project is expected to provide a clear picture of the usability of polarimetric IR sensors for different use cases, such as detection of drones, cars and roads from land and from air, and detection of oil spill at sea. The usability is expected to further increase thanks to implementation of AI algorithms to automatically recognize such anomalies. The use of this technology is expected to incresase in both civilian and military applications, by increasing the awareness of this technology.

Approach and implementation:

A polarimetric IR camera will be integrated into a gimbal and mounted on a car and/or an aircraft to detect anomalies from both land and from air. Data will be collected to train AI algorithms to automatically detect man-made objects.

In parallel, a high-resolution polarimetric sensor will be developed through design, simulations and fabrication of IR sensors.

A reference group within the Swedish IR ecosystem will be formed to disseminate results and to share conclusions from the project.