-
1.
公开(公告)号:US08401279B2
公开(公告)日:2013-03-19
申请号:US12849450
申请日:2010-08-03
申请人: Tsukasa Sugino , Ryujin Watanabe
发明人: Tsukasa Sugino , Ryujin Watanabe
IPC分类号: G06K9/00
CPC分类号: B25J9/1697 , G05B2219/37567 , G06K9/00664 , G06K9/209 , G06K9/6289 , G06T3/0006
摘要: A system capable of recognizing position, a shape, a posture and the like of an object present in a marginal environment of a device such as a robot in order to make the device perform operations on the object as a subject. In an environment recognition system, 3D information and physical information (color information and the like) of a subject are associated by using camera parameters of each of a 3D image sensor and a 2D image sensor. Thereby, the position, the posture and the shape related to the subject and the physical information of the subject present in the environment of a robot are obtained.
摘要翻译: 能够识别存在于诸如机器人的装置的边缘环境中的物体的位置,形状,姿势等的系统,以使得装置对作为被摄体的物体进行操作。 在环境识别系统中,通过使用3D图像传感器和2D图像传感器中的每一个的相机参数来关联对象的3D信息和物理信息(颜色信息等)。 由此,获得与被检体有关的位置,姿势和形状以及存在于机器人的环境中的被检体的物理信息。
-
2.
公开(公告)号:US5522015A
公开(公告)日:1996-05-28
申请号:US384732
申请日:1995-02-10
申请人: Ryujin Watanabe
发明人: Ryujin Watanabe
CPC分类号: G06N3/04
摘要: A neural network has an input layer, a hidden layer, and an output layer. The neural network includes a lower neural network model composed of hidden layer neurons and input layer neurons for learning a plurality of linearly separable patterns, and a higher neural network model composed of hidden layer neurons and output layer neurons for combining the linearly separable patterns into a linearly unseparable pattern.
摘要翻译: 神经网络具有输入层,隐藏层和输出层。 神经网络包括由隐层神经元组成的较低神经网络模型和用于学习多个线性可分离模式的输入层神经元,以及由隐层神经元和输出层神经元组成的较高神经网络模型,用于将线性可分离模式组合成 线性不可分割的模式。
-
公开(公告)号:US20090148034A1
公开(公告)日:2009-06-11
申请号:US12315411
申请日:2008-12-03
申请人: Nobuo Higaki , Ryujin Watanabe
发明人: Nobuo Higaki , Ryujin Watanabe
IPC分类号: G06K9/00
CPC分类号: G06K9/3241 , G06K9/6857
摘要: There is disclosed a mobile robot including an image processor that generates recognition information regarding a target object included in a taken image, and a main controller integrally controlling the robot based on this recognition information. The image processor executes steps of: generating a low-resolution image and at least one high-resolution image whose resolution higher than that of the low-resolution image; generating first target object information regarding the target object from the low-resolution image; determining which high-resolution image should be processed if two or more high-resolution images are generated, and then defining a resolution process region in the low-resolution image; processing a region in the high-resolution region corresponding to the resolution process region in the low-resolution image, so as to generate second target object information in the high-resolution image; and determining whether or not the first and the second target object information are matched; and based on this determination, using at least either of the first and the second target object information, thereby to generate the recognition information.
摘要翻译: 公开了一种移动机器人,其包括:图像处理器,其生成关于包括在拍摄图像中的目标对象的识别信息;以及主控制器,基于该识别信息来一体地控制机器人。 图像处理器执行以下步骤:产生分辨率高于低分辨率图像的分辨率的低分辨率图像和至少一个高分辨率图像; 从所述低分辨率图像生成关于所述目标对象的第一目标对象信息; 如果生成两个或多个高分辨率图像,则确定应该处理哪个高分辨率图像,然后在低分辨率图像中定义分辨率处理区域; 处理对应于低分辨率图像中的分辨率处理区域的高分辨率区域中的区域,以便在高分辨率图像中产生第二目标对象信息; 以及确定所述第一和第二目标对象信息是否匹配; 并且基于该确定,使用第一和第二目标对象信息中的至少一个,从而生成识别信息。
-
公开(公告)号:US08090193B2
公开(公告)日:2012-01-03
申请号:US12315411
申请日:2008-12-03
申请人: Nobuo Higaki , Ryujin Watanabe
发明人: Nobuo Higaki , Ryujin Watanabe
CPC分类号: G06K9/3241 , G06K9/6857
摘要: There is disclosed a mobile robot including an image processor that generates recognition information regarding a target object included in a taken image, and a main controller integrally controlling the robot based on this recognition information. The image processor executes steps of: generating a low-resolution image and at least one high-resolution image whose resolution higher than that of the low-resolution image; generating first target object information regarding the target object from the low-resolution image; determining which high-resolution image should be processed if two or more high-resolution images are generated, and then defining a resolution process region in the low-resolution image; processing a region in the high-resolution region corresponding to the resolution process region in the low-resolution image, so as to generate second target object information in the high-resolution image; and determining whether or not the first and the second target object information are matched; and based on this determination, using at least either of the first and the second target object information, thereby to generate the recognition information.
摘要翻译: 公开了一种移动机器人,其包括:图像处理器,其生成关于包括在拍摄图像中的目标对象的识别信息;以及主控制器,基于该识别信息来一体地控制机器人。 图像处理器执行以下步骤:产生分辨率高于低分辨率图像的分辨率的低分辨率图像和至少一个高分辨率图像; 从所述低分辨率图像生成关于所述目标对象的第一目标对象信息; 如果生成两个或多个高分辨率图像,则确定应该处理哪个高分辨率图像,然后在低分辨率图像中定义分辨率处理区域; 处理对应于低分辨率图像中的分辨率处理区域的高分辨率区域中的区域,以便在高分辨率图像中产生第二目标对象信息; 以及确定所述第一和第二目标对象信息是否匹配; 并且基于该确定,使用第一和第二目标对象信息中的至少一个,从而生成识别信息。
-
公开(公告)号:US5200898A
公开(公告)日:1993-04-06
申请号:US614194
申请日:1990-11-15
申请人: Hiromitsu Yuhara , Ryujin Watanabe
发明人: Hiromitsu Yuhara , Ryujin Watanabe
CPC分类号: F02D41/1405 , F02D41/045 , F02D2041/1433 , Y10S706/905
摘要: A motor vehicle is controlled with a neural network which has a data learning capability. A present value of the throttle valve opening of the engine on the motor vehicle and a rate of change of the present value of the throttle valve opening are periodically supplied to the neural network. The neural network is controlled to learn the present value of the throttle valve opening when the rate of change of the present value of the throttle valve opening becomes zero so that a predicted value of the throttle valve opening approaches the actual value of the throttle valve opening at the time the rate of change thereof becomes zero. An operating condition of the motor vehicle is controlled based on the predicted value of the throttle valve opening, which is represented by a periodically produced output signal from the neural network.
摘要翻译: 用具有数据学习能力的神经网络控制汽车。 机动车辆上的发动机的节气门开度的当前值和节气门开度的当前值的变化率周期性地提供给神经网络。 当节气门开度的当前值的变化率变为零时,控制神经网络以了解节流阀开度的当前值,使得节流阀开度的预测值接近节气门开度的实际值 在其变化率为零时。 基于由神经网络周期性产生的输出信号表示的节气门开度的预测值来控制机动车辆的运转状态。
-
-
-
-