公开(公告)号：US11699247B2

公开(公告)日：2023-07-11

申请号：US12647192

申请日：2009-12-24

Applicant: Xiangyun Ye ,  David Y. Li ,  Guruprasad Shivaram ,  David J. Michael

Inventor： Xiangyun Ye ,  David Y. Li ,  Guruprasad Shivaram ,  David J. Michael

IPC: G06T7/80

CPC classification number: G06T7/85 ,  G06T2207/10012 ,  G06T2207/30108

Abstract: This invention provides a system and method for runtime determination (self-diagnosis) of camera miscalibration (accuracy), typically related to camera extrinsics, based on historical statistics of runtime alignment scores for objects acquired in the scene, which are defined based on matching of observed and expected image data of trained object models. This arrangement avoids a need to cease runtime operation of the vision system and/or stop the production line that is served by the vision system to diagnose if the system's camera(s) remain calibrated. Under the assumption that objects or features inspected by the vision system over time are substantially the same, the vision system accumulates statistics of part alignment results and stores intermediate results to be used as indicator of current system accuracy. For multi-camera vision systems, cross validation is illustratively employed to identify individual problematic cameras. The system and method allows for faster, less-expensive and more-straightforward diagnosis of vision system failures related to deteriorating camera calibration.

公开(公告)号：US09121751B2

公开(公告)日：2015-09-01

申请号：US13296950

申请日：2011-11-15

Applicant: David J. Michael

Inventor： David J. Michael

IPC: G01G23/38 ,  G01G19/03 ,  G01G19/393 ,  G01G19/387

CPC classification number: G01G23/38 ,  G01G19/035 ,  G01G19/387 ,  G01G19/393

Abstract: The disclosure relates to weighing moving objects in a weighing platform functionally coupled to a computer-vision tracking platform. The objects can translate, rotate, and translate and rotate. Weighing of the objects can be accomplished through combination of object imaging and upstream weighing. Object imaging can permit tracking, through computer vision, a logical object moving in a trajectory from the first location to the second location, wherein a logical object is a formal representation of one or more physical objects. Upstream weighing can permit updating a record indicative of weight of the one or more physical objects associated with the tracked logical object. As a part of weighing termination, data integrity check(s) can be performed on a plurality of records indicative of a weight of a single physical object. Based on outcome of the data integrity check(s), a record indicative of the weight of the single physical object can be supplied.

Abstract translation: 本公开涉及在功能上耦合到计算机视觉跟踪平台的称重平台中称量移动物体。 对象可以进行平移，旋转，平移和旋转。 物体的称重可以通过对象成像和上游称重的组合来实现。 对象成像可以允许通过计算机视觉跟踪从第一位置到第二位置的轨迹移动的逻辑对象，其中逻辑对象是一个或多个物理对象的正式表示。 上游称重可以允许更新指示与跟踪的逻辑对象相关联的一个或多个物理对象的权重的记录。 作为称重终端的一部分，可以在指示单个物理对象的权重的多个记录上执行数据完整性检查。 基于数据完整性检查的结果，可以提供表示单个物理对象的权重的记录。

公开(公告)号：US20080298672A1

公开(公告)日：2008-12-04

申请号：US11754948

申请日：2007-05-29

Applicant: Aaron S. Wallack ,  David J. Michael

Inventor： Aaron S. Wallack ,  David J. Michael

IPC: G06K9/00 ,  G06K9/62

CPC classification number: G06K9/32 ,  G06T7/73 ,  G06T2207/30164

Abstract: This invention provides a system and method for determining position of a viewed object in three dimensions by employing 2D machine vision processes on each of a plurality of planar faces of the object, and thereby refining the location of the object. First a rough pose estimate of the object is derived. This rough pose estimate can be based upon predetermined pose data, or can be derived by acquiring a plurality of planar face poses of the object (using, for example multiple cameras) and correlating the corners of the trained image pattern, which have known coordinates relative to the origin, to the acquired patterns. Once the rough pose is achieved, this is refined by defining the pose as a quaternion (a, b, c and d) for rotation and a three variables (x, y, z) for translation and employing an iterative weighted, least squares error calculation to minimize the error between the edgelets of trained model image and the acquired runtime edgelets. The overall, refined/optimized pose estimate incorporates data from each of the cameras' acquired images. Thereby, the estimate minimizes the total error between the edgelets of each camera's/view's trained model image and the associated camera's/view's acquired runtime edgelets. A final transformation of trained features relative to the runtime features is derived from the iterative error computation.

