Research On Technologies Of Moving Concealed Targets In Passive Millimeter Wave Imaging

Passive millimeter wave imaging system can receive millimeter-wave radiation energy from the scene and targets, and use differences from radiations among variety of targets for the realization of imaging. Besides, this kind of imaging is characteristic by radiationless, concealment, non-contact and right resolution, and poses no threat to detectors’ physical condition and individual privacy. Consequently, it can apply to security inspection. Thus it can be seen that exploiting Passive Millimeter-Wave(PMMW) Imaging Technology aims at realizing the detection of the hidden carry-on contraband’s forms images, which is a hot research area and a mainstream of researching development in recent years.Technologies of moving object detection which makes use of the difference between the features of hidden target(such as the brightness temperature, scale, shape, motion of target, etc.) and background to detect, segment and locate the hidden target. Recent years, the PMMW Imaging Technology is used to detect the hidden contraband(such as gun, weapon and plastic explosive) and it has applied to camp, airport, bank, venues and other important places. Hence, technologies of moving object detection are a key to the realization of PMMW Imaging Technology.Based on the scientific research project, the article studies Technologies of Moving Object Detection in PMMW Imaging Technology. The major contents in this thesis include:Firstly, this article studies PMMW Imaging Technology, analyzes the radiation characteristic of the hidden target, and researches the detection method and theory of moving object via millimeter-wave imaging technology.Secondly, this thesis analyzes the feature of millimeter-wave imaging, and studies the denoting technology, Background Suppression, and segmentation technology of the millimeter-wave imaging. Besides, this paper introduces the wavelet transform for improving conventional Otsu algorithm, which proves that Otsu algorithm can suppress the interference from complex background towards image segmentation.Thirdly, regarding the issue that complex scene create large numbers of false alarm targets, this paper puts forward Detect Before Track(DBT) algorithm based on region of interest(ROI). And DBT algorithm can reduce target searching area effectively and increase the detection rate.Fourthly, this thesis researches the Particle Filter Theory and verifies the effectiveness in trailing the object trajectory through simulation.Finally, on the issues of mobile hiding spot and area target in the near-field millimeter-wave imaging, this paper studies Track Before Detect(TBD) algorithm based on Particle filter. In addition, author adopts eigenvector and construct observation model according to the different features of point target and area target. And this paper also puts forward an adaptive window tracking method based on particle filter tracking before detection algorithm(PF-TBD), to solve the problems of the traditional fixed window unsound, which increases the validity of algorithm.