3-D Imaging: Seeing Deeper, Farther, and QuickerJuly 2012
Topics: Image Processing, Optoelectronics
3-D imaging is familiar to the generations of moviegoers who have sat in theaters wearing funny glasses. But less familiar are the complex 3-D imaging technologies used under controlled laboratory environments to provide highly precise measurements of objects. Researchers are now exploring new optical techniques that will allow for simpler and more mobile high-precision 3-D measurement systems.
The Full Measure of 3-D Imaging
The use of imaging systems to precisely measure an object in three dimensions has important applications in such fields as engineering, architecture, medical diagnosis, and biometrics. However, these systems require a controlled environment, complex equipment, and a significant amount of time to collect their measurements.
Laser radar (LADAR) operates the same way as conventional radar, but rather than radio waves, it tracks the reflection of short pulses of laser light scanned across a scene. Currently, LADAR systems are used for terrain mapping as well as for the precise measurement of complex parts and architecture. However, the bulky size and slow operation of LADAR systems limit their applications.
Improving imaging technology to make such systems more portable and faster, yet still capable of providing precise 3-D measurements, could open up many new uses. For instance, we can imagine 3-D versions of programs like Google Maps that combine their scalable terrain images with their drive-through video. Users could view an aerial perspective of a cityscape and then zoom in to enjoy the smallest details of their surroundings as they take a virtual stroll down one of the city's streets.
To achieve the full potential of 3-D measurement, we need to develop new methodologies that will allow imaging systems to perform high-precision measurements outside of a lab and in a matter of seconds.
Conventional imaging systems use an external light source (such as a camera's flash) when available light is inadequate or when the motion of an object needs to be frozen. 3-D measurement systems also make use of an external light source. The various methods employed in producing that light help determine the abilities and utility of the system. Illumination methods that project light with controlled characteristics onto the object being measured are called "active illumination" methods.
One active illumination method, holographic imaging, uses laser illumination. Variations in the surface of the object cause changes in the reflecting laser wave's timing and amplitude. These changes are compared to a steady reference wave, and the differences are then decoded to provide 3-D images and measurements.
Instead of a steady beam, 3-D measurement systems can set their lasers to produce an ultra-short light pulse that illuminates the entire scene at once. The system simultaneously measures the time delay of the pulse as it reflects off all objects in the path of the pulse and converts the time intervals into 3-D measurement data for the entire scene.
"Structured light" is a type of active illumination where a geometric pattern of light is projected onto an object's surface. The surface's hills and valleys display as distortions of the light pattern, and these distortions can be converted into 3-D data. In a related approach called "coded light," each spot of projected light is encoded with identifying characteristics. Differences measured between the reflections of the many coded spots enable a rapid calculation of a surface profile or a depth map.
Scientists originated the concepts for these technologies years ago. However, recent and rapid advances in optoelectronic technologies—cheap and compact lasers, megapixel semiconductor photodetectors, high-speed computer processing—have provided us with the ability to exploit these techniques to produce 3-D imaging systems that are more compact and inexpensive without sacrificing performance.
One of these emerging 3-D imaging systems is a new type of camera called a plenoptic or light-field camera. A light-field camera captures a 2-D image but also records the paths of the multiple light rays that produced it. In this "4-D" approach, information on the direction of the light paths allows for taking additional measurements and establishing different focus points in post-production.
In a recent demonstration, researchers constructed a light-field camera by synchronizing 100 inexpensive cameras to capture multiple images from widely different viewpoints. Combining these images generated high-quality 3-D measurements of the scene.
3-D on the Front Lines
Commercial use of 3-D imaging technologies is increasing at a rapid pace. The automotive safety market depends on precision 3-D measurement for design and testing. And the entertainment industry is never at a loss for creative uses of 3-D imaging. For instance, Microsoft's video game system Kinect employs coded light.
Despite having different and more stringent requirements, defense and security applications can nonetheless exploit the imaging resources the commercial sector is bringing to market. MITRE's military and security sponsors require 3-D measurement systems capable of timely recognition of furtive targets from great distances in unpredictable conditions with minimal errors. To help the military identify or develop such systems, MITRE researchers are analyzing current 3-D imaging and measurement methodologies, as well as following current research into developing new methodologies.
Further, we are determining which methods may yield the highest benefit vs. cost improvements through well-chosen technological investment. We believe that our best hope to meet the military's challenges is to use active illumination approaches that will shift the burden of imaging away from traditional lens-mounted camera systems toward computational image generation.
—by Robert Latham