当前的汽车安全辅助驾驶和无人驾驶汽车是图像领域的研究热点,针对汽车在启动或行驶时车前存在行人可能导致的安全问题,着重研究了基于双目视觉的车前行人检测方法。进行了双目相机的相机标定和立体标定;通过改进后半全局立体匹配算法...当前的汽车安全辅助驾驶和无人驾驶汽车是图像领域的研究热点,针对汽车在启动或行驶时车前存在行人可能导致的安全问题,着重研究了基于双目视觉的车前行人检测方法。进行了双目相机的相机标定和立体标定;通过改进后半全局立体匹配算法获取深度图,确定车前行人所处位置的感兴趣区域(Region of Interest,ROI),剔除冗余的背景信息;分割并提取了图像的降维梯度直方图(Histogram of Gradients,HOG)特征信息;将特征输入到支持向量机(Support Vector Machine,SVM)分类器训练,检测并标记出车前的行人目标。实验证明,所提算法对车前场景下的动态行人可以更为有效地检测,具备更优的检率精度、时效性和鲁棒性。展开更多
针对行人被障碍物部分遮挡导致的检测准确率降低问题,提出了基于多特征融合的树形路径半全局立体匹配的部分遮挡行人检测算法。使用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法进行超像素分割,提升行人的轮廓信息,...针对行人被障碍物部分遮挡导致的检测准确率降低问题,提出了基于多特征融合的树形路径半全局立体匹配的部分遮挡行人检测算法。使用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法进行超像素分割,提升行人的轮廓信息,并使用多特征融合的树形路径半全局立体匹配算法生成深度图;对行人信息和背景信息及障碍物信息使用自适应分割算法进行分离,获取感兴趣区域;将感兴趣区域放置在行人特征明显且稳定的头肩部,进行感兴趣区域的约束;使用降维梯度直方图特征(histogram of gradient,HOG)进行特征提取并生成样本集,训练支持向量机(support vector machines,SVM)分类器,最终实现部分遮挡的行人检测。实验表明,所提算法与其他行人检测算法相比,在行人部分遮挡场景下,有着更高的行人检测准确率,证明所提算法的有效性。展开更多
针对行人航位推算(Pedestrian Dead Reckoning,PDR)室内定位系统的累计误差问题,提出了一种多维信息感知地标匹配的PDR定位算法(PDR positioning algorithm based Multi-imensional Information Perception Landmark Matching,MIPLM)。...针对行人航位推算(Pedestrian Dead Reckoning,PDR)室内定位系统的累计误差问题,提出了一种多维信息感知地标匹配的PDR定位算法(PDR positioning algorithm based Multi-imensional Information Perception Landmark Matching,MIPLM)。算法利用行人在室内走廊环境下的众包轨迹,并基于突出性路口结构,从位置、航向、影响范围以及WiFi特征指纹等方面构建多维信息感知地标库。给出的自适应地标检测算法,结合航向约束轨迹相似度匹配模型,更新行人位置和航向,避免了本地化匹配过程对空间位置的强依赖性。实验结果表明,相比于其他地标构建及匹配算法,所提算法更好地反映了行人活动与室内空间结构的相关性,且在未知起始位置时,算法能够快速收敛并提供较高的定位精度,对于室内行人连续定位具有较高的应用价值。展开更多
Navigation systems play an important role in many vital disciplines. Determining the location of a user relative to its physical environment is an important part of many indoor-based navigation services such as user n...Navigation systems play an important role in many vital disciplines. Determining the location of a user relative to its physical environment is an important part of many indoor-based navigation services such as user navigation, enhanced 911 (E911), law enforcement, location-based and marketing services. Indoor navigation applications require a reliable, trustful and continuous navigation solution that overcomes the challenge of Global Navigation Satellite System (GNSS) signal unavailability. To compensate for this issue, other navigation systems such as Inertial Navigation System (INS) are introduced, however, over time there is a significant amount of drift especially in common with low-cost commercial sensors. In this paper, a map aided navigation solution is developed. This research develops an aiding system that utilizes geospatial data to assist the navigation solution by providing virtual boundaries for the navigation trajectories and limits its possibilities only when it is logical to locate the user on a map. The algorithm develops a Pedestrian Dead Reckoning (PDR) based on smart-phone accelerometer and magnetometer sensors to provide the navigation solution. Geospatial model for two indoor environments with a developed map matching algorithm was used to match and project navigation position estimates on the geospatial map. The developed algorithms were field tested in indoor environments and yielded accurate matching results as well as a significant enhancement to positional accuracy. The achieved results demonstrate that the contribution of the developed map aided system enhances the reliability, usability, and accuracy of navigation trajectories in indoor environments.展开更多
