一种运动恢复结构和航位推算结合的室内行人视觉定位方法

1 引言

随着智能移动终端的广泛普及,基于位置的服务（Location Based Service,LBS）^{[ 1]}与室内外一体化行人导航技术受到越来越多的关注。目前,利用GPS的定位导航已经能够满足室外行人的服务需求;而室内环境下由于接收到的GPS卫星信号差、多路径等影响,导致传统的定位方法无法直接使用,对于室内行人的定位及导航方法尚缺乏统一的标准。通常使用的室内定位方法可以分为2类：基于几何（Geometry）的定位和基于场景分析（Scene Analysis）的定位,使用的技术主要包括：无线传感网（Wireless Sensor Network, WSN）^{[ 2]}、无线局域网络（Wireless Local Area Networks, WLAN）^{[ 3]}、蓝牙^{[ 4]}、红外^{[ 5]}、射频识别（Radio Frequency Identification, RFID）^{[ 6]}等。其中,基于几何的定位方式很大程度上依赖于硬件设备,信号在传播过程中容易出现衰减等现象,导致定位结果存在系统偏差。因此,大多数室内定位的研究集中在基于场景位置的指纹匹配以及多传感器集成等方面。视觉定位^{[ 7]}是一种流行的定位技术,它具有成本低、无需外接设备、便捷性强、容易让人理解等优点。目前已开展关于视觉定位的研究^{[ 8- 11]},但是利用视觉进行室内定位仍然面临着建立（更新）位置指纹库费时费力、图像匹配结果不准确等困难。

行人航位推算（PDR）方法^{[ 12]}利用手机内置的陀螺仪、加速度计等传感器信息,可以实时计算行人的位置坐标,为室内多源信号位置指纹的自动标注提供了解决方案。但是,由于手机内置传感器的精度不高,航位角度计算容易产生累积误差,导致航位推算结果与真值偏差较大、不能直接用于位置指纹自动标注。为进一步提高航位推算位置精度,文献[ 13]提出了基于行为感知的行人室内定位方法,通过行人感知技术^{[ 14]}提取时间序列内的多个行为,结合室内路网拓扑结构^{[ 15]},匹配行为点到相应的特殊位置,从而消除了行人航位推算的累积误差。文献[16]采用UKF（Unscented Kalman Filter）融合Wi-Fi指纹匹配定位和PDR定位,得到了很好的室内定位结果。文献[17]针对多种环境下无线信号可以充分利用的特点,提出了一种泛在无线信号辅助的无缝定位方法,并对无缝定位技术的特点和发展趋势进行了讨论。文献[18]使用粒子滤波结合室内平面地图信息,对行人航位推算（PDR）结果进行修正,使室内行人的定位精度有所提高。上述这些室内定位系统或方法都在一定程度上提高了行人航位推算（PDR）的定位结果,但是往往需要添加额外的硬件设施或需要已知室内平面地图信息,降低了室内多源信号位置指纹自动标注的实际可操作性。文献[19]借助多个室内基站的实时数据,提出一种面向无线局域网的位置指纹动态自适应模型,该模型可以自适应更新室内位置指纹数据库,但是带来了更多的计算负荷。

本文基于智能手机平台实现了一种室内行人视觉定位方法,主要包括离线阶段的影像位置数据库自动标注和在线阶段的图像匹配定位,整个系统流程如图1所示。本文提出一种多约束图像匹配方法,该方法的匹配结果既用于查询影像与数据库影像间的匹配定位,也用于自动标注影像位置时的方位角计算。

显示原图| 下载原图ZIP| 生成PPT

图1 系统流程示意图

Fig. 1 The workflow of our system

2 多约束图像匹配及视觉定位

广义上图像匹配是指建立影像间最小空间分辨率单元一一对应关系的过程,是摄影测量、计算机视觉等领域的基础研究。图像匹配的结果可用于计算影像间的相对位姿^{[ 20]},也可直接反映图像间的相似程度^{[ 8- 11]}。一般情况下,2张影像的匹配点个数越多,其相似程度越高。本文提出一种多约束图像匹配算法,该算法在SIFT特征匹配的基础上增加了3种约束条件,可以剔除传统图像匹配中存在的大部分错误匹配结果,有助于恢复图像位姿和提高视觉定位的准确率。在定位阶段,使用本文方法计算查询影像与数据库影像间的匹配点数量,最多匹配点的K张影像位置坐标加权平均作为查询影像的定位结果。

目前,常用的图像匹配特征包括：Harris特征^{[ 21]}、SIFT特征^{[ 22]}、SURF特征^{[ 23]}等。其中,SIFT特征点具有尺度、旋转不变的特性,被广泛的用于图像拼接、影像跟踪等领域。利用特征点的高维特征描述子（特征向量）来衡量图像间的相似性是目前常用的方法,通常使用欧式距离、曼哈顿距离、夹角余弦等方式进行计算,然而仅仅使用距离最小作为相似性度量方式往往会发生错误的匹配结果（图2（a））。因此,本文以SIFT特征为例,相继使用了3种约束条件用于剔除图像匹配过程中产生错误的匹配点对,但是这3种约束并不局限于SIFT特征,所有基于向量的特征描述都可使用。具体约束条件如下：

显示原图| 下载原图ZIP| 生成PPT

图2 图像匹配结果示意图

Fig. 2 Schematic diagram of image matching

（1）比例约束。在计算特征点P₀的最佳匹配特征点时,假设与该点最近的2个特征点的欧式距离分别为d₁、d₂,欧式计算公式为：

$\begin{matrix} d_{i} = \sqrt{\overset{128}{\sum_{j = 1}} (v_{j} - v_{j}^{'})^{2}} \end{matrix}$ （1）

式中：v表示特征点的描述向量;j表示特征向量的维度。当最近匹配点与次近匹配点的距离比值小于阈值,即 $\begin{matrix} \frac{d_{1}}{d_{2}} < r 时 \end{matrix}$ ,则认为特征点P₀找到了一个最佳匹配点。使用比例约束能够消除那些特征距离较近的错误匹配。

（2）对称约束。使用比例约束后得到的匹配点对会出现一对多现象,一张图像上的特征点对应另一张图像上的多个特征点。为了消除这种错误匹配,再从另一张图像上的特征点出发,计算与它匹配的特征点,保留2个映射关系中一一对应的特征点对。

（3）RANSAC约束。RANSAC是指随机采样一致性^{[ 23]},其原理是从观测数据中随机挑选部分数据得出模型参数,使用模型参数验证其他数据是否符合,通过不断的选取观测数据得到最优模型参数的方法。在本方法中,首先选择4对匹配点计算得到单应矩阵H,其中：

$\begin{matrix} H = [\begin{matrix} h_{11} & h_{12} & h_{13} \\ h_{21} & h_{22} & h_{23} \\ h_{31} & h_{32} & h_{33} \end{matrix}] \end{matrix}$ （2）

使用H可以描述平移、旋转、仿射等坐标变换。如果一对匹配点的图像坐标为 $\begin{matrix} A (x_{1}, y_{1}) \end{matrix}$ , $\begin{matrix} B (x_{2}, y_{2}) \end{matrix}$ ,则通过单应矩阵二者的转换关系为：

$\begin{matrix} \{\begin{matrix} x_{2} = \frac{h_{11} x_{1} + h_{12} y_{1 +} h_{13}}{h_{33} x_{1} + h_{32} y_{1 +} 1} \\ y_{2} = \frac{h_{21} x_{1} + h_{22} y_{1 +} h_{23}}{h_{33} x_{1} + h_{32} y_{1 +} 1} \end{matrix} \end{matrix}$ （3）

