Millimeter-wave Passive Imaging Super-resolution Algorithm And Dsp Implementation

Passive millimeter-wave(PMMW) imaging system utilizes the differences of radiant intensity of different objects in the scene to achieve the goal of imaging. Compared with infrared, optic and microwave imaging, Millimeter-wave imaging has unique advantages. PMMW imaging has important application value in the application areas such as remote sensing, navigation, security check and battlefield scrutiny. However, PMMW imaging always has low-pass filtering effect due to the finite size of the antenna and the underlying diffraction limits. Hence, acquired images have poor resolution. The resolution of acquired images can be enhanced through using super-resolution image recovery algorithms, besides increasing aperture dimensions.This paper focuses on millimeter-wave super-resolution image recovery and DSP implementation, carried on research and experimental verification. The main contents are as follows:1. Aim at different super-resolution recovery algorithms restoration effectiveness for millimeter wave image,comparative analysis of single-frame statistics-optimized super-resolution algorithms, including maximum likelihood estimation based Lucy-Richardson algorithm, statistics-optimized frequency domain correction algorithm and other three algorithms.2. Aim at the disadvantages of single-frame super-resolution algorithms not using complementary information between image sequence, research on multiple-frames subpixel shift super-resolution algorithms, image restoration ability can be further improved.3. Estimate computational complexity and memory requirements of statistics-optimized super-resolution algorithms based on system real-time requirements and practical hardware platform. Implement these algorithms on high performance DSP, and ameliorate and optimize algorithms on real-timing of system and image recovery effect.4. Accomplish the design of real-time signal processing unit of PMMW imaging system, using ADSP TS-201 as the core processor, achieving the functions of communications with data acquisition FPGA, channels calibration, image rearrange pre-treatment, super-resolution image recovery function library.Through research above, effectiveness of super-resolution algorithms is verified by simulation,and it is validated by experiments that real-time signal processing unit can meet the requirements of actual system.