文摘当前的汽车安全辅助驾驶和无人驾驶汽车是图像领域的研究热点,针对汽车在启动或行驶时车前存在行人可能导致的安全问题,着重研究了基于双目视觉的车前行人检测方法。进行了双目相机的相机标定和立体标定;通过改进后半全局立体匹配算法获取深度图,确定车前行人所处位置的感兴趣区域(Region of Interest,ROI),剔除冗余的背景信息;分割并提取了图像的降维梯度直方图(Histogram of Gradients,HOG)特征信息;将特征输入到支持向量机(Support Vector Machine,SVM)分类器训练,检测并标记出车前的行人目标。实验证明,所提算法对车前场景下的动态行人可以更为有效地检测,具备更优的检率精度、时效性和鲁棒性。
文摘针对行人被障碍物部分遮挡导致的检测准确率降低问题,提出了基于多特征融合的树形路径半全局立体匹配的部分遮挡行人检测算法。使用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法进行超像素分割,提升行人的轮廓信息,并使用多特征融合的树形路径半全局立体匹配算法生成深度图;对行人信息和背景信息及障碍物信息使用自适应分割算法进行分离,获取感兴趣区域;将感兴趣区域放置在行人特征明显且稳定的头肩部,进行感兴趣区域的约束;使用降维梯度直方图特征(histogram of gradient,HOG)进行特征提取并生成样本集,训练支持向量机(support vector machines,SVM)分类器,最终实现部分遮挡的行人检测。实验表明,所提算法与其他行人检测算法相比,在行人部分遮挡场景下,有着更高的行人检测准确率,证明所提算法的有效性。
文摘针对行人航位推算(Pedestrian Dead Reckoning,PDR)室内定位系统的累计误差问题,提出了一种多维信息感知地标匹配的PDR定位算法(PDR positioning algorithm based Multi-imensional Information Perception Landmark Matching,MIPLM)。算法利用行人在室内走廊环境下的众包轨迹,并基于突出性路口结构,从位置、航向、影响范围以及WiFi特征指纹等方面构建多维信息感知地标库。给出的自适应地标检测算法,结合航向约束轨迹相似度匹配模型,更新行人位置和航向,避免了本地化匹配过程对空间位置的强依赖性。实验结果表明,相比于其他地标构建及匹配算法,所提算法更好地反映了行人活动与室内空间结构的相关性,且在未知起始位置时,算法能够快速收敛并提供较高的定位精度,对于室内行人连续定位具有较高的应用价值。
文摘Navigation systems play an important role in many vital disciplines. Determining the location of a user relative to its physical environment is an important part of many indoor-based navigation services such as user navigation, enhanced 911 (E911), law enforcement, location-based and marketing services. Indoor navigation applications require a reliable, trustful and continuous navigation solution that overcomes the challenge of Global Navigation Satellite System (GNSS) signal unavailability. To compensate for this issue, other navigation systems such as Inertial Navigation System (INS) are introduced, however, over time there is a significant amount of drift especially in common with low-cost commercial sensors. In this paper, a map aided navigation solution is developed. This research develops an aiding system that utilizes geospatial data to assist the navigation solution by providing virtual boundaries for the navigation trajectories and limits its possibilities only when it is logical to locate the user on a map. The algorithm develops a Pedestrian Dead Reckoning (PDR) based on smart-phone accelerometer and magnetometer sensors to provide the navigation solution. Geospatial model for two indoor environments with a developed map matching algorithm was used to match and project navigation position estimates on the geospatial map. The developed algorithms were field tested in indoor environments and yielded accurate matching results as well as a significant enhancement to positional accuracy. The achieved results demonstrate that the contribution of the developed map aided system enhances the reliability, usability, and accuracy of navigation trajectories in indoor environments.