本文利用单应矩阵计算匹配点坐标转换的误差值;通过不断的选择匹配点和计算误差值,得到一个最优的坐标转换关系H,使大部分匹配点以最小的坐标误差实现互相转换。剔除不满足最优转换关系的匹配点对,保留满足约束的匹配点对。

多约束匹配算法的伪代码描述如算法1,相继使用上述3种约束得到的图像匹配结果如图2（b）所示。相比于传统的匹配方法,本文提出的多约束图像匹配方法可以消除大部分的错误匹配,利用匹配点个数衡量图像间的相似性,可用于视觉定位。一般情况下,与查询影像最相似的数据库影像位置即查询影像的定位结果。为了增强视觉定位的鲁棒性,本文使用PDR结合KNN算法计算查询影像的室内位置坐标。首先使用PDR方法估计出行人的大致位置,再利用本文的多约束图像匹配方法计算查询影像 $\begin{matrix} Q \end{matrix}$ 与数据库影像 $\begin{matrix} P_{i} \end{matrix}$ 的匹配点个数 $\begin{matrix} N_{i} \end{matrix}$ ,降序排列 $\begin{matrix} N_{i} \end{matrix}$ 所对应的数据库影像,最后选取匹配点个数最多的前K个数据库影像,用加权平均的方法计算查询影像的位置坐标。

$\begin{matrix} C_{q} = \overset{k}{\sum_{i = 1}} w_{i} C_{i} \end{matrix}$ （4）

式中： $\begin{matrix} C_{q} \end{matrix}$ 是查询影像的位置坐标; $\begin{matrix} C_{i} \end{matrix}$ 是数据库影像的位置坐标; $\begin{matrix} w_{i} \end{matrix}$ 是权重因子,如式（5）所示。

$\begin{matrix} w_{i} = \frac{N_{i}}{\overset{k}{\sum_{i = 1}} N_{i}} \end{matrix}$ （5）

算法1 多约束匹配算法
input: image1, image2 input: r //比例约束参数 input: K //KNN参数 output: matches //匹配矩阵 detector->detect(image1,keypoints1); //SIFT算子计算特征点 detector->detect(image2,keypoints2); extractor->compute(image1,keypoints1,descriptors1); //计算特征点的描述子 extractor->compute(image2,keypoints2,descriptors2); matcher->knnMatch(descriptors1,descriptors2,matches1, 2); //KNN方法计算特征匹配 matcher->knnMatch(descriptors2,descriptors1,matches2, 2); removed=0; //删除匹配的点个数 removed=ratioTest(matches1);//比例约束
removed=ratioTest(matches2); vector<DMatch> symMatches; //对称约束 for matchIterator1= matches1.begin(); matchIterator1!= matches1.end(); ++matchIterator1 for matchIterator2= matches2.begin(); matchIterator2!= matches2.end(); ++matchIterator2 if ((matchIterator1)[0].queryIdx == (matchIterator2) [0].trainIdx && (matchIterator2)[0].queryIdx ==(matchItera tor1)[0]. trainIdx)) { symMatches.push_back((*matchIterator1)); } end if end for end for vector<DMatch> matches; ransacTest(symMatches,keypoints1, keypoints2, matches); //RANSAC约束

算法1 多约束匹配算法

input: image1, image2
input: r //比例约束参数
input: K //KNN参数
output: matches //匹配矩阵
detector->detect(image1,keypoints1); //SIFT算子计算特征点
detector->detect(image2,keypoints2);
extractor->compute(image1,keypoints1,descriptors1); //计算特征点的描述子
extractor->compute(image2,keypoints2,descriptors2);
matcher->knnMatch(descriptors1,descriptors2,matches1, 2); //KNN方法计算特征匹配
matcher->knnMatch(descriptors2,descriptors1,matches2, 2);
removed=0; //删除匹配的点个数
removed=ratioTest(matches1);//比例约束

removed=ratioTest(matches2);
vector<DMatch> symMatches; //对称约束
for matchIterator1= matches1.begin(); matchIterator1!=
matches1.end(); ++matchIterator1
for matchIterator2= matches2.begin(); matchIterator2!=
matches2.end(); ++matchIterator2
if ((*matchIterator1)[0].queryIdx == (*matchIterator2)
[0].trainIdx &&
(*matchIterator2)[0].queryIdx ==(*matchItera tor1)[0].
trainIdx))
{
symMatches.push_back((*matchIterator1));
}
end if
end for
end for
vector<DMatch> matches;
ransacTest(symMatches,keypoints1, keypoints2, matches); //RANSAC约束

3 影像位置自动标注

利用图像匹配的方法进行室内定位很大程度上依赖于影像位置数据库,而传统的建立影像位置数据库的方法是基于划分网格的人工标注方法,需要在每个坐标已知的网格中心点采集影像,其定位精度取决于网格划分的密度,建立高密度的位置影像数据库视需要花费大量的人力和时间开销。因此,本文提出自动标注影像位置的方法,利用视频输出的连续图像序列计算相邻图像间的运动方位角,将手机陀螺仪计算得到的方位角作为图像匹配失败时的补偿,利用行人航位推算（PDR）实现影像位置自动标注。

3.1 图像序列方位角计算

本文计算图像序列运动的方位角主要使用了从运动恢复结构（Structure From Motion, SFM）方法^{[ 24]},其基本原理是：利用双视几何中的极线约束,使用基础矩阵表达影像同名点间的对应关系,在已知相机内参的情况下,分解本质矩阵得到图像间的相对位置姿态。

如图3所示,C、C′为相机中心,X为三维空间中的一点,C、C′之间的连线被称为基线,x、x′分别为X在像平面I、I′中所成的像。则有图像点x、x′、空间点X和相机中心C、C′在同一个平面π内,平面与像平面的交线为I和I’被称为极线。可以看出,寻找图像上一点x在另一副图像上的同名点x′时,无需在整个图像上搜索,只需在极线l′上搜索即可,这种几何约束把同名点的搜索空间从二维降到了一维。

显示原图| 下载原图ZIP| 生成PPT

图3 图像双视几何关系

Fig. 3 Epipolar geometry of two images.

根据立体投影变换,空间一点 $\begin{matrix} X (X_{1}, Y_{1}, Z_{1})^{T} \end{matrix}$ 投影到图像平面上的点 $\begin{matrix} x (x_{1}, y_{1})^{T} \end{matrix}$ ,二者的坐标转换关系为：

$\begin{matrix} x = K [R | T] X = PX \end{matrix}$ （6）

式中：K为相机内参矩阵,通常情况下 $\begin{matrix} K = [\begin{matrix} f & 0 & cx \\ 0 & f & cy \\ 0 & 0 & 1 \end{matrix}] \end{matrix}$ ;f为相机焦距;cx、cy为相机光心在图像上的坐标;R为3×3旋转矩阵;T为3×1平移向量;P为投影矩阵。

根据双目视觉原理,可利用图像匹配点坐标和内参矩阵求得R和T,其中R和T分别体现当前图像相对与上一张图像的旋转和平移。2张图像间的相对运动角度可以通过进一步分解旋转矩阵得到,有如下关系：

$\begin{matrix} R = [\begin{matrix} cosθ & 0 & sinθ \\ sinφsinθ & cosφ & - sinφcosθ \\ - cosφsinθ & sinφ & cosφcosθ \end{matrix}] \end{matrix}$ （7）

