公开(公告)号：US20110044543A1

公开(公告)日：2011-02-24

申请号：US12451078

申请日：2008-05-30

申请人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Motoki Kanba ,  Osamu Aisaka ,  Kiyokazu Okada

发明人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Motoki Kanba ,  Osamu Aisaka ,  Kiyokazu Okada

IPC分类号： G06K9/46

CPC分类号： G06K9/00798

摘要： A feature extraction method includes: the step of grouping a cluster of features, in which an internal of the respective features is less than or equal to a predetermined grouping interval, to form a feature group, for a plurality of features of which feature information including at least information of a position and a feature type is included in a predetermined feature information storage unit; the step of excluding the feature, not suitable for use in an image recognition process of the feature with respect to image information, from the cluster of the features within the feature group; and the step of extracting a part or all of one or more of the features within the feature group remaining as a result of the exclusion step as a target feature suitable for the use in the image recognition process.

摘要翻译： 特征提取方法包括：对于多个特征，对特征的一组特征进行分组，其中各个特征的内部小于或等于预定的分组间隔，以形成特征组，其特征信息包括 位置和特征类型的至少信息被包括在预定特征信息存储单元中; 从特征组中的特征的集群排除不适合用于相对于图像信息的特征的图像识别处理的特征的步骤; 以及将作为排除步骤的结果剩余的特征组中的一个或多个特征的一部分或全部的部分或全部提取为适合于在图像识别处理中使用的目标特征的步骤。

公开(公告)号：US08825364B2

公开(公告)日：2014-09-02

申请号：US12213040

申请日：2008-06-13

申请人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

发明人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

IPC分类号： G01C21/30

CPC分类号： G01C21/30

摘要： A vehicle location recognition apparatus includes a link determination unit which accurately determines a link traveled by a vehicle from among a plurality of links meeting at a narrow-angle branch point, on the basis of vehicle location information and road information, travel history information representing a route traveled by the vehicle at the branch point. The vehicle location recognition apparatus includes a history information generation unit for generating travel history information for the determined link. The travel history information is stored in a learning database. On the basis of the travel history information, a learning unit generates learned priority information representing the priorities of the respective links meeting at a branch point and used in the determination by the link determination unit.

摘要翻译： 车辆位置识别装置包括链路确定单元，其基于车辆位置信息和道路信息，精确地确定从窄角分支点处的多个链路中的车辆行驶的链路，行驶历史信息表示 车辆在分支点行驶的路线。 车辆位置识别装置包括用于生成所确定的链路的旅行历史信息的历史信息生成单元。 旅行历史信息存储在学习数据库中。 基于旅行历史信息，学习单元生成表示在分支点处会议的各个链路的优先级的学习优先级信息，并且由链接确定单元进行确定。

公开(公告)号：US08155873B2

公开(公告)日：2012-04-10

申请号：US12213074

申请日：2008-06-13

申请人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

发明人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

IPC分类号： G01C21/36

CPC分类号： G01C21/28 ,  G06K9/00798

摘要： Vehicle-position-recognition apparatuses, methods, and programs acquire current position information indicating a current position of a vehicle and acquire road information indicating at least one road based on a connection relationship between plural links. The apparatuses, methods, and programs determine a link where the vehicle travels based on the current position information and the road information and generate travel-history information indicating a route that the vehicle travels at a point where a link is branched based on the determined link. The apparatuses, methods, and programs acquire an image of an area surrounding the vehicle, recognize a recognition target in the acquired image, and store the generated travel-history information in the memory in association with information indicating the recognition of the recognition target. The apparatuses, methods, and programs generate learning-priority information indicating a priority of each link branched out at the branch point based on the stored travel-history information and the stored recognition information, the generated learning-priority information usable to determine on which of the links branched out at the branch point that the vehicle is traveling.

