Project Abstract:

ClearFast aims at improving the efficiency and safety of humanitarian demining operation by (1) reducing the false alarm rate, (2) increasing the clearance rate, (3) increasing the rate of survey.

CLEARFAST proposes 3 main innovations: (1) A new concept for minefield survey; (2) for the 1st step, implementation of a novel analysis of image sequence under changing natural illumination, using a Multi-temporal & Multi-spectral IR(MSIR) sensor optimised for background rejection; (3) for the 2nd step, a motion-driven data fusion using Ground Penetrating Radar, electromagnetic Metal Detector with sensitivity adaptation and MSIR reducing the well-known weaknesses of IR imaging. CLEARFAST proposes a system qualification both in a controlled environment and in realistic conditions. Although Level 2 survey is the main application area, CLEARFAST may be used for Level 3 of quality assurance.

Objectives:

The ClearFast objective is to improve the efficiency and safety of demining operation by (1) reducing the false alarm rate, (2) increasing the clearance rate by providing more accurate survey data, (3) increasing the rate of clearance. To achieve these objectives, ClearFast aims at developing a robust mine detection system, relying on fusion of multiple sensor data - GPR, IR, MD. The system performance, however, ultimately depends on the performance of the individual sensors. Hence ClearFast will reduce the well-known weaknesses of IR imaging by implementing a novel multi-temporal/multi-spectral IR system. The MD array will be adapted in sensitivity to achieve detection and imaging capabilities. Besides the technological innovations, ClearFast proposes a novel concept for Level 2 surveys. It will be validated in a controlled environment with realistic conditions.

Description of the Work:

ClearFast develops a transportable mast-mounted MSIR sensor and a multisensor platform mounted on a vehicle. Both systems are based on the requirements of a two-phase Level 2 survey model. The post-project exploitation, aims at covering the full range of activities required in mine clearance, also envisages airborne (Level 2) and hand-held systems. ClearFast includes several phases, each of them ended by a formal review. Phase 1 defines the system according to the ClearFast concept, the end-user requirements and the technology constraints. Phase 2 is planned for signal processing enhancement of individual sensors and adjustment to the system interface requirements. It includes data fusion for the multi-temporal MSIR sensor. Phase 3 concerns multi-sensor data fusion. Two developments are carried out in parallel for risk mitigation. The first design involves the adaptation to APL detection of an existing probabilistic algorithm for target detection based on a learning process of the local environment. The second design is based on algorithms for deterministic shape/attribute verification. Phase 4 is the integration of the system components in the Level 2A survey platform and the multi-sensor platform. Phase 5 is dedicated to tests and performance evaluation in realistic conditions. This phase is driven by a commercial demining end-user, to demonstrate that (1) Level 2A surveys increase efficiency, speed and safety of subsequent Level 2B minefield delineation and area reduction, and (2) the Level 2B multi-sensor suite and associated data fusion allows for reducing the false alarm rate. Special attention is given to the selection of a representative set of objects, normally triggering false alarms in the target area within the Balkan. For efficiency evaluation, the main parameters are number of undetected mines, number of false alarms, potential improvement that can be obtained between prototype and industrial system, operational cost, speed and safety.