式中： $\begin{matrix} θ \end{matrix}$ 是相邻2张图像的方向角; $\begin{matrix} φ \end{matrix}$ 是俯仰角。手机载体的三轴坐标方向如图4所示,绕Z轴旋转的角度为方向角,绕X轴旋转的角度为俯仰角,绕Y轴旋转的角度为横滚角。本文忽略横滚角的影响,主要利用相邻2张图像之间的方向角估计运动轨迹中的方向变化。假设第一张图像的坐标方位角为 $\begin{matrix} θ_{i} \end{matrix}$ ,则第k张图像的坐标方位角为 $\begin{matrix} A_{k} = \overset{k}{\sum_{i = 1}} θ_{i} \end{matrix}$ ,如图5所示。

显示原图| 下载原图ZIP| 生成PPT

图4 手机内置三轴方向

Fig. 4 The three-axis direction of smartphone

显示原图| 下载原图ZIP| 生成PPT

图5 连续图像方位角计算原理

Fig. 5 The principle of azimuth angle from consecutive images

3.2 图像序列位置标注

由于手机内置惯性传感器的精度不高,导致行人航位推算（PDR）方法得到的标注位置与真实值存在偏差,并且随着时间的推移累积误差会越来越大,其主要误差源是方向角度值的漂移。本文利用图像序列计算得到的角度作为方向角,使用加速度波峰检测算法^{[ 25]}探测行走步数、步态检测算法^{[ 26]}估计行走步长。在已知起始坐标的情况下,利用PDR方法可以标注行走轨迹上的图像序列坐标（式（8））。

$\begin{matrix} \{\begin{matrix} x_{k} = x_{k - 1} + S ∙ \sin θ_{k} \\ y_{k} = y_{k - 1} + S ∙ \cos θ_{k} \end{matrix} \end{matrix}$ （8）

式中：（ $\begin{matrix} x_{k}, y_{k} \end{matrix}$ ）是第k张图像的平面坐标; $\begin{matrix} θ_{k} \end{matrix}$ 是利用图像序列计算得到的方向角,S是估计距离。本文采用文献[28]所述的步态检测方法,检测波峰计算每一步的起止时刻,使用线性模型来确定行走过程中的步长。本文基于图像序列的位置标注流程如图6所示,对采集得到的视频按照等时间间隔输出图像序列 $\begin{matrix} p_{i} \end{matrix}$ ,分别计算图像序列的运动方向角 $\begin{matrix} θ_{i} \end{matrix}$ 和陀螺仪积分方向角 $\begin{matrix} α_{j} \end{matrix}$ ;由于传感器采样频率不同,陀螺仪的采样频率高于图像采样频率。当检测到相邻两张图像间的匹配点个数过少无法计算 $\begin{matrix} θ_{i} \end{matrix}$ 时,使用相同时刻下的陀螺仪瞬时积分角度作为补偿。

$\begin{matrix} θ_{i} = \overset{T_{i 2}}{\sum_{T_{i 1}}} α_{j} \end{matrix}$ （9）

式中： $\begin{matrix} T_{i 1} \end{matrix}$ 表示第i张采样图像所对应的时刻, $\begin{matrix} T_{i 2} \end{matrix}$ 表示第i+1张采样图像所对应的时刻。步子检测和步长约束代入到式（8）得到图像序列的位置坐标。

显示原图| 下载原图ZIP| 生成PPT

图6 图像序列位置标注流程图

Fig. 6 The workflow of calibration fingerprint of the image series

4 实验分析

4.1 实验场景

为了验证本文提出的方法,分别选取了4个实验区域进行了验证。图7（a）为深圳大学科技楼第14层,该层主要由走廊区域构成,区域面积为52.5 m×52.5 m; 图7（b）为深圳大学科技楼第 1层,该层存在较多开阔区域,行人行走自由度高,楼层面积为106.3 m×61.5 m; 图7（c）为深圳大学教学楼,主要由走廊区域构成; 图7（d）为深圳南山医院第2层,区域面积为62 m × 60 m。实验流程如图8所示,首先在每个实验区域分别采集运动轨迹数据,使用本文方法估计航向角度和轨迹位置,构建室内影像数据库;利用在已知坐标点位置拍摄的查询影像计算在线图像匹配定位误差;最后对实验结果进行分析。本文以三星GALAXY Note III（Android 4.3平台）型号智能手机为平台开发了实验原型系统,建立影像数据库过程中,实验者持手机沿着指定采集路线行走并保持手机与身体不发生相对移动;在线匹配定位阶段使用客户端-服务器模式,首先由用户上传采集到的查询影像到服务器,服务器进行匹配定位后再将位置结果发送给用户。

显示原图| 下载原图ZIP| 生成PPT

图7 实验区域概况

Fig. 7 Layout of the test area

显示原图| 下载原图ZIP| 生成PPT

图8 实验流程图

Fig. 8 The workflow of experiments

4.2 图像运动角度估计试验

图像运动角度估计是影像序列位置标注的核心问题,高精度的图像运动角度估计是构建室内位置影像数据库的基础。本文提出的多约束图像匹配方法可以剔除大部分的错误匹配,匹配结果用于估计图像运动角度可得到准确的图像运动方位角。为验证本文方法的准确性,选择上述实验区域内的固定位置,使用带刻度的360°旋转云台固定手机拍摄图像序列。保证手机沿着一个方向旋转拍摄影像,分别在固定间隔5°、10°、15°、20°下记录摄序列影像和陀螺仪数据,使用本文方法计算旋转角度和陀螺仪角度,分别统计角度误差及误差的方差。从误差分布的实验结果（图9）可看出,本文利用图像计算角度的方法得到的角度误差均值约为0.5°,最大误差不超过1°。本文算法得到运动方位角相比陀螺仪计算出的角度具有更高的精度和稳定性。

显示原图| 下载原图ZIP| 生成PPT

图9 角度误差分布

Fig. 9 The results of motion angle errors

4.3 运动轨迹标注试验

实验者沿着上述4条路径行走并采集手机惯导数据和视频数据,利用本文方法处理可得到图像序列的位置坐标,在构建室内影像位置数据库的同时,每张采样影像还包含了其他定位信号强度信息如Wi-Fi、地磁、气压等,即构建了多源信号的位置指纹数据库。在数据处理过程中,本文使用0.5 s的时间采样间隔从运动视频输出图像序列,计算相同采样频率下的陀螺仪积分角度,分别使用图像序列角度和陀螺仪积分角度恢复行走轨迹、计算位置标注误差,得到标注位置的累积误差分布（Cumulative Distribution Function, CDF）,如图10所示。

显示原图| 下载原图ZIP| 生成PPT

图10 位置标注累积误差曲线

Fig. 10 The cumulative distribution function of calibrated fingerprints

可以看出,使用本文方法得到的标注点累积误差明显限小于使用陀螺仪的方法。使用图像序列计算得到的角度代入PDR方法得到的标注误差80%小于1.3 m,平均标注误差为0.58 m。使用本文方法得到的影像位置标注最大误差不超过2.5 m,最小误差小于0.2 m,标注误差最大点一般发生在有较大转弯运动的位置。位置标注后的影像序列可以作为影像位置数据库,用于视觉定位。

