基于像素分割算法的超宽谱生物雷达目标识别定位技术研究  

Research on targets distinguishing and locating technique via UWB bioradar based on pixel segmentation algorithm

作　　者：马洋洋 于霄[1] 吕昊 李钊[1] 梁福来[1] 薛惠君 张华[1] 张杨[1] 

机构地区：[1]第四军医大学生物医学工程学院,西安710032

出　　处：《医疗卫生装备》2017年第7期1-5,共5页Chinese Medical Equipment Journal

基　　金：国家科技支撑计划(2014BAK12B02);国家自然科学基金(61327805);第四军医大学军事医学人才资助课题

摘　　要：目的:探索一维距离区分结合像素分割算法在多通道超宽谱生物雷达多目标识别定位中的可行性和适用条件。方法:对雷达接收到的信号先进行分解、重构、滤波,应用一维距离区分算法对多目标进行距离区分,然后采用像素分割算法,根据角度确定原理实现多目标的二维定位。最后征集10名志愿者,利用该方法进行了目标定位实验并给出结果。结果:实验说明像素分割算法对于单目标的定位效果较好,而对于多目标出现误判、漏判的几率则较大。结论:证明先采用一维距离区分、再采用基于像素分割的二维定位算法对3个以内的多静止目标进行定位是可行的,但对不同数量目标的定位精度不同。Objective To explore the feasibility and applicable conditions of one-dimensional distance distinction combining pixel segmentation algorithm in multi-target recognition and identification of multi-channel ultra wide-band bioradar. Methods The signals the radar received were decomposed, reconstructed and filtered, and one-dimensional distance distinction algorithm was applied to achieving multi-target distance discrimination, then multi-target two-dimensional positioning was achieved based on the principle of angle determination applying pixel segmentation algorithm. Finally, a target positioning experiment was executed by collecting 10 volunteers using the method above. Results The experiment indicated that pixel segmentation algorithm gained advantages when used for positioning a single target while disadvantages for multi target.Conclusion It＇s proved that it＇s feasible to locate three or less targets with one-dimensional distance distinction as well as two-dimensional locating based on pixel segmentation algorithm. Positioning accuracies are different in case of numbers of targets.

关 键 词：超宽谱 生物雷达 一维距离区分 像素分割 

分 类 号：TN957.51[电子电信—信号与信息处理]