公开(公告)号：US20060176209A1

公开(公告)日：2006-08-10

申请号：US11054757

申请日：2005-02-10

申请人： David Shu ,  Cynthia Daniell

发明人： David Shu ,  Cynthia Daniell

IPC分类号： G01S13/90

CPC分类号： G01S7/411 ,  G01S13/90 ,  G06K9/3241 ,  G06K9/4642

摘要： A method for classification of a target object having a periphery comprises the steps of: selecting a plurality of random first chords D1 across the periphery of the target object; measuring each of the first chords D1 to obtain a plurality of first dimensions; computing for each of the first chords D1 a second chord D2 across the periphery thus forming a plurality of D1, D2 pairs of chords; measuring the second chords D2 for all pairs to obtain second dimensions; computing for each of the pairs of first chords D1 and second chords D2 the ratio D1/D2 of the first dimension to the second dimension to obtain a plurality of first values; computing the logarithm of the first values to obtain a plurality of logarithmic values; computing difference values by subtracting the second dimension from the first dimension for each of the pairs; recording the first values, logarithmic values and difference values in histograms; extracting a vertex lists from the histograms; combining one or more of the vertex lists to obtain a signature indicative of the classification of the object target. For each of the pairs, the second chords are orthogonal to the first chords, and the second chords share a point with the first chords and the periphery.

公开(公告)号：US06897802B1

公开(公告)日：2005-05-24

申请号：US10704841

申请日：2003-11-10

申请人： Cynthia Daniell ,  Narayan Srinivasa

发明人： Cynthia Daniell ,  Narayan Srinivasa

IPC分类号： G01S13/90 ,  G01S7/41

CPC分类号： G01S13/9005

摘要： A plurality of image chips (202) (over 100), each of the chips containing the same, known target of interest, such as, for example an M109 tank are presented to the system for training. Each image chip of the known target is slightly different than the next, showing the known target at different aspect angles and rotation with respect to the moving platform acquiring the image chip.The system extract multiple features of the known target from the plurality of image chips (202) presented for storage and analysis, or training. These features distinguish a known target of interest from the nearest similar target to the M109 tank, for example a Caterpillar D7 bulldozer. These features are stored for use during unknown target identification. When an unknown target chip is presented, the recognition algorithm relies on the features stored during training to attempt to identify the target.The tools used for extracting features of the known target of interest as well as the unknown target presented for identification are the same and include the Haar Transform (404), and entropy measurements (410) generating coefficient locations. Using the Karhunen-Loeve (KL) transform 406, eigenvectors are computed. A Gaussian mixture model (GMM) (507) is used to compare the extracted coefficients and eigenfeatures from the known target chips with that of the unknown target chips. Thus the system is trained initially by presenting to it known target chips for classification. Subsequently, the system uses the training in the form of stored eigenfeatures and entropy coefficients fused with multiscale features to identify unknown targets.

摘要翻译： 将多个图像芯片（202）（超过100个），包含相同的每个芯片，已知的目标目标，例如M109坦克被呈现给用于训练的系统。 已知目标的每个图像芯片与下一个图像芯片稍微不同，示出了在获取图像芯片的移动平台上以不同方位角度和旋转的已知目标。 该系统从呈现用于存储和分析或训练的多个图像芯片（202）中提取已知目标的多个特征。 这些功能将已知的目标目标与最近的类似目标区别到M109坦克，例如卡特彼勒D7推土机。 这些功能存储在未知目标识别期间使用。 当出现未知的目标芯片时，识别算法依赖于训练期间存储的特征来尝试识别目标。 用于提取已知目标目标的特征以及用于识别的未知目标的工具是相同的，并且包括Haar变换（404）和产生系数位置的熵测量（410）。 使用Karhunen-Loeve（KL）变换406，计算特征向量。 高斯混合模型（GMM）（507）用于比较已知目标芯片的提取系数和本征特征与未知目标芯片的特征特征。 因此，系统最初通过向已知的目标芯片呈现分类来进行训练。 随后，系统采用与多尺度特征融合的存储的特征特征和熵系数的形式的训练来识别未知目标。