背散射人体成像中隐藏物识别方法研究-计算机科学与技术专业论文.docxVIP

II ABSTRACT Image recognition has been one of the most hot research fields in computer graphics. This paper mainly studies the recognition of backscattering body concealed objects, includes: image preprocessing, feature extraction and the recognition of concealed objects. According to the characteristics of the human body image, firstly, we adopt a combination denoising method, which is composed of median filter and Gaussian filter method. This method not only can remove the noise, but also can retain filtering image edge information. Then, An multi-level segmentation method，which is based on the global segmentation and the local segmentation, is proposed. The global segmentation separates the body and the concealed objects area from the CBS image, and the local segmentation segments the concealed objects from the body area. Experimental results show that the method can effectively segment the concealed objects. After the analysis of the concealed objects characteristic, we design a combination feature extraction method, which is composed of the classical moment invariants, geometric parameters and objects space characteristic value. We use principal component analysis method for multidimensional data dimension reduction process. Finally, The BP neural network recognition technology is adopted to recognize concealed metal objects of human body images and good experimental result is obtained. Key words: X-rays; Concealed objects; Image processing; Feature extraction; Image recognition  PAGE 4 目 录  HYPERLINK \l _bookmark0 第一章 绪论1  HYPERLINK \l _bookmark1 1.1 课题背景1  HYPERLINK \l _bookmark2 1.2 X 射线背散射技术 2  HYPERLINK \l _bookmark3 1.2.1 X 射线背散射探测原理 2  HYPERLINK \l _bookmark4 1.2.2 X 射线背散射成像对人体检查的应用 2  HYPERLINK \l _bookmark5 1.3 国内外研究现状3  HYPERLINK \l _bookmark6 1.3.1 X 射线技术的研究现状 3  HYPERLINK \l _bookmark7 1.3.2 图像识别技术研究现状4  HYPERLINK \l _bookmark8 1.4 本文研究的主要内容及各章节安排5  HYPERLINK \l _bookmark9 第二章 人体图像预处理7  HYPERLINK \l _bookmark10