4.4 视觉定位试验

本文提出的多约束图像匹配方法可以剔除大部分的错误匹配结果,相比文献[21]-[23]在像素级别上提高了匹配精度。目前的图像匹配研究很少从像素匹配拓展到位置匹配,文献[8]、[11]利用图像匹配技术进行室内定位的研究。为了验证本文方法的正确性,在标注后的影像轨迹中选择了106个已知位置拍摄查询图像,使用本文提出的图像匹配定位方法计算查询影像的位置坐标,量化的视觉定位结果如表1所示。

表1 视觉匹配定位结果

Tab. 1 The results of image matching positioning

匹配指标	匹配正确率/%	平均定位误差	时间开销/s
本文方法	94	0.6 m	3
文献[8]方法	80	房间级别	4.1
文献[11]方法	94	1 m	10-12

新窗口打开

从表1可看出,使用本文方法进行视觉匹配定位的平均定位误差为0.6 m,匹配正确率为94%。其中,本文方法的最大定位误差为1.4 m,发生匹配错误的原因在于影像数据库中存在图像内容上相似的不同位置,利用运动跟踪的方法可进一步提高匹配正确率。与其他室内视觉定位方法相比,文献[8]的方法定位精度只能识别出不同的房间,对于开阔区域无法给出量化的定位结果,原因是该方法在图像匹配阶段没有剔除错误的图像匹配点。文献[11]具有较高的匹配正确率,但是需要额外的激光扫描设施构建室内三维模型,计算复杂度高。

算法的时间开销与机器性能、图像分辨率、初始搜索位置等因素有关。文献[8]在客户端-服务器模式下,传输大小为128 k低分辨率图像,在历史轨迹约束情况下匹配10张最近的数据库影像的时间开销为4.2 s。文献[11]在笔记本电脑上解算一张约100万像素的查询图像平均需要10~12 s,时间开销较大的原因在于该方法不仅计算了图像的平面位置,还计算出图像的三维姿态信息。本文方法的原型系统在台式机上实施,匹配一张约600万像素的查询影像平均时间为3 s,使用GPU加速的方法可进一步降低匹配时间开销至1 s以内。利用历史轨迹约束的方法匹配10张最近邻影像,计算得到查询影像位置的时间少于10 s。如何加速图像匹配、提高系统实时性能也是本文将要继续研究的问题之一。

5 结论

在室内环境下,本文通过采集手机惯导数据和视频数据,进行了自动构建影像位置数据库和视觉定位的试验。利用一种多约束图像匹配方法剔除大部分的错误匹配点,匹配结果用于SFM方法可以得到图像序列的运动方位角,利用PDR方法计算图像序列位置坐标、构建图像位置数据库;图像匹配结果可直接用于室内定位。通过真实场景试验,本文方法对室内采样点的位置标注误差不超过2.5 m,图像匹配定位精度能够满足导航需求。该方法可以减少影像匹配室内定位时建库、更新劳力开销,还可用于多源定位信号的室内指纹自动标注。

The authors have declared that no competing interests exist.

