Passive Millimeter Wave Imaging Algorithm Research

Passive millimeter wave (PMMW) imaging makes use of the differences in the distribution of millimeter-wave radiation energy of scenarios and different target to achieve imaging. The passive millimeter wave imaging has the ability of passing through clouds with rain and fog, passing through battlefield smoke, working at any weather and any time and penetrate parcel layer to detectc non-metallic targets with a certain thickness. It is the passive detection system, so it can not easily be found. These advantages make it has great value for application in the scene to monitor the aircraft blind landing, battlefield surveillance, military reconnaissance, anti-terrorism and other areas. However, due to the finite size of the antenna and the underlying diffraction limits, PMMW imaging system always gets the low resolution image. This is not conducive to observe and analysis for the objective. Therefore, in order to enhance the image resolution and ensure the Observations and analysis of the images can carry out successfully, we need to do super-resolution processing for the passive millimeter-wave image. Supported by National Natural Science Foundation of China, super-resolution algorithms which are the key technology in passive millimeter wave imaging are researched in this dissertation. The main contents include:1. The similarities and differences of passive millimeter wave imaging super-resolution, image restoration, image resolution enhancement, image sequence super-resolution is analyzed. The conditions and ideas of passive millimeter wave imaging super-resolution are summarized. This will be the guiding role for the design of super-resolution algorithm.2. To solve the problem of the inaccurate priori model and the difficulty to choose proper relax parameter, we researched Projected Wavelet-domain Maximum A Posteriori (PWMAP) estimation super-resolution algorithm. This algorithm based on maximum a posteriori estimate, extrapolates the spectrum by using the non-linear projection operation and achieves parameter adaptive updates. It gets a good super-resolution effect in the experimental results.3. Using the sparsity of an image decomposes in over-complete basis, we researched a Super-resolution Algorithm Based on Sparse Prior. This algorithm gets a good super-resolution effect in the experimental results. Especially, it has a better semi-convergence effect in the small SNR than the other algorithms.4. The problems of Maximum Likelihood Frequency Domain Correction Super-resolution Algorithm are analyzed. By constructing a new optical transfer function and a new spectrum correction formula, we researched a Modified Maximum Likelihood Frequency Domain Correction Super-resolution Algorithm. And this algorithm gets a good super-resolution effect in the experimental results.5. The necessity of multi-frame processing in passive millimeter wave imaging is analyzed. Using the multi-frame image sequence with sub-pixel shifting of the same scene, we researched a multi-frame super-resolution algorithm for passive millimeter wave imaging to solve the spectrum aliasing issues in passive millimeter wave imaging. And this algorithm gets a good super-resolution effect in the experimental results.In conclusion, all the research and researched algorithms in this article not only can realize the PMMW super-resolution, but also have important guiding significance for the design and improve of the related algorithms.