Abstract translation: 本发明提供了一种用于通过在对象的多个平面中的每一个上采用2D机器视觉过程来确定三维视图对象的位置的系统和方法，从而改善对象的位置。 首先推导出对象的粗略姿态估计。 这种粗略姿态估计可以基于预定的姿态数据，或者可以通过获取对象的多个平面脸部姿势（使用例如多个摄像机）并且将已训练图像图案的角部相关联来获得，其具有已知的坐标相对 到原产地，获得的模式。 一旦实现了粗糙的姿态，就通过将姿态定义为旋转的四元数（a，b，c和d）和用于平移的三个变量（x，y，z）并采用迭代加权最小二乘误差 计算以最小化训练模型图像的边缘与所获取的运行时间边缘之间的误差。 整体，精细/优化的姿态估计包含来自每个摄像机所获取图像的数据。 因此，估计使每个相机/视图的训练模型图像的边缘与相关联的相机/视图的获取的运行时间边缘之间的总误差最小化。 相对于运行时特征的训练特征的最终变换是从迭代误差计算得出的。

公开(公告)号：US5912984A

公开(公告)日：1999-06-15

申请号：US769739

申请日：1996-12-19

Applicant: David J. Michael ,  Juha Koljonen ,  Sanjay Nichani ,  Peter Roberts

Inventor： David J. Michael ,  Juha Koljonen ,  Sanjay Nichani ,  Peter Roberts

IPC: B23K3/08 ,  G06T7/00 ,  H05K1/02 ,  H05K3/12 ,  H05K3/34 ,  G06K9/00

CPC classification number: G06T7/001 ,  B23K3/08 ,  G06T7/0028 ,  G06T7/2053 ,  H05K3/3484 ,  B23K2201/40 ,  G06T2207/30152 ,  H05K1/0269 ,  H05K2203/163 ,  H05K3/1216

Abstract: Inspection of solder paste on a printed circuit board using a before printing image (pre-image) to normalize an after printing image (post-image) of the printed circuit board. Existing lighting and optics used for alignment of the screen printing stencil to the printed circuit board are used for the solder paste inspection. A stencil in the screen printing process is also inspected using a before printing image (pre) to normalize an after printing image (post) of the stencil.

Abstract translation: 使用之前打印的图像（预图像）检查印刷电路板上的焊膏以使印刷电路板的后打印图像（后图像）归一化。 用于将丝网印刷模板对准印刷电路板的现有照明和光学装置用于焊膏检查。 丝网印刷过程中的模板也使用之前的打印图像（预）进行检查，以对印版图像（后）进行归一化。

公开(公告)号：US5825913A

公开(公告)日：1998-10-20

申请号：US503574

申请日：1995-07-18

Applicant: Fariborz Rostami ,  James Harald Clark ,  David J. Michael ,  David J. Wilson

Inventor： Fariborz Rostami ,  James Harald Clark ,  David J. Michael ,  David J. Wilson

IPC: G06T7/00 ,  G06K9/36

CPC classification number: G06T7/0042 ,  G06T2207/30148 ,  H01L2223/54453

Abstract: The invention can be used to find the orientation of a semiconductor wafer without wafer handling, i.e., in a non-contact manner. The invention uses knowledge of the position of a semiconductor wafer, and the position of an orientation feature of the wafer, to find the orientation of the wafer. According to the invention, a curved band image is formed that includes an image of an orientation feature. The curved band image is then transformed into a straight band image. The longitudinal position of the orientation feature is then determined in the coordinate system of the straight band image, which longitudinal position is then converted into an angular displacement in the coordinate system of the curved band image to provide the orientation of the wafer.

Abstract translation: 本发明可以用于找到半导体晶片的方位，而不需要晶片处理，即以非接触方式。 本发明使用半导体晶片的位置以及晶片的取向特征的位置的知识来找到晶片的取向。 根据本发明，形成包括取向特征的图像的弯曲带图像。 然后将弯曲的带状图像变换为直带图像。 然后在直带图像的坐标系中确定取向特征的纵向位置，然后将该纵向位置转换成在弯曲带图像的坐标系中的角位移以提供晶片的取向。

公开(公告)号：US4431744A

公开(公告)日：1984-02-14

申请号：US384319

申请日：1982-06-02

Applicant: Mark C. Kernion ,  David J. Michael

Inventor： Mark C. Kernion ,  David J. Michael

IPC: C04B35/01 ,  C04B35/52

CPC classification number: C04B35/013

Abstract: Carbon bonded refractory brick containing 1 to 6% by weight liquid thermosetting resin binder consisting of polyhydroxydiphenyl resin and a curing agent and the balance being refractory aggregate.