参考文献

原文顺序

文献年度倒序

文中引用次数倒序

被引期刊影响因子

[1]	Chen X H, Pang J. Protecting query privacy in location based services [J]. GeoInformatica, 2014,18(1):95-133. https://doi.org/10.1007/s10707-013-0192-0 URL Magsci [本文引用: 1] 摘要 The popularity of location-based services (LBSs) leads to severe concerns on users’ privacy. With the fast growth of Internet applications such as online social networks, more user information becomes available to the attackers, which allows them to construct new contextual information. This gives rise to new challenges for user privacy protection and often requires improvements on the existing privacy-preserving methods. In this paper, we classify contextual information related to LBS query privacy and focus on two types of contexts— user profiles and query dependency : user profiles have not been deeply studied in LBS query privacy protection, while we are the first to show the impact of query dependency on users’ query privacy. More specifically, we present a general framework to enable the attackers to compute a distribution on users with respect to issuing an observed request. The framework can model attackers with different contextual information. We take user profiles and query dependency as examples to illustrate the implementation of the framework and their impact on users’ query privacy. Our framework subsequently allows us to show the insufficiency of existing query privacy metrics, e.g., k -anonymity, and propose several new metrics. In the end, we develop new generalisation algorithms to compute regions satisfying users’ privacy requirements expressed in these metrics. By experiments, our metrics and algorithms are shown to be effective and efficient for practical usage.
[2]	Kwon O H, Song H J, Park S. The effects of stitching orders in Patch-and-stitch WSN localization algorithms [J]. IEEE Transactions on Parallel and Distributed Systems, 2009, 20(9): 1380-1391. https://doi.org/10.1109/TPDS.2008.226 URL [本文引用: 1] 摘要 A "patch-and-stitchrdquo localization algorithm divides the network into small overlapping subregions. Typically, each subregion consists of a node and all or some of its neighbors. For each subregion, the algorithm builds a local map, called a patch, which is actually an embedding of the nodes it spans in a relative coordinate system. Finally, the algorithm stitches those patches to form a single global map. In a patch-and-stitch algorithm, the stitching order makes an influence on both the performance and the complexity of the algorithm. In this paper, we present a formal framework to deal with stitching orders in patch-and-stitch localization algorithms. In our framework, the stitching order is determined by a stitching scheme and the stitching scheme consists of a stitching policy and a potential function. The potential function is to predict how well a patch will be stitched if patches are stitched according to a given partial order. The stitching policy is a mechanism that determines the stitching order based on the predictions by the potential function. We present various stitching schemes and evaluate them through simulations. In addition, we apply the patch-and-stitch strategy into the anchor-based localization and propose a clustering-based localization algorithm. A potential function is used to partition the network into clusters each of which is centered at an anchor node. For each cluster, a cluster map is constructed via the anchor-free localization algorithm. Then, those cluster maps are combined to form a single global map. We propose a stitching technique for combining those cluster maps and analyze the performance of the algorithm by simulations.
[3]	Fang S H, Lin T N, Lee K C. A novel algorithm for multipath fingerprinting in indoor WLAN environments [J]. IEEE Transactions on Wireless Communications, 2008,7(9):3579-3588. https://doi.org/10.1109/TWC.2008.070373 URL [本文引用: 1] 摘要 Positioning in indoor wireless environments is growing rapidly in importance and gains commercial interests in context-awareness applications. The essential challenge in localization is the severe fluctuation of receive signal strength (RSS) for the mobile client even at a fixed location. This work explores the major noisy source resulted from the multipath in an indoor wireless environment and presents an advanced positioning architecture to reduce the disturbance. Our contribution is to propose a novel approach to extract the robust signal feature from measured RSS which is provided by IEEE 802.11 MAC software so that the multipath effect can be mitigated efficiently. The dynamic multipath behavior, which can be modeled by a convolution operation in the time domain, can be transformed into an additive random variable in the logarithmic spectrum domain. That is, the convolution process becomes a linear and separable operation in the logarithmic spectrum domain and then can be effectively removed. To our best knowledge, this work is the first to enhance the robustness to a multipath fading condition, which is common in the environments of an indoor wireless LAN (WLAN) location fingerprinting system. Our approach is conceptually simple and easy to be implemented for practical applications. Neither a new hardware nor an extra sensor network installation is required. Both analytical simulation and experiments in a real WLAN environment demonstrate the usefulness of our approach to significant performance improvements. The numerical results show that the mean and the standard deviation of estimated error are reduced by 42% and 29%, respectively, as compared to the traditional maximum likelihood based approach. Moreover, the experimental results also show that fewer training samples are required to build the positioning models. This result can be attributed to that the location related information is effectively extracted by our algorithm.
[4]	Kotanen A, HÄNNIKÄINEN M, LEPPÄKOSKI H, HÄMÄLÄINEN T. D. Experiments on local positioning with Bluetooth [C]// IEEE International Conference on Information Technology: Coding and Computing [Computers and Communications], ITCC, 2003:297-303. [本文引用: 1]
[5]	Harter A, Hopper A. A distributed location system for the active office [J]. IEEE Network, 1994,8(1):62-70. https://doi.org/10.1109/65.260080 URL [本文引用: 1] 摘要 Distributed systems for locating people and equipment will be at the heart of tomorrow's active offices. Computer and communications systems continue to proliferate in the office and home. Systems are varied and complex, involving wireless networks a...
[6]	Ni L M, Liu Y, Lau Y C, et al. LANDMARC: Indoor location sensing using active RFID [J]. Wireless networks, 2004,10(6):701-710. https://doi.org/10.1023/B:WINE.0000044029.06344.dd URL Magsci [本文引用: 1] 摘要 Growing convergence among mobile computing devices and embedded technology sparks the development and deployment of "context-aware" applications, where location is the most essential context. In this paper we present LANDMARC, a location sensing prototype system that uses Radio Frequency Identification (RFID) technology for locating objects inside buildings. The major advantage of LANDMARC is that it improves the overall accuracy of locating objects by utilizing the concept of reference tags. Based on experimental analysis, we demonstrate that active RFID is a viable and cost-effective candidate for indoor location sensing. Although RFID is not designed for indoor location sensing, we point out three major features that should be added to make RFID technologies competitive in this new and growing market.
[7]	Schroth G, Huitl R, Chen D, et al. Mobile visual location recognition [J]. IEEE Signal Processing Magazine, 2011, 28(4):77-89. https://doi.org/10.1109/MSP.2011.940882 Magsci [本文引用: 1]
[8]	Ravi N, Shankar P, Frankel A, et al. Indoor localization using camera phones [C]//Mobile Computing Systems and Applications, 2006. WMCSA'06. Proceedings. 7^th IEEE Workshop on. IEEE, 2006:49-49. [本文引用: 2]
[9]	Liu T, Li Q, Zhang X. Automatic construction of Wi-Fi radio map using smartphones [J]. ISPRS-International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2016,XLI-B4:309-312. https://doi.org/10.5194/isprsarchives-XLI-B4-309-2016 URL 摘要 Abstract Indoor positioning could provide interesting services and applications. As one of the most popular indoor positioning methods, location fingerprinting determines the location of mobile users by matching the received signal strength (RSS) which is location dependent. However, fingerprinting-based indoor positioning requires calibration and updating of the fingerprints which is labor-intensive and time-consuming. In this paper, we propose a visual-based approach for the construction of radio map for anonymous indoor environments without any prior knowledge. This approach collects multi-sensors data, e.g. video, accelerometer, gyroscope, Wi-Fi signals, etc., when people (with smartphones) walks freely in indoor environments. Then, it uses the multi-sensor data to restore the trajectories of people based on an integrated structure from motion (SFM) and image matching method, and finally estimates location of sampling points on the trajectories and construct Wi-Fi radio map. Experiment results show that the average location error of the fingerprints is about 0.53 m.
[10]	Ruotsalainen L, Kuusniemi H, Bhuiyan M Z H, et al. A two-dimensional pedestrian navigation solution aided with a visual gyroscope and a visual odometer [J]. Gps Solutions, 2013,17(4):575-586. https://doi.org/10.1007/s10291-012-0302-8 URL Magsci 摘要 Integration of different positioning systems such as wireless local area networks (WLAN) and INS has proven to be the most promising approach for navigation in global navigation satellite system challenging environments. However, the integrated solution suffers from the errors induced by the sensors in the INS and the low availability and infrastructure dependency of the WLAN. Visual aiding is a complementary method for augmenting these systems, because it suffers from errors and availability problems of a different nature. We introduce the concept of a “visual gyroscope” and a “visual odometer,” based on recovering user information by tracking the feature motion between consecutive images. All calculations are of sufficiently low complexity to be adoptable for navigation with current smartphones. The camera orientation is observed using the vanishing point locations, and this information is transformed into user heading change information and also used for visual odometer calculations. The visual odometer retrieves the camera translation information based on a special camera configuration and the motion of the matched scale-invariant feature transform features. The performance of both of these tools is evaluated. Experiments in selected environments have proven that visual aiding with a visual gyroscope and visual odometer improves the two-dimensional navigation solution significantly.
[11]	Liang J Z, Corso N, Turner E, et al. Image based localization in indoor environments [C]// IEEE International Conference on Computing for Geospatial Research and Application (COM. Geo), 2013:70-75. [本文引用: 2]
[12]	Beauregard S, Haas H. Pedestrian dead reckoning: A basis for personal positioning [C]//Proceedings of the 3rd Workshop on Positioning, Navigation and Communication. 2006:27-35. [本文引用: 1]
[13]	周宝定,李清泉,毛庆洲,等. 用户行为感知辅助的室内行人定位 [J].武汉大学学报·信息科学版,2014,39(6):719-723. https://doi.org/10.13203/j.whugis20140100 URL [本文引用: 1] 摘要提出了一种用户行为感知辅助的室内行人定位方法。该方法采用行为感知技术，通过智能手机内置的传感器获取行人经过室内特殊位置（如电梯、转角等）的连续多个位置相关的行为状态，结合室内路网的拓扑结构，将行人的位置匹配到相应的特殊位置，以消除行人航位推算（PDR）的累积误差，同时可以在起始点未知的情况下实现行人的自主定位。试验验证了该方法的有效性。 [ Zhou B D,Li Q Q,Mao Q Z,et al. User activity awareness assisted indoor pedestrian localization [J]. Geomatics and Information Science of Wuhan Univers, 2014,39(6):719-723. ] https://doi.org/10.13203/j.whugis20140100 URL [本文引用: 1] 摘要提出了一种用户行为感知辅助的室内行人定位方法。该方法采用行为感知技术，通过智能手机内置的传感器获取行人经过室内特殊位置（如电梯、转角等）的连续多个位置相关的行为状态，结合室内路网的拓扑结构，将行人的位置匹配到相应的特殊位置，以消除行人航位推算（PDR）的累积误差，同时可以在起始点未知的情况下实现行人的自主定位。试验验证了该方法的有效性。
