Sensors used for security purposes have to cover the non invasive control of men as well as surroundings of buildings and camps. Military checkpoint protection against terrorists carrying weapons or bomb belts is one of the most important tasks to be taken in this domain. Special work has to be done to solve the disadvantages of commonly used systems like metal detectors. Another request is to develop systems, which are lighter, more modular and have a higher sensitivity. The work described in this paper concentrates mainly on sensors at 0.1 and 0.2 THz, which in contrast to the above mentioned systems are able to detect non metallic objects like ceramic knifes. In addition standoff surveillance is possible which is of high importance with regard to suicide bombers. To help operators of future sensor systems images with high resolution will be generated or even overlaid on optical images. The used systems active as well as passive are based on state-of-the-art solid-state components. Since the resulting detection probability depends not only on the sensor properties, but also on the signal processing, more and more effort is put into the enhancement of the data and its fusion with other sensor outputs. 1.0 INTRODUCTION Security checks are getting more and more common in several areas of life. These can be a nuisance for passengers or visitors undergoing the security checks, as today's approaches tend to take more and more time compared to the flight or venue itself. On the other hand, the security personnel face increasing numbers of people to be screened. Taking into account that each technical aid, like metal detectors and Xray machines, can only deliver a limited set of information, each new system can add an indication which may help to distinguish actual threat material from other items. This, in turn, can help to increase overall security while reducing the annoyance that a security check can be today. As in former times, when passengers at airports spent most of their time waiting for their luggage or mounting the airplane, now more and more time has to be spent on the security checks. Also, the threat has changed, because terrorists prepare to mislead classical metal detectors. This is why new systems are necessary, which are capable of detecting concealed weapons and explosives, even if they are not made of metal. Furthermore, mm-wave imaging, as well as active radar techniques, can be used to operate in a standoff scenario. This means that persons can be screened at distances of up to 20 meters. This field of application will be very important in the future, to detect a potential danger already before the hazardous person can enter a sensitive area or building. 2.0 RADIOMETRIC IMAGING 2.1 Main principle One possible approach to detect concealed weapons is by using radiometric systems. In principle, a scanner using this technique measures the thermal noise of the radiation reflected by the body. This is equivalent to the temperature on the surface of the body. The detection and imaging of objects concealed RTO-MP-SET-125 48 1 Imaging of Concealed Weapons and IED using mmW Technology under the clothing is based on the fact, that different materials exhibit different radiation temperatures TB depending on their own temperature and their physical properties. It is a linear combination of absorptive/emissive (ε), reflective (ρ) and transmittive (τ) characteristics: TB = εTO + ρTU + τTE. TO is the physical temperature of the object, TU the environmental temperature and TE the background temperature. At mm-wave frequencies, the radiation is capable of penetrating clothes and other materials, with transmittivities of τ = 0.3...0.8. It is obvious that due to these different properties, objects differing from skin and fabric can be detected. This is possible over a wide range of frequencies, but a compromise has to be found between best contrast conditions at lower frequencies and better real beam imaging properties with a given antenna diameter at higher frequencies. In addition, some materials like explosives are transparent in lower frequencies, so to detect them as they reflect the surrounding radiation, higher frequencies should be used. In contrast, fabrics where these explosives may be hidden under are also more transparent at lower frequencies. Here, a good agreement between highand low frequencies should be found as well. 2.2 Experimental setup Up to now, most of our experimental setups were well suited for close distance scanning, with the advantage of givin very clear and detailed images of measured persons and concealed weapons. In this paper we concentrate mainly on standoff detection. The systems presented here work at 220 GHz, with some 94 GHz results being shown as well for comparison. Although the atmospheric attenuation is generally considered to increase with frequency in the mm-wave band, there are frequency ranges around 220 GHz, which lend themselves to imaging purposes. because they