摘要翻译： 车辆位置识别装置，方法和程序获取指示车辆的当前位置的当前位置信息，并且基于多个链路之间的连接关系获取指示至少一条道路的道路信息。 设备，方法和程序根据当前位置信息和道路信息确定车辆行驶的链接，并根据所确定的链路，生成表示车辆在分支点处行驶的路线的旅行历史信息 。 装置，方法和程序获取车辆周围的区域的图像，识别所获取的图像中的识别目标，并将生成的旅行历史信息与指示识别目标的识别的信息相关联地存储在存储器中。 所述装置，方法和程序基于所存储的旅行历史信息和所存储的识别信息，生成指示在分支点分支的每个链接的优先级的学习优先级信息，所生成的学习优先级信息可用于确定哪个 链路在车辆行驶的分支点处分支。

公开(公告)号：US08346473B2

公开(公告)日：2013-01-01

申请号：US12593772

申请日：2008-05-23

申请人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Osamu Aisaka ,  Motoki Kanba ,  Kiyokazu Okada

发明人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Osamu Aisaka ,  Motoki Kanba ,  Kiyokazu Okada

IPC分类号： G01C21/32

CPC分类号： G06K9/00798 ,  G01C21/30 ,  G01C21/3658

摘要： A lane determining device includes: a feature information obtaining unit for obtaining feature information F of a target feature existing in each lane in the traveling direction of a vehicle 30, based on vehicle position information; an image recognizing unit for conducting image recognition processing of a feature type of the target feature in a vehicle lane, with respect to image information; and a lane accuracy determining unit for determining a vehicle lane accuracy for the respective lanes based on a type of a recognized feature fa shown in an image recognition result, feature types Ft1 to Ft4 of the target feature of the respective lanes shown in the feature information F, in accordance with a level of probability for each of the feature types Ft1 to Ft4 of the target feature of the respective lanes is recognized as the type of the recognized feature fa.

摘要翻译： 车道确定装置包括：特征信息获取单元，用于基于车辆位置信息获得车辆30的行驶方向上存在于每个车道中的目标特征的特征信息F; 图像识别单元，用于相对于图像信息进行车道中的目标特征的特征类型的图像识别处理; 以及车道精度确定单元，用于基于图像识别结果中所示的识别特征fa的类型来确定各车道的车道准确度，特征信息中所示的各车道的目标特征的特征类型Ft1至Ft4 F根据各车道的目标特征的特征类型Ft1至Ft4的概率的等级被识别为识别的特征fa的类型。

公开(公告)号：US20090005979A1

公开(公告)日：2009-01-01

申请号：US12213040

申请日：2008-06-13

申请人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

发明人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

IPC分类号： G01C21/34

CPC分类号： G01C21/30

摘要： A vehicle location recognition apparatus includes a link determination unit which accurately determines a link traveled by a vehicle from among a plurality of links meeting at a narrow-angle branch point, on the basis of vehicle location information and road information, travel history information representing a route traveled by the vehicle at the branch point. The vehicle location recognition apparatus includes a history information generation unit for generating travel history information for the determined link. The travel history information is stored in a learning database. On the basis of the travel history information, a learning unit generates learned priority information representing the priorities of the respective links meeting at a branch point and used in the determination by the link determination unit.

摘要翻译： 车辆位置识别装置包括链路确定单元，其基于车辆位置信息和道路信息，精确地确定从窄角分支点处的多个链路中的车辆行驶的链路，行驶历史信息表示 车辆在分支点行驶的路线。 车辆位置识别装置包括用于生成所确定的链路的旅行历史信息的历史信息生成单元。 旅行历史信息存储在学习数据库中。 基于旅行历史信息，学习单元生成表示在分支点处会议的各个链路的优先级的学习优先级信息，并且由链接确定单元进行确定。

公开(公告)号：US20080243312A1

公开(公告)日：2008-10-02

申请号：US12076067

申请日：2008-03-13

申请人： Masaki Nakamura ,  Koichi Nakao ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

发明人： Masaki Nakamura ,  Koichi Nakao ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

IPC分类号： B60W40/12 ,  G01C21/26

CPC分类号： G01C21/26 ,  B60W2420/42 ,  B60W2550/10

摘要： Vehicle behavior learning apparatuses, methods, and programs store pieces of feature information including position information and attribute information of a plurality of target features. The apparatuses, methods, and programs obtain vehicle position information that shows a current position of a vehicle, obtaining image information of surroundings of the vehicle, and obtain one of the stored pieces of feature information corresponding to the surroundings of the vehicle based on the vehicle position information. The apparatuses, methods, and programs perform an image recognition for recognizing a target feature contained in the image information that corresponds to the obtained piece of feature information, detect a behavior of the vehicle that is performed within a predetermined range from a position of the recognized target feature, and store, based on the vehicle position information, the detected behavior in correspondence with information of a position in which the detected behavior was detected. The apparatuses, methods, and programs extract, based the detected behavior being stored a plurality of times at the same position, the detected behavior as a learned behavior, output learned behavior information including attribute information of the detected behavior and the position information of the detected behavior, each of which are kept in correspondence with the piece of feature information of the recognized target feature.