[14]	Zhou B, Li Q, Mao Q, et al. Activity sequence-based indoor pedestrian localization using smartphones [J]. IEEE Transactions on Human-Machine Systems, 2015,45(5):562-574. https://doi.org/10.1109/THMS.2014.2368092 URL [本文引用: 1] 摘要 This paper presents an activity sequence-based indoor pedestrian localization approach using smartphones. The activity sequence consists of several continuous activities during the walking process, such as turning at a corner, taking the elevator, taking the escalator, and walking stairs. These activities take place when a user walks at some special points in the building, like corners, elevators, escalators, and stairs. The special points form an indoor road network. In our approach, we first detect the user's activities using the built-in sensors in a smartphone. The detected activities constitute the activity sequence. Meanwhile, the user's trajectory is reckoned by Pedestrian Dead Reckoning (PDR). Based on the detected activity sequence and reckoned trajectory, we realize pedestrian localization by matching them to the indoor road network using a Hidden Markov Model. After encountering several special points, the location of the user would converge on the true one. We evaluate our proposed approach using smartphones in two buildings: an office building and a shopping mall. The results show that the proposed approach can realize autonomous pedestrian localization even without knowing the initial point in the environments. The mean offline localization error is about 1.3 m. The results also demonstrate that the proposed approach is robust to activity detection error and PDR estimation error.
[15]	刘涛, 张星,李清泉,等. 顾及地标可视性的室内导航路径优化算法 [J].武汉大学学报·信息科学版,2017,42(1):43-48. https://doi.org/10.13203/j.whugis20150387 URL [本文引用: 1] 摘要提出了一种构建室内行人通行网络的方法,利用矢量建筑图自动构建室内建筑、地标的可视关系,建立行人导航通行规则,支持室内导航路径规划。实验结果表明,此方法能够有效描述室内行人通行规则,并满足拓扑网络构建的实时性需求,减少大规模存储与维护室内路网的压力。在此基础上提出了一种多目标导航路径优化算法,该算法时间开销较低,能够实时地进行路径规划,得到的最优路径与最短路径相比具有更高的地标可见性和覆盖率。 [ Liu T, Zhang X, Li Q Q, et al. An indoor pedestrian route planning algorithm based on landmark visibility [J]. Geomatics and Information Science of Wuhan Univers, 2017,42(1):43-48. ] https://doi.org/10.13203/j.whugis20150387 URL [本文引用: 1] 摘要提出了一种构建室内行人通行网络的方法,利用矢量建筑图自动构建室内建筑、地标的可视关系,建立行人导航通行规则,支持室内导航路径规划。实验结果表明,此方法能够有效描述室内行人通行规则,并满足拓扑网络构建的实时性需求,减少大规模存储与维护室内路网的压力。在此基础上提出了一种多目标导航路径优化算法,该算法时间开销较低,能够实时地进行路径规划,得到的最优路径与最短路径相比具有更高的地标可见性和覆盖率。
[16]	陈国良,张言哲,汪云甲,等. WiFi-PDR 室内组合定位的无迹卡尔曼滤波算法 [J].测绘学报,2015,44(12):1314-1321. https://doi.org/10.11947/j.AGCS.2015.20140691 URL 摘要针对当前室内定位的应用需求和亟待解决的关键问题,结合城市室内环境下广泛存在的WiFi无线信号以及智能手机传感器信息,提出了一种WiFi无线信号联合行人航迹推算(PDR)的室内定位方法.该方法采用无迹卡尔曼滤波(UKF)算法对WiFi和PDR定位信息进行融合处理,有效克服了WiFi单点定位精度低和PDR存在累计误差的问题.针对融合算法中WiFi指纹匹配计算量大的问题,用k-means聚类算法对WiFi指纹库进行聚类处理,降低了指纹匹配算法的计算量,提高了算法的实时性.通过在华为P6-U06智能手机平台上实际测试,在时间效率上经过聚类处理后系统定位耗时有很大程度的改善,平均降幅为51％,其中最大降幅达到64％,最小的也达到了36％;在定位精度上,当室内人员为行走状态时WiFi定位平均误差为7.76 m,PDR定位平均误差为4.57 m,UKF滤波融合后平均定位误差下降到1.24 m. [ Chen G L, Zhang Y Z, Wang Y J, et al. Unscented kalman filter algorithm for WiFi-PDR integrated indoor positioning [J]. Acta Geodaetica et Cartographica Sinica, 2015,44(12):1314-1321. ] https://doi.org/10.11947/j.AGCS.2015.20140691 URL 摘要针对当前室内定位的应用需求和亟待解决的关键问题,结合城市室内环境下广泛存在的WiFi无线信号以及智能手机传感器信息,提出了一种WiFi无线信号联合行人航迹推算(PDR)的室内定位方法.该方法采用无迹卡尔曼滤波(UKF)算法对WiFi和PDR定位信息进行融合处理,有效克服了WiFi单点定位精度低和PDR存在累计误差的问题.针对融合算法中WiFi指纹匹配计算量大的问题,用k-means聚类算法对WiFi指纹库进行聚类处理,降低了指纹匹配算法的计算量,提高了算法的实时性.通过在华为P6-U06智能手机平台上实际测试,在时间效率上经过聚类处理后系统定位耗时有很大程度的改善,平均降幅为51％,其中最大降幅达到64％,最小的也达到了36％;在定位精度上,当室内人员为行走状态时WiFi定位平均误差为7.76 m,PDR定位平均误差为4.57 m,UKF滤波融合后平均定位误差下降到1.24 m.
[17]	田辉,夏林元,莫志明,等. 泛在无线信号辅助的室内外无缝定位方法与关键技术 [J].武汉大学学报·信息科学版,2009,34(11):1372-1376. URL 摘要针对无缝定位的应用需求和亟待解决的关键技术问题,结合城市环境下多种泛在无线信号可以充分利用的特点,提出了一种泛在无线信号辅助的无缝定位新方法,并对无缝定位技术的原理、特点和发展趋势进行了讨论。特别针对香港典型城市环境,就多种泛在无线信号辅助的无缝定位进行了初步实验,给出了一个能够满足大多数无缝定位服务应用的原型框架。 [ Tian H, Xia L Y, Esmond M, et al. Signal of opportunity assisted ubiquitous positioning and its key elements for outdoor/indoor environment [J]. Geomatics and Information Science of Wuhan University, 2009,34(11):1372-1376. ] URL 摘要针对无缝定位的应用需求和亟待解决的关键技术问题,结合城市环境下多种泛在无线信号可以充分利用的特点,提出了一种泛在无线信号辅助的无缝定位新方法,并对无缝定位技术的原理、特点和发展趋势进行了讨论。特别针对香港典型城市环境,就多种泛在无线信号辅助的无缝定位进行了初步实验,给出了一个能够满足大多数无缝定位服务应用的原型框架。
[18]	Evennou F, Marx F, Novakov E. Map-aided indoor mobile positioning system using particle filter [C]//Wireless Communications and Networking Conference, 2005 IEEE. IEEE, 2005,4:2490-2494.
[19]	吴东金,夏林元. 面向室内WLAN定位的动态自适应模型 [J].测绘学报,2015,44(12):1322-1330. https://doi.org/10.11947/j.AGCS.2015.20130780 URL 摘要提出了一种面向无线局域网（wireless local area network,WLAN）位置指纹匹配定位的动态自适应模型,借助多个基站的实时数据为稳健室内定位提供更新的匹配数据库——radio map。基于基站与radio map采样点之间的空间关联性,将基站数据和radio map分别作为多层神经网络的输入和输出,动态更新radio map;利用多元数据异常探测技术检验基站数据捕获环境的时空变化;根据探测结果采用顾及室内布局的数值内插和再训练的方式更新模型,从而使其适应环境的时空变化。在室内动态环境中进行了多次验证试验。试验结果显示,在时变因素作用下,相比较传统方法,采用所提模型的定位方法的平均误差至少下降10%;在空间变化因素（以信标移动为例）作用下,其他方法平均定位误差大幅增加了至少165%,而采用所提模型方法的平均定位误差只增加了10%~20%,定位精度维持在3m左右（即原始精度）。结果证明采用了所提模型的定位系统能够自适应环境的时空变化而保持原有定位表现。不过,与传统位置指纹匹配定位方法相比,该模型带来了更多的计算负荷。 [ Wu D J, Xia L Y. Dynamic adaptive model for indoor WLAN localization [J]. Acta Geodaetica et Cartographica Sinica, 2015,44(12):1322-1330. ] https://doi.org/10.11947/j.AGCS.2015.20130780 URL 摘要提出了一种面向无线局域网（wireless local area network,WLAN）位置指纹匹配定位的动态自适应模型,借助多个基站的实时数据为稳健室内定位提供更新的匹配数据库——radio map。基于基站与radio map采样点之间的空间关联性,将基站数据和radio map分别作为多层神经网络的输入和输出,动态更新radio map;利用多元数据异常探测技术检验基站数据捕获环境的时空变化;根据探测结果采用顾及室内布局的数值内插和再训练的方式更新模型,从而使其适应环境的时空变化。在室内动态环境中进行了多次验证试验。试验结果显示,在时变因素作用下,相比较传统方法,采用所提模型的定位方法的平均误差至少下降10%;在空间变化因素（以信标移动为例）作用下,其他方法平均定位误差大幅增加了至少165%,而采用所提模型方法的平均定位误差只增加了10%~20%,定位精度维持在3m左右（即原始精度）。结果证明采用了所提模型的定位系统能够自适应环境的时空变化而保持原有定位表现。不过,与传统位置指纹匹配定位方法相比,该模型带来了更多的计算负荷。
[20]	Sturm P, Triggs B. A factorization based algorithm for multi-image projective structure and motion [M]//Computer Vision-ECCV'96. Springer Berlin Heidelberg, 1996:709-720. [本文引用: 1]
[21]	Harris C, Stephens M. A combined corner and edge detector [C]//Alvey vision conference. 1988. [本文引用: 1]
[22]	Lowe D G. Distinctive image features from scale-invariant keypoints [J]. International journal of computer vision, 2004,60(2):91-110. https://doi.org/10.1023/B:VISI.0000029664.99615.94 Magsci [本文引用: 1] 摘要 <a name="Abs1"></a>This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are shown to provide robust matching across a substantial range of affine distortion, change in 3D viewpoint, addition of noise, and change in illumination. The features are highly distinctive, in the sense that a single feature can be correctly matched with high probability against a large database of features from many images. This paper also describes an approach to using these features for object recognition. The recognition proceeds by matching individual features to a database of features from known objects using a fast nearest-neighbor algorithm, followed by a Hough transform to identify clusters belonging to a single object, and finally performing verification through least-squares solution for consistent pose parameters. This approach to recognition can robustly identify objects among clutter and occlusion while achieving near real-time performance.
[23]	Bay H, Tuytelaars T, Van Gool L. Surf: Speeded up robust features [M]//Computer vision-ECCV 2006. Springer Berlin Heidelberg, 2006:404-417. [本文引用: 2]
[24]	Fischler M A, Bolles R C. Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography [J]. Communications of the ACM, 1981,24(6):381-395. https://doi.org/10.1145/358669.358692 URL [本文引用: 1] 摘要 We introduce a new paradigm, Random Sample Consensus (RANSAC),for fitting a model to experimental data; and illustrate its use in scene analysis and auto- mated cartography. The application discussed, the loca- tion determination problem (LDP), is treated at a level beyond that of a mere example of the use of the RANSAC paradigm; new basic findings concerning the conditions under which the LDP can be solved are presented and a comprehensive approach to the solution of this problem that we anticipate will have near-term practical appli- cations is described.To a large extent, scene analysis (and, in fact, science in general) is concerned with the interpretation of sensed data in terms of a set of predefmed models. Conceptually, interpretation involves two distinct activities: First, there is the problem of finding the best match between the data and one of the available models (the classification problem); Second, there is the problem of computing the best values for the free parameters of the selected model (the parameter estimation problem). In practice, these two problems are not independent--a solution to the parameter estimation problem is often required to solve the classification problem.Classical techniques for parameter estimation, such as least squares, optimize (according to a specified ob- jective function) the fit of a functional description (model) to all of the presented data. These techniques have no internal mechanisms for detecting and rejecting gross errors. They are averaging techniques that rely on the assumption (the smoothing assumption) that the maximum expected deviation of any datum from the assumed model is a direct function of the size of the data set, and thus regardless of the size of the data set, there will always be enough good values to smooth out any gross deviations.In many practical parameter estimation problems the smoothing assumption does not hold; i.e., the data con- tain uncompensated gross errors. To deal with this situ- ation, several heuristics have been proposed. The tech- nique usually employed is some variation of first using all the data to derive the model parameters, then locating the datum that is farthest from agreement with the instantiated model, assuming that it is a gross error, deleting it, and iterating this process until either the maximum deviation is less then some preset threshold or until there is no longer sufficient data to proceed.It can easily be shown that a single gross error ("poisoned point"), mixed in with a set of good data, can
[25]	Faugeras O D, Lustman F. Motion and structure from motion in a piecewise planar environment [J]. International Journal of Pattern Recognition and Artificial Intelligence, 1988,2(3):485-508. https://doi.org/10.1142/S0218001488000285 URL [本文引用: 1] 摘要 We show in this article that when the environment is piecewise linear, it provides a powerful constraint on the kind of matches that exist between two images of the scene when the camera motion is unknown. For points and lines located in the same plane, the correspondence between the two cameras is a collineation. We show that the unknowns (the camera motion and the plane equation) can be recovered, in general, from an estimate of the matrix of this collineation. The two-fold ambiguity that remains can be removed by looking at a second plane, by taking a third view of the same plane, or by using a priori knowledge about the geometry of the plane being looked at. We then show how to combine the estimation of the matrix of collineation and the obtaining of point and line matches between the two images, by a strategy of Hypothesis Prediction and Testing guided by a Kalman filter. We finally show how our approach can be used to calibrate a system of cameras.
[26]	Golub G H, Reinsch C. Singular value decomposition and least squares solutions [J]. Numerische mathematik, 1970,14(5):403-420. https://doi.org/10.1007/978-3-662-39778-7_10 URL [本文引用: 1] 摘要 Let A be a real m × n matrix with m ≧ n. It is well known (cf. [4]) that $$A = U\sum {{V^T}} $$ (1) where $${U^T}U = {V^T}V = V{V^T} = {I_n}\;and\;\sum { = diag({\sigma _1},...,{\sigma _n})} $$ .
[27]	Mladenov M, MockM. “A step counter service for java-enabled devices using a built-in accelerometer,” in Proc. 1st Int [C]. Workshop Context-Aware Middleware Services, Afﬁliated 4^th Int. Conf. Commun. Syst. Softw. Middleware, 2009:1-5.
[28]	Rai A, Chintalapudi K K, Padmanabhan V N, et al. Zee: zero-effort crowdsourcing for indoor localization [C]//Proceedings of the 18th annual international conference on Mobile computing and networking. ACM, 2012:293-304.

