Research On Channel Equalization And Contour Extraction Algorithm For Passive Millimeter Wave And Terahertz Imaging

The passive millimeter-wave and terahertz imaging systems do not emit electromagnetic waves,but they can achieve image and realize the concealed objects through detecting the radiation energy that is emitted and reflected by objects in the experimental scene.The system will not injury to the human body.Furthermore,they not only enable rapid imaging detection,but also does not result in personal privacy issues,which makes the technology has great application prospects and commercial value in the security inspection field.However,due to the limitation of the manufacturing process and its high cost,the response between each detection channel on the focal plane is not exactly the same,which seriously affects the imaging quality of the detection system.As a result,the accuracy of the edge detection is not high.It is necessary to study the channel equalization technique and contour extraction algorithm.The research content of this paper is based on the actual research projects.The main contents of the research are as following:(1)The basic theory of millimeter-wave and terahertz passive detection imaging technology is studied,the radiation characteristics of the object under this frequency band are studied,and the imaging models of millimeter-wave and terahertz passive detection are analyzed.(2)Researched the image channel equalization algorithm based on Kalman filter.The problem of Kalman-filter-based non-uniformity correction algorithm is resolved,and the fast equalization based on steady-state Kalman filter is studied.The algorithm improves the overall real-time performance by improving the update mode of the algorithm’s gain matrix.(3)The channel equalization algorithm based on neural network is studied.By constructing a new cost function on the basis of this algorithm,adaptive adjustment of iterative step size factor is realized.An improved neural network algorithm is studied and proposed.The channel equalization algorithm has achieved excellent millimeter-wave equalization image results.(4)For the traditional single-threshold segmentation algorithm in the edge detection of millimeter-wave images appears "blush" and "broken leg" phenomenon,an improved contour extraction algorithm based on Canny operator is proposed.(5)The contour matching algorithm based on the shape context is studied.By using the corner point detection algorithm based on the curvature scale space to initially select the target to be matched,the shape matching accuracy rate is ensured and the recognition rate of the original algorithm is improved.Experiments show that the algorithm can quickly and accurately match the hidden target category.In summary,all the improved algorithms in this paper can effectively improve the detection performance of hidden targets in millimeter-wave passive imaging,and have practical engineering application value for millimeter-wave passive detection imaging technology.