Research On Key Techniques Of Millimeter Wave And Infrared Composite Imaging

Millimeter wave imaging can penetrate rain, fog, smoke, dust etc. Infrared imaging has higher resolution. A composite imaging system using both millimeter wave and infrared can be effectively applied in severe environment such as battlefield detection. The researches around the imaging acquisition, processing and fusion of the composite imaging with millimeter wave and infrared are discussed in this paper. The detailed contents are as below:Passive millimeter wave (PMW) and Infrared (IR) thermal imaging model are analyzed and established based on radiometry and in reference to the W-band PMW and uncooled IR thermal imaging system in laboratory. The image restoration with original data is using Wiener filter method in real imaging system. The reasons for PMMW and IR image degradation are mentioned and a method of wavelet transform based on soft threshold is implemented to eliminate the noise in image. In order to resolve the image distortion caused by the cone scanning of millimeter wave imaging system, the nonlinear distortion correction method is used to correct the image. The feature extraction method based on the Otsu image segmentation theory is applied for image registration.In the base of comparing the performance of various types of traditional multi-scare decomposition fusion algorithm such as Laplace Pyramid, Ratio of low-pass Pyramid, Gradient Pyramid etc, the pre-processed and registered images of PMMW and IR by Discrete Wavelet Transform (DWT) are fused as the methods mentioned below need high accuracy for image registration. The measuring standard is determined by combining the two time fusion method based on section characteristics and the features of original data acquired by real imaging system.The simulation program is realized for image reconstruction, preprocessing, registration and fusion. The result of image fusion for experimental data is evaluated subjectively and objectively. The effectiveness of methods mentioned in this article is verified with the simulation result of original data.