exhibit relative low attenuation. This makes it possible to look through clothes even under standoff conditions. The radiometric system consists of a Cassegrain antenna mounted on top of a pedestal. The aperture has to be moved to point the single beam of the system to each measured spot on the target to be observed. Up to now, this movement gives the limitation in scanning time. Figure 1 shows the system setup. The radiometric part consists of three LNA stages, as shown in Figure 2. Figure 1: Radiometric System for standoff detection of concealed weapons working at 220 GHz. The radiometer used here is a Dicke-type direct receiver using either a PIN-switch or – as shown in Figure 1 – a chopper wheel. So, for stability reasons, the received signal is compared to a fixed noise temperature, and drifts caused by the very sensitive amplifiers are cancelled. The 220 GHz LNAs were 48 2 RTO-MP-SET-125 Imaging of Concealed Weapons and IED using mmW Technology supplied by FhG-IAF and represent the most advanced MMIC LNA type available with respect to noise figure [1]. Figure 2: Photograph of the 220 GHz front-end. From left to right: Receiving horn, first LNA stage, second LNA stage, high-pass filter, third LNA stage, attenuator and detector. Chopper Wheel Frequency 200 Hz LNAs Bandwidth >10 GHz Noise Figure 10 dB Radiometer Gain >50 dB System Noise Temperature 2000 K Temperature Resolution ~0.12 K Table 1: Performance of individual components and overall radiometer channel. Figure 2 shows a photograph of the receiver, when using a chopper wheel for calibration. Alternatively, a PIN-switch is mounted between the horn and the first LNA stage. The total gain of the receiver is more than 50 dB, giving a noise temperature of about 2000 K. Figure 3 shows a photograph of a four-stage 220 GHz low-noise amplifier MMIC. The G-band LNA was designed by FhG-IAF to achieve high gain in combination with a low noise figure at 220 GHz, which is also shown in Figure 3. Figure 3: Left: Chip photograph of a four-stage 220 GHz cascode low-noise amplifier MMIC. The chip-size is 1x2.5mm2. Right: Measured gain and noise figure (NF) of a 220 GHz low-noise amplifier module. RTO-MP-SET-125 48 3 Imaging of Concealed Weapons and IED using mmW Technology 2.3 Typical measurement results The shown standoff measurements were carried out outdoors. Because the system was equipped with a single receiver, the two feet dish had to be moved completely for real aperture scans. Therefore, a single measurement at 220 GHz took about 20 minutes to complete. The person was standing 10 m away from the measurement device. Figure 4: Scan of a person at 10 m distance. Left: Without bomb vest or weapon. Right: Person wearing a bomb vest under the coat. The person measured in Figure 4 and Figure 5 was wearing different weapons or alternatively a bomb vest beneath the clothes. The bomb vest was fabricated by putting debris from metal together with screws into plastic cylinders. By filling up the cylinders with plastic glue, the signature of a real bomb vest could be simulated. Figure 5: Scan of a person. Left: Without weapon. Right: With hidden weapon. The gun worn at the chest can be clearly identified. In all shown radiometric images (Figure 4 and 5) you can see a bright band going from the throat down to the belly. This is the part of the jacket where the zipper was closed. For additional tests, optical images and radiometric measurements were superimposed to get a better understanding of where hazardous objects have been worn. To implement this feature, a picture was taken from the exactly same direction as the mm-wave camera was looking. The radiometric images were then 48 4 RTO-MP-SET-125 Imaging of Concealed Weapons and IED using mmW Technology piped through a thresholding algorithm and then superimposed onto the black and white optical image. Figure 7 shows examples of these measurements. In the left picture the bomb vest can be seen clearly, together with the belt-buckle. The right picture shows a hidden gun worn in front of the chest, as well as some metallic object hidden in the right pocket. Figure 6: Optical image of scanned persons. The prominent parts of the radiometric images have been superimposed on the images. In the left picture the bomb vest worn beneath the jacket can clearly be seen. Additionally, you can see the belt-buckle. On the right image the hidden gun can be clearly identified together with some metallic object hidden in the right pocket. Figure 7 shows a standoff scene which was scanned for test purposes. To increase the contrast in this measurement, the person and the bicycle were put in front of an absorber material, so the reflection of the diffuse illumination gives a very good image quality. By deconvolving the data with the measured beam, a better image quality can be achieved. The measured beam size is 0.23° FWHM, which is close to the theoretical value of 0.18° using the above configuration at 220 GHz. Figure 7: Radiometric imag