Research On Passive Millimeter Wave Imaging Theory And Super-resolution Signal Processing Technology

Passive millimeter-wave (PMMW) imaging system forms images by detecting the millimeter-wave radiation energy from the scene and utilizing the differences of the radiation intensity. It is widely used in airport security, scene monitoring, plane blind landing, medical diagnosis and environmental detection, et al. However, the obtained images usually have poor resolution due to the weakness of the target's self-radiation energy and the finite size of the antenna and the underlying diffraction limits, which leads to difficulty for analysis and detection. In this paper, we focuss on the PMMW imaging theory and the super-resolution algorithms and propose some methods to improve the spatial resolution and to enhance target detection ability based on the focal plane linear array scanning process, which mainly contain the following contents:1.According to the millimeter-wave radiation characteristics and the PMMW imaging mechanism, several different PMMW imaging forms are analyzed in the framework of the limited spatial spectrum inversion, and the mathematical nature is also described. This work establishs the theoretical foundation for the design and analysis of the PMMW system.2.A spatial-variant image mathematical model is built up by introducing a focal image plane transformation operator and deriving the spherical projection relationship and the non-uniform multiple-beam antenna model. This model can separately express the process of propagation, accumulation and reception, which are the three different attributesis space variant factors on the effects of the imaging quality.3.A spherical back-projection transformation method is proposed to eliminate the space variant and geometric distortion. This method transforms the original space variant problem into a space-invariant problem, thus the difficulty and calculation amount of the super-resolution algorithm is significantly reduced, which enhances the performance. To solve the problem that the nonlinear operation of the super-resolution algorithms corrupts the passband frequency, a frequency spectrum correction algorithm is proposed. By using this correction algorithm, the frequency spectrum within the pass-band can be effectively restored, and the ringing effects are notably reduced. 4.A maximum likelihood multi-grid super-resolution algorithm based on wavelet decomposition (WMLMG) is proposed. The problem of the spectrum aliasing between extrapolated frequencies and passband frequency components is resolved. Thus, the aliasing artifacts in restored images are reduced and the resolution of PMMW images is enhanced.5.To detect the dim small moving targets, an improved dynamic programming algorithm based on truncated sequential probability ratio test (SPRTD_IDPA) is proposed. The algorithm can not only detect dim object and get target trajectory but also overcome the problem of the conventional DPA algorithm, which is not accurate enough at the beginning of tracks and exists a small number of false alarm.To sum up, research on the PMMW imaging theory and super-resolution signal processing algorithm, not only can enhance the resolution of PMMW images, but also have important guiding significance for the design of PMMW imaging system.