摘要翻译： 车辆行为学习装置，方法和程序存储包括多个目标特征的位置信息和属性信息的特征信息。 所述装置，方法和程序获得显示车辆的当前位置的车辆位置信息，获得车辆周围环境的图像信息，并且基于车辆获得所存储的与车辆周围相对应的特征信息之一 位置信息。 设备，方法和程序执行图像识别，以识别包含在与获得的特征信息相对应的图像信息中的目标特征，从识别的位置检测在预定范围内执行的车辆的行为 目标特征，并且基于车辆位置信息存储与检测到的检测到的行为的位置的信息相对应的检测到的行为。 设备，方法和程序基于检测到的行为在相同位置被多次存储的情况下提取检测到的行为作为学习行为，输出学习行为信息，包括检测到的行为的属性信息和检测到的行为的位置信息 行为，其中的每一个保持与识别的目标特征的特征信息对应。

公开(公告)号：US08452101B2

公开(公告)日：2013-05-28

申请号：US12451078

申请日：2008-05-30

申请人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Motoki Kanba ,  Osamu Aisaka ,  Kiyokazu Okada

发明人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Motoki Kanba ,  Osamu Aisaka ,  Kiyokazu Okada

IPC分类号： G06K9/66

CPC分类号： G06K9/00798

摘要： A feature extraction method includes: the step of grouping a cluster of features, in which an internal of the respective features is less than or equal to a predetermined grouping interval, to form a feature group, for a plurality of features of which feature information including at least information of a position and a feature type is included in a predetermined feature information storage unit; the step of excluding the feature, not suitable for use in an image recognition process of the feature with respect to image information, from the cluster of the features within the feature group; and the step of extracting a part or all of one or more of the features within the feature group remaining as a result of the exclusion step as a target feature suitable for the use in the image recognition process.

摘要翻译： 特征提取方法包括：对于多个特征，对特征的一组特征进行分组，其中各个特征的内部小于或等于预定的分组间隔以形成特征组，其特征信息包括 位置和特征类型的至少信息被包括在预定特征信息存储单元中; 从特征组中的特征的集群排除不适合用于相对于图像信息的特征的图像识别处理的特征的步骤; 以及将作为排除步骤的结果剩余的特征组中的一个或多个特征的一部分或全部的部分或全部提取为适合于在图像识别处理中使用的目标特征的步骤。

公开(公告)号：US08155826B2

公开(公告)日：2012-04-10

申请号：US12076067

申请日：2008-03-13

申请人： Masaki Nakamura ,  Koichi Nakao ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

发明人： Masaki Nakamura ,  Koichi Nakao ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

IPC分类号： G01M17/00

CPC分类号： G01C21/26 ,  B60W2420/42 ,  B60W2550/10

摘要： Vehicle behavior learning apparatuses, methods, and programs store pieces of feature information including position information and attribute information of a plurality of target features. The apparatuses, methods, and programs obtain vehicle position information that shows a current position of a vehicle, obtaining image information of surroundings of the vehicle, and obtain one of the stored pieces of feature information corresponding to the surroundings of the vehicle based on the vehicle position information. The apparatuses, methods, and programs perform an image recognition for recognizing a target feature contained in the image information that corresponds to the obtained piece of feature information, detect a behavior of the vehicle that is performed within a predetermined range from a position of the recognized target feature, and store, based on the vehicle position information, the detected behavior in correspondence with information of a position in which the detected behavior was detected. The apparatuses, methods, and programs extract, based the detected behavior being stored a plurality of times at the same position, the detected behavior as a learned behavior, output learned behavior information including attribute information of the detected behavior and the position information of the detected behavior, each of which are kept in correspondence with the piece of feature information of the recognized target feature.

