Segmentation Method And Implementation For Detecting Concealed Objects In Compton Backscatter Images

Compton back-scattering(CBS)is able to generate body scan image under complex background as a widely used security check technology.However,due to the complexity of background environment in the process of generating CBS image,images of materials carried by passengers are segmented with great difficulty.In order to fit the growing security requirement,a body concealed object detection and recognition system is developed to offer real-time detection and identification services for security equipment,and intensive analyze and research the segmentation of material image which is the key technology in the system.Enhance the detection rate of the system by a segmentation method based on CBS body image.An iterative segmentation method based on CBS body image is designed according to the distribution of gray value of the human body image and the comparison of the applied range and the advantage and disadvantage of various image processing technology.Firstly,the image is filtered mainly on background filter and noise filter.Secondly,the edge of the image is detected by Canny algorithm.Finally,a Binary segment image of carried material is recovered,extracted and padded by expansion and etching method and connected component labeling method.Experimental results show that the segmentation method in this paper has a greater detection rate of metal with high atomic number and non-metallic material with low atomic number.The paper completes the system requirements analysis and modeling based on the software engineering thought and Unified Modeling Language.In order to apply to the public security detection,the human concealed object detection and recognition system is developed,with the method and BP neural network pattern recognition technology,to intelligently identify the extracted concealed object.