Passive Millimeter Wave Imaging for detection of concealed contraband has been demonstrated for decades in various forms. One vexing problem is that when some emissive materials reach body temperature they lose contrast. A solution to this problem, along with results of an implementation, is presented.
We have developed a novel approach to performing automatic detection of concealed threat objects in
passive MMW imagery of people scanned in a portal setting. It is applicable to the significant class of
imaging scanners that use the protocol of having the subject rotate in front of the camera in order to image
them from several closely spaced directions. Customary methods of dealing with MMW sequences rely on
the analysis of the spatial images in a frame-by-frame manner, with information extracted from separate
frames combined by some subsequent technique of data association and tracking over time. We contend
that the pooling of information over time in traditional methods is not as direct as can be and potentially
less efficient in distinguishing threats from clutter. We have formulated a more direct approach to
extracting information about the scene as it evolves over time.
We propose an atypical spatio-temporal arrangement of the MMW image data - to which we give the
descriptive name Row Evolution Image (REI) sequence. This representation exploits the singular aspect of
having the subject rotate in front of the camera. We point out which features in REIs are most relevant to
detecting threats, and describe the algorithms we have developed to extract them. We demonstrate results
of successful automatic detection of threats, including ones whose faint image contrast renders their
disambiguation from clutter very challenging. We highlight the ease afforded by the REI approach in
permitting specialization of the detection algorithms to different parts of the subject body. Finally, we
describe the execution efficiency advantages of our approach, given its natural fit to parallel processing.
We describe a low-cost passive millimeter wave (MMW) scanning camera for detecting concealed weapons and contraband. It is based on a focal plane array of 64 radiometric channels that employ MMICs operating at 94 GHz. Equipped with a 125 mm primary optic, the camera achieves a 26 26 degrees field of view by means of a rotating optic that performs 10 conical scans of the scene per second. The resulting 10 Hz rate images are of size 28 by 28, yielding a spatial resolution of 5 cm at a range of 1.6 meters from the camera. The radiometric sensitivity, at the maximum frame rate, is given by a median of under 3 Kelvin. With a size of 8 in. 8 in. 22 in. and a weight of 26 lbs., the camera is very compact and portable. This development may constitute the first affordable, commercially available passive MMW scanning camera.
When operated at the slower frame rate of 1 Hz, the resulting time integration improves the image to less than 1 Kelvin, making the camera well suited for the detection of a wide variety of threats at security checkpoints. At finer camera sensitivity levels, the possibility arises of the exposure of anatomic details of the scanned subjects. In view of this, we have developed specialized display software that allows the presentation of the MMW scanning results in a manner that overcomes privacy concerns.
The semi-automated film video reader system (SAFVR) is an integrated system for motion sequence analysis, including acquisition, qualitative analysis, quantitative analysis, and storage of tracks and images. The SAFVR system can digitize high resolution images from film and video, save the digitized images to disk, perform object tracking for rigid bodies, and produce video tapes for presentation of analysis results. The tracking is based on a hierarchical correlation matching algorithm.
The application of computer technology to the automation of visual tasks is a difficult and time consuming process. Until recently, the lack of powerful software tools has required that the developer of computer based imaging applications be capable both of programming computer systems and of conducting imaging research. The characteristics of the development process that should be captured in such an environment are related to human-machine interface issues. This paper reports on the user interface issues that were encountered during the implementation and ongoing development of a commercial product, designed to aid in the construction of image understanding applications