Research On Detection Algorithm Of Targets For Passive Millimeter-Wave Image

The passive millimeter wave(PMMW) imaging system obtain image through detecting the difference of millimeter energy that all kinds of objects radiate. It can penetrate the clothes and other coverings effectively and imaging of the concealed metal target. PMMW imaging system didn't emit electromagnetic waves, so it's no radiation to the human body, and has good concealment. Secondly, because of less loss of millimeter wave propagation in rain and fog, it can penetrate them, and the PMMW imaging system can work around the clock. Based on the above advantages, currently in the field of security and military application, it has great potential applications using this system to detect concealed target. The target detection technology is still a key technology of the passive millimeter wave imaging system in practice, so research target detection algorithm for PMMW imaging is of great significance.Based on the specific research project, this paper carries out a thorough research for the target detection technology of passive millimeter wave images. The main research contents are as follows:(1) Described the detection theory of PMMW, analysis the radiation characteristics of metal objects in millimeter wave band, and the characteristics of the imaging results.(2) For PMMW imaging with no interference target in background, we researched a novel algorithm based on Otsu method for target detection in PMMW images. Aiming at the traditional single-threshold Otsu method can't segment PMMW image effectively, we proposed double-threshold Otsu method, and introduce the area information of each region in PMMW images to modifies weighting coefficient in the objective function, improve the accurate in target detection. The simulation results show that the algorithm has better performance in target detection than other traditional threshold segmentation algorithm based on Otsu.(3) For the PMMW image has low signal-to-noise ratio, we study the DBSCAN(Density-Based Spatial Clustering of Applications with Noise) algorithm, which can overcome the noise. But the DBSCAN algorithm has poor effect for image that has fuzzy boundary, aiming at this problem, this paper puts forward a fast clustering algorithm based on local density to improve the detection accuracy of the algorithm. The simulation results demonstrate the proposed algorithm has better performance in target detection than DBSCAN algorithm.(4) For the PMMW imaging with interference target in background, use the threshold segment method will generate false alarm problem, we research the target detection algorithm based on self-organizing background modeling method, to eliminate interference target in background. This paper also improved algorithm to improve the convergence rate of the algorithm. The simulation results show that the algorithm has good effect on target detection.