摘要翻译： 车辆行为学习装置，方法和程序存储包括多个目标特征的位置信息和属性信息的特征信息。 所述装置，方法和程序获得显示车辆的当前位置的车辆位置信息，获得车辆周围环境的图像信息，并且基于车辆获得所存储的与车辆周围相对应的特征信息之一 位置信息。 设备，方法和程序执行图像识别，以识别包含在与获得的特征信息相对应的图像信息中的目标特征，从识别的位置检测在预定范围内执行的车辆的行为 目标特征，并且基于车辆位置信息存储与检测到的检测到的行为的位置的信息相对应的检测到的行为。 设备，方法和程序基于检测到的行为在相同位置被多次存储的情况下提取检测到的行为作为学习行为，输出学习行为信息，包括检测到的行为的属性信息和检测到的行为的位置信息 行为，其中的每一个保持与识别的目标特征的特征信息对应。

公开(公告)号：US20100121569A1

公开(公告)日：2010-05-13

申请号：US12593772

申请日：2008-05-23

申请人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Osamu Aisaka ,  Motoki Kanba ,  Kiyokazu Okada

发明人： Masaki Nakamura ,  Tomoaki Ishikawa ,  Koichi Nakao ,  Osamu Aisaka ,  Motoki Kanba ,  Kiyokazu Okada

IPC分类号： G01C21/26 ,  G06K9/00 ,  G08G1/123

CPC分类号： G06K9/00798 ,  G01C21/30 ,  G01C21/3658

摘要： A lane determining device includes: a feature information obtaining unit for obtaining feature information F of a target feature existing in each lane in the traveling direction of a vehicle 30, based on vehicle position information; an image recognizing unit for conducting image recognition processing of a feature type of the target feature in a vehicle lane, with respect to image information; and a lane accuracy determining unit for determining a vehicle lane accuracy for the respective lanes based on a type of a recognized feature fa shown in an image recognition result, feature types Ft1 to Ft4 of the target feature of the respective lanes shown in the feature information F, in accordance with a level of probability for each of the feature types Ft1 to Ft4 of the target feature of the respective lanes is recognized as the type of the recognized feature fa.

摘要翻译： 车道确定装置包括：特征信息获取单元，用于基于车辆位置信息获得车辆30的行驶方向上存在于每个车道中的目标特征的特征信息F; 图像识别单元，用于相对于图像信息进行车道中的目标特征的特征类型的图像识别处理; 以及车道精度确定单元，用于基于图像识别结果中所示的识别特征fa的类型来确定各车道的车道准确度，特征信息中所示的各车道的目标特征的特征类型Ft1至Ft4 F根据各车道的目标特征的特征类型Ft1至Ft4的概率的等级被识别为识别的特征fa的类型。

公开(公告)号：US20090005980A1

公开(公告)日：2009-01-01

申请号：US12213074

申请日：2008-06-13

申请人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

发明人： Koichi Nakao ,  Masaki Nakamura ,  Tomoaki Ishikawa ,  Motoki Kanba ,  Osamu Aisaka

IPC分类号： G01C21/36

CPC分类号： G01C21/28 ,  G06K9/00798

摘要： Vehicle-position-recognition apparatuses, methods, and programs acquire current position information indicating a current position of a vehicle and acquire road information indicating at least one road based on a connection relationship between plural links. The apparatuses, methods, and programs determine a link where the vehicle travels based on the current position information and the road information and generate travel-history information indicating a route that the vehicle travels at a point where a link is branched based on the determined link. The apparatuses, methods, and programs acquire an image of an area surrounding the vehicle, recognize a recognition target in the acquired image, and store the generated travel-history information in the memory in association with information indicating the recognition of the recognition target. The apparatuses, methods, and programs generate learning-priority information indicating a priority of each link branched out at the branch point based on the stored travel-history information and the stored recognition information, the generated learning-priority information usable to determine on which of the links branched out at the branch point that the vehicle is traveling.

摘要翻译： 车辆位置识别装置，方法和程序获取指示车辆的当前位置的当前位置信息，并且基于多个链路之间的连接关系获取指示至少一条道路的道路信息。 设备，方法和程序根据当前位置信息和道路信息确定车辆行驶的链接，并根据所确定的链路，生成表示车辆在分支点处行驶的路线的旅行历史信息 。 装置，方法和程序获取车辆周围的区域的图像，识别所获取的图像中的识别目标，并将生成的旅行历史信息与指示识别目标的识别的信息相关联地存储在存储器中。 所述装置，方法和程序基于所存储的旅行历史信息和所存储的识别信息，生成指示在分支点分支的每个链接的优先级的学习优先级信息，所生成的学习优先级信息可用于确定哪个 链路在车辆行驶的分支点处分支。