Aries

The ARIES project (Advanced Resolution and Intelligence for Explosive Sensing) tackles the growing problem of surface-landmine detection in challenging environments where conventional methods fail due to vegetation, soil variability, camouflage and noise. It proposes a context-adaptive, multi-modal sensing framework that fuses data from passive visual and geophysical sensors, including hyperspectral, thermal infrared, polarization imaging and magnetometry. The goal is to develop a lightweight, mobile and adaptable detection system with high accuracy and low false-alarm rates, suitable for deployment on UAVs, UGVs or soldier-carried platforms. The project follows a two-phase data collection and development process, first gathering a representative dataset across diverse environments and conditions to prototype and evaluate sensor fusion and mine detection algorithms, then acquiring a larger dataset to optimize performance and robustness. The detection system is refined for mobile deployment through neural network compression and GPU acceleration, with validation carried out on datasets and field trials. Work is organized around data acquisition, sensor fusion development, validation and benchmarking, and optical filter design, in close collaboration with project partners. IPI is involved in the acquisition, design of the sensor rig, the development of the sensor fusion and detection algorithms, the validation and the GPU acceleration.

DEFRA, 3/2026 – 2/2029