Design and phenomenology of a polarization sensitive passive millimeter-wave sensor based on optical up-conversion

Passive millimeter-wave (mmW) sensors utilize a unique imaging modality. They operate by detecting the electromagnetic reflections and emissions of objects at wavelengths of 1-10 mm. This wavelength range is in-between the wavelengths of infrared (IR) and radar systems and shares some similarities to both. Passive mmW sensors are able to image through common obscurants such as dust, clouds, textiles and some plastics like radar systems while also being able to achieve resolutions that are sufficient to create images like IR systems. This unique ability to image through obscurants with sufficient resolution has opened up a variety of applications for which mmW sensors are well suited. Security and defense applications have been the most prominent with the desire to have an imager that is capable of scanning through clothes at stand-off distances without emitting a signal. Dust storm mitigation for helicopter landings is another application that is well suited to this imaging modality due to the low attenuation of dust storms at these wavelengths and the ability to resolve common helicopter landing obstacles. The work of this dissertation focuses on some of the unique phenomenology of this wavelength range. My work has benefited from having access to several passive mmW sensors that have been built at the University of Delaware and from numerous data collection trips to a variety of locations across the United States. These trips have provided a large collection of data from which to observe the different phenomena associated with this wavelength. From observations about the cold radiometric sky temperature and how it affects images to the diurnal changes in radiometric temperature due to thermal cycles; these effects have been observed and reported in this dissertation. Particular emphasis has been placed on the role of polarization and how it affects mmW imagery. A mmW sensor that is capable of detecting two linear polarizations states is described and characterized in this dissertation and then upgrades are discussed which allow it to detect all polarization states. Later chapters discuss the problem of displaying the information from these types of sensors. Several approaches are demonstrated although the specific approach used for a given image is dependent on the statistics of the scene being imaged.