Protecting query privacy in location based services

2014

... 随着智能移动终端的广泛普及,基于位置的服务（Location Based Service,LBS）^[1]与室内外一体化行人导航技术受到越来越多的关注.目前,利用GPS的定位导航已经能够满足室外行人的服务需求;而室内环境下由于接收到的GPS卫星信号差、多路径等影响,导致传统的定位方法无法直接使用,对于室内行人的定位及导航方法尚缺乏统一的标准.通常使用的室内定位方法可以分为2类：基于几何（Geometry）的定位和基于场景分析（Scene Analysis）的定位,使用的技术主要包括：无线传感网（Wireless Sensor Network, WSN）^[2]、无线局域网络（Wireless Local Area Networks, WLAN）^[3]、蓝牙^[4]、红外^[5]、射频识别（Radio Frequency Identification, RFID）^[6]等.其中,基于几何的定位方式很大程度上依赖于硬件设备,信号在传播过程中容易出现衰减等现象,导致定位结果存在系统偏差.因此,大多数室内定位的研究集中在基于场景位置的指纹匹配以及多传感器集成等方面.视觉定位^[7]是一种流行的定位技术,它具有成本低、无需外接设备、便捷性强、容易让人理解等优点.目前已开展关于视觉定位的研究^[8-11],但是利用视觉进行室内定位仍然面临着建立（更新）位置指纹库费时费力、图像匹配结果不准确等困难. ...

