Buried landmine detection using microwave radiometry

At present many million anti-personnel mines are polluting our environment in many countries. They cause a considerable limitation of the living space and make the land unuseable for agricultural purposes. Because the mine clearance procedures are much slower than the mine laying operations the number of polluting mines and the related contanimated areas are nowadays still increasing. For the current mine detection technologies and for the most modern mines, which become continuously smaller and have less metal content, a high false alarm rate of up to 1000 to 1 is very likely. Classical mine detection procedures like using metal detectors or dogs, or brute-force methods like mechanical plowing through the soil by tank-like vehicles are not able to solve satisfactorily the mine contamination problem. They suffer mainly from many drawbacks like a high false-alarme rate, unclear reliability, high costs, and a destroying mode of operation, which cannot be tolerated in each situation. Thus an advanced mine detection system consists of multiple sensors based on different physical phenomena and principles and benefits from fusing their data. By doing this the different advantages of each sensor can be isolated from its shortcomings and merged to match as comprehensive as possible the specific mine detection requirements. In the past we developed a broadband microwave radiometer operating in the frequency range of about 1-7GHz. The radiometer was designed as a part of a three-sensor system within the HOPE project using additionally a metal detector and a ground penetrating radar. Using a common search head, the system was intended to be hand-held operated as the conventional metal detector, but allowing the application of imaging for all three sensors in conjunction with an optical-camera-based position monitoring system. In addition the capability of measuring at multiple low-bandwidth channels by sweeping the center frequency for the radiometer allows the examination of the layered structure as in the case of a hidden object. Taking this signature as a finger print of the specific depth profile, this information can help to discriminate landmines form other objects. This contribution discusses comparisons of measured and theoretically simulated signatures of typical scenes