公开(公告)号：US07149356B2

公开(公告)日：2006-12-12

申请号：US10193019

申请日：2002-07-10

申请人： Walter D. Clark ,  Richard R. Ramroth

发明人： Walter D. Clark ,  Richard R. Ramroth

IPC分类号： G06K9/48 ,  G06K9/62

CPC分类号： G06K9/3241 ,  G06K9/48

摘要： The present invention includes a method for matching a stored template with an image of an object. This is a final stage identification device in the at least one candidate array is received from a prescreener and/or intermediate discriminator. The candidate array includes a respective plurality of direction values at least partially defining a respective contour of an object within a two-dimensional image containing a possible target. The candidate array is compared with at least one stored template array. The stored template array includes a respective plurality of direction values at least partially defining an outline of a predetermined object of interest. The template array may include a plurality of template segment arrays such that various portions of the overall template are compared one at a time with the entire candidate array. The match between candidate and stored template may be scored based upon how well the template segments correspond to the candidate. Furthermore, the candidate may also be scored based upon a positional relationship between the template segments, while they are aligned with respective portions of the candidate.

摘要翻译： 本发明包括用于将存储的模板与对象的图像进行匹配的方法。 这是从预处理器和/或中间鉴别器接收至少一个候选阵列中的最后阶段识别装置。 候选阵列包括相应的多个方向值，其至少部分地限定包含可能目标的二维图像内的对象的相应轮廓。 将候选阵列与至少一个存储的模板阵列进行比较。 存储的模板阵列包括相应的多个方向值，至少部分地限定预定的感兴趣对象的轮廓。 模板阵列可以包括多个模板段阵列，使得整个模板的各个部分一次一个地与整个候选阵列进行比较。 基于模板片段与候选者对应的多少，可以对候选和存储的模板之间的匹配进行评分。 此外，还可以基于模板段之间的位置关系对候选者进行评分，同时它们与候选者的各个部分对齐。

公开(公告)号：US07627170B2

公开(公告)日：2009-12-01

申请号：US11247422

申请日：2005-10-11

申请人： Omar Aboutalib ,  Richard R. Ramroth ,  Bea Thai

发明人： Omar Aboutalib ,  Richard R. Ramroth ,  Bea Thai

IPC分类号： G06K9/00

CPC分类号： G06K9/00201 ,  G06K2209/29

摘要： The invention is a process for identifying an unknown object. In detail, the process includes the steps of: 1) compiling data on selected features on a plurality of segments of a plurality of known objects; 2) illuminating the unknown object with a laser radar system; 3) dividing the unknown object into a plurality segments corresponding to each of the segments of the known objects; 4) sequentially measuring selected features of each of the plurality of segments of the unknown object; and 5) comparing the sequentially measuring selected features of each of the plurality of segments of the unknown object to the selected features on the plurality of segments of the plurality of known objects.

摘要翻译： 本发明是识别未知物体的过程。 具体而言，该过程包括以下步骤：1）在多个已知对象的多个段上编译所选特征上的数据; 2）用激光雷达系统照亮未知物体; 3）将未知对象划分成对应于已知对象的每个段的多个段; 4）顺序地测量未知物体的多个片段中的每一个的所选特征; 以及5）将所述未知对象的所述多个片段中的每一个的所选特征顺序地测量到所述多个已知对象的所述多个片段上的所选特征。