The effects of stitching orders in Patch-and-stitch WSN localization algorithms

2009

A novel algorithm for multipath fingerprinting in indoor WLAN environments

2008

Experiments on local positioning with Bluetooth

2003

A distributed location system for the active office

1994

LANDMARC: Indoor location sensing using active RFID

2004

Mobile visual location recognition

2011

Indoor localization using camera phones

2006

... 广义上图像匹配是指建立影像间最小空间分辨率单元一一对应关系的过程,是摄影测量、计算机视觉等领域的基础研究.图像匹配的结果可用于计算影像间的相对位姿^[20],也可直接反映图像间的相似程度^[8-11].一般情况下,2张影像的匹配点个数越多,其相似程度越高.本文提出一种多约束图像匹配算法,该算法在SIFT特征匹配的基础上增加了3种约束条件,可以剔除传统图像匹配中存在的大部分错误匹配结果,有助于恢复图像位姿和提高视觉定位的准确率.在定位阶段,使用本文方法计算查询影像与数据库影像间的匹配点数量,最多匹配点的K张影像位置坐标加权平均作为查询影像的定位结果. ...

Automatic construction of Wi-Fi radio map using smartphones

2016

A two-dimensional pedestrian navigation solution aided with a visual gyroscope and a visual odometer

2013

Image based localization in indoor environments

2013

Pedestrian dead reckoning: A basis for personal positioning

2006

... 行人航位推算（PDR）方法^[12]利用手机内置的陀螺仪、加速度计等传感器信息,可以实时计算行人的位置坐标,为室内多源信号位置指纹的自动标注提供了解决方案.但是,由于手机内置传感器的精度不高,航位角度计算容易产生累积误差,导致航位推算结果与真值偏差较大、不能直接用于位置指纹自动标注.为进一步提高航位推算位置精度,文献[13]提出了基于行为感知的行人室内定位方法,通过行人感知技术^[14]提取时间序列内的多个行为,结合室内路网拓扑结构^[15],匹配行为点到相应的特殊位置,从而消除了行人航位推算的累积误差.文献[16]采用UKF（Unscented Kalman Filter）融合Wi-Fi指纹匹配定位和PDR定位,得到了很好的室内定位结果.文献[17]针对多种环境下无线信号可以充分利用的特点,提出了一种泛在无线信号辅助的无缝定位方法,并对无缝定位技术的特点和发展趋势进行了讨论.文献[18]使用粒子滤波结合室内平面地图信息,对行人航位推算（PDR）结果进行修正,使室内行人的定位精度有所提高.上述这些室内定位系统或方法都在一定程度上提高了行人航位推算（PDR）的定位结果,但是往往需要添加额外的硬件设施或需要已知室内平面地图信息,降低了室内多源信号位置指纹自动标注的实际可操作性.文献[19]借助多个室内基站的实时数据,提出一种面向无线局域网的位置指纹动态自适应模型,该模型可以自适应更新室内位置指纹数据库,但是带来了更多的计算负荷. ...

用户行为感知辅助的室内行人定位

2014

用户行为感知辅助的室内行人定位

2014

Activity sequence-based indoor pedestrian localization using smartphones

2015

顾及地标可视性的室内导航路径优化算法

2017

顾及地标可视性的室内导航路径优化算法

2017

WiFi-PDR 室内组合定位的无迹卡尔曼滤波算法

2015

WiFi-PDR 室内组合定位的无迹卡尔曼滤波算法

2015

泛在无线信号辅助的室内外无缝定位方法与关键技术

2009

泛在无线信号辅助的室内外无缝定位方法与关键技术

2009

Map-aided indoor mobile positioning system using particle filter

2005

面向室内WLAN定位的动态自适应模型

2015

面向室内WLAN定位的动态自适应模型

2015

A factorization based algorithm for multi-image projective structure and motion

1996

A combined corner and edge detector

1988

... 目前,常用的图像匹配特征包括：Harris特征^[21]、SIFT特征^[22]、SURF特征^[23]等.其中,SIFT特征点具有尺度、旋转不变的特性,被广泛的用于图像拼接、影像跟踪等领域.利用特征点的高维特征描述子（特征向量）来衡量图像间的相似性是目前常用的方法,通常使用欧式距离、曼哈顿距离、夹角余弦等方式进行计算,然而仅仅使用距离最小作为相似性度量方式往往会发生错误的匹配结果（图2（a））.因此,本文以SIFT特征为例,相继使用了3种约束条件用于剔除图像匹配过程中产生错误的匹配点对,但是这3种约束并不局限于SIFT特征,所有基于向量的特征描述都可使用.具体约束条件如下： ...

Distinctive image features from scale-invariant keypoints

2004

Surf: Speeded up robust features

2006

... （3）RANSAC约束.RANSAC是指随机采样一致性^[23],其原理是从观测数据中随机挑选部分数据得出模型参数,使用模型参数验证其他数据是否符合,通过不断的选取观测数据得到最优模型参数的方法.在本方法中,首先选择4对匹配点计算得到单应矩阵H,其中： ...

Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography

1981

... 本文计算图像序列运动的方位角主要使用了从运动恢复结构（Structure From Motion, SFM）方法^[24],其基本原理是：利用双视几何中的极线约束,使用基础矩阵表达影像同名点间的对应关系,在已知相机内参的情况下,分解本质矩阵得到图像间的相对位置姿态. ...

Motion and structure from motion in a piecewise planar environment

1988

... 由于手机内置惯性传感器的精度不高,导致行人航位推算（PDR）方法得到的标注位置与真实值存在偏差,并且随着时间的推移累积误差会越来越大,其主要误差源是方向角度值的漂移.本文利用图像序列计算得到的角度作为方向角,使用加速度波峰检测算法^[25]探测行走步数、步态检测算法^[26]估计行走步长.在已知起始坐标的情况下,利用PDR方法可以标注行走轨迹上的图像序列坐标（式（8））. ...

Singular value decomposition and least squares solutions

1970

“A step counter service for java-enabled devices using a built-in accelerometer,” in Proc. 1st Int

2009

Zee: zero-effort crowdsourcing for indoor localization

2012

一种运动恢复结构和航位推算结合的室内行人视觉定位方法

An Indoor Pedestrian Positioning Approach Based on the Integration of SFM and PDR

1 引言

2 多约束图像匹配及视觉定位

3 影像位置自动标注

3.1 图像序列方位角计算

3.2 图像序列位置标注

4 实验分析

4.1 实验场景

4.2 图像运动角度估计试验

4.3 运动轨迹标注试验

4.4 视觉定位试验

5 结论

参考文献 文献选项 原文顺序 文献年度倒序 文中引用次数倒序 被引期刊影响因子

参考文献

原文顺序

文献年度倒序

文中引用次数倒序

被引期刊影响因子