高光谱技术在血迹分类识别中的应用  

Application of hyperspectral technology in classification and recognition of bloodstains

作　　者：李成成 赵明富 汤斌 罗彬彬 邹雪 王博思 LI Chengcheng;ZHAO Mingfu;TANG Bin;LUO Binbin;ZOU Xue;WANG Bosi(School of Electrical and Electronic Engineering,Chongqing University of Technology,Chongqing 400054,China;Chongqing Key Laboratory of Optical Fiber Sensor and Photodetector,Chongqing 400054,China;Chongqing Key Laboratory of Modern Photoelectric Detection Technology and Instrument,Chongqing 400054,China;China Merchants Chongqing Communications Technology Research & Designing Co.,Chongqing 400067,China;The 8th Research Institute of China Electronics Technology Group Corporation,Hefei Anhui 230000,China)

机构地区：[1]重庆理工大学电气与电子工程学院,重庆400054 [2]光纤传感与光电检测重庆市重点实验室,重庆400054 [3]重庆市现代光电检测技术与仪器重点实验室,重庆400054 [4]招商局重庆交通科研设计院有限公司,重庆400067 [5]中国电子科技集团公司第八研究所,安徽合肥230000

出　　处：《太赫兹科学与电子信息学报》2019年第3期489-494,共6页Journal of Terahertz Science and Electronic Information Technology

基　　金：国家自然科学基金资助项目(51276209);重庆市科委基础与前沿研究资助项目(cstc2014jcyj A0081)

摘　　要：利用高光谱技术对血迹种类进行无损识别研究。采用小波变换技术对 400~950 nm 之间的原始光谱进行去噪处理,并对处理后的光谱进行特征波段选择,建立全波段和特征波长下的血迹种类识别模型。结果表明,利用特征波长与支持向量机(SVM)结合建立的血迹种类识别模型的识别准确率及识别时间分别为 98%和 0.2 s,优于全波段建立的模型。研究表明,采用高光谱技术对血迹种类识别是可行的。Nondestructive identification of blood type is studied by hyperspectral technology in this paper. The original spectrum between 400-950 nm is denoised by wavelet transform, and then the characteristic spectrum of the processed spectrum is selected, finally the bloodstain type recognition models in the whole band and characteristic wavelengths are established. The results show that the model established in the characteristic wavelength combined with Support Vector Machine(SVM) can realize recognition accuracy of 98% and recognition time of 0.2 s, which is better than that of the model in full band. Research indicates that the hyperspectral technology is feasible for identifying the bloodstain types.

关 键 词：血迹 高光谱 分类 

分 类 号：TN911.74[电子电信—通信与信息系统]