Abstract translation: 含有1〜6重量％液体热固性树脂粘合剂的碳粘结耐火砖，由聚羟基二苯基树脂和固化剂组成，余量为耐火骨料。

公开(公告)号：US08126260B2

公开(公告)日：2012-02-28

申请号：US11754948

申请日：2007-05-29

Applicant: Aaron S. Wallack ,  David J. Michael

Inventor： Aaron S. Wallack ,  David J. Michael

IPC: G06K9/00

CPC classification number: G06K9/32 ,  G06T7/73 ,  G06T2207/30164

Abstract: This invention provides a system and method for determining position of a viewed object in three dimensions by employing 2D machine vision processes on each of a plurality of planar faces of the object, and thereby refining the location of the object. First a rough pose estimate of the object is derived. This rough pose estimate can be based upon predetermined pose data, or can be derived by acquiring a plurality of planar face poses of the object (using, for example multiple cameras) and correlating the corners of the trained image pattern, which have known coordinates relative to the origin, to the acquired patterns. Once the rough pose is achieved, this is refined by defining the pose as a quaternion (a, b, c and d) for rotation and a three variables (x, y, z) for translation and employing an iterative weighted, least squares error calculation to minimize the error between the edgelets of trained model image and the acquired runtime edgelets. The overall, refined/optimized pose estimate incorporates data from each of the cameras' acquired images. Thereby, the estimate minimizes the total error between the edgelets of each camera's/view's trained model image and the associated camera's/view's acquired runtime edgelets. A final transformation of trained features relative to the runtime features is derived from the iterative error computation.

Abstract translation: 本发明提供了一种用于通过在对象的多个平面中的每一个上采用2D机器视觉过程来确定三维视图对象的位置的系统和方法，从而改善对象的位置。 首先推导出对象的粗略姿态估计。 这种粗略姿态估计可以基于预定的姿态数据，或者可以通过获取对象的多个平面脸部姿势（使用例如多个摄像机）并且将已训练图像图案的角部相关联来获得，其具有已知的坐标相对 到原产地，获得的模式。 一旦实现了粗糙的姿态，就通过将姿态定义为旋转的四元数（a，b，c和d）和用于平移的三个变量（x，y，z）并采用迭代加权最小二乘误差 计算以最小化训练模型图像的边缘与所获取的运行时间边缘之间的误差。 整体，精细/优化的姿态估计包含来自每个摄像机所获取图像的数据。 因此，估计使每个相机/视图的训练模型图像的边缘与相关联的相机/视图的获取的运行时间边缘之间的总误差最小化。 相对于运行时特征的训练特征的最终变换是从迭代误差计算得出的。

公开(公告)号：US20090274361A1

公开(公告)日：2009-11-05

申请号：US12113492

申请日：2008-05-01

Applicant: John W. Schwab ,  Gang Liu ,  David J. Michael

Inventor： John W. Schwab ,  Gang Liu ,  David J. Michael

IPC: G06K9/00

CPC classification number: H01L31/022425 ,  H01L21/681 ,  H01L21/682 ,  H01L31/18 ,  Y02E10/50

Abstract: A vision system is provided to determine a positional relationship between a photovoltaic device wafer on a platen and a printing element, such as a printing screen, on a remote side of the photovoltaic device wafer from the platen. A source emits ultraviolet light along a path that is transverse to a longitudinal axis of an aperture through the platen, and a diffuser panel is located along that path. A reflector directs the light from the diffuser panel toward the aperture. A video camera is located along the longitudinal axis of the aperture and produces an image using light received from the platen aperture, wherein some of that received light was reflected by the wafer. A band-pass filter is placed in front of the camera to block ambient light. The use of diffused ultraviolet light enhances contrast in the image between the wafer and the printing element.

Abstract translation: 提供了一种视觉系统，用于确定平台上的光伏器件晶片与光伏器件晶片远离印版的远程侧的印刷元件（例如印刷丝网）之间的位置关系。 源沿着与穿过压板的孔的纵向轴线横向的路径发射紫外光，并且漫射板位于沿着该路径的位置。 反射器将来自扩散板的光引向光圈。 摄像机沿着孔的纵向轴线定位，并使用从压板孔接收的光产生图像，其中一些接收的光被晶片反射。 带通滤波器放置在相机的前面以阻挡环境光。 使用扩散的紫外光增强了晶片和印刷元件之间的图像的对比度。

公开(公告)号：US06421458B2

公开(公告)日：2002-07-16

申请号：US09141932

申请日：1998-08-28

Applicant: David J. Michael ,  Igor Reyzin

Inventor： David J. Michael ,  Igor Reyzin

IPC: G06K932

CPC classification number: G06K9/00 ,  G06T7/001 ,  G06T7/33 ,  G06T2207/30108

Abstract: During statistical training and automated inspection of objects by a machine vision system, a General Affine Transform is advantageously employed to improve system performance. During statistical training, the affine poses of a plurality of training images are determined with respect to an alignment model image. Following filtering to remove high frequency content, the training images and their corresponding affine poses are applied to an affine transformation. The resulting transformed images are accumulated to compute template and threshold images to be used for run-time inspection. During run-time inspection, the affine pose of the run-time image relative to the alignment model image is determined. Following filtering of the run-time image, the run-time image is affine transformed by its affine pose. The resulting transform image is compared with the template and threshold images computed during statistical training to determine object status. In this manner, automated training and inspection is relatively less demanding on system storage, and results in an improvement in system speed and accuracy.

Abstract translation: 在通过机器视觉系统的统计训练和物体的自动检查期间，有利地采用通用仿射变换来提高系统性能。 在统计训练期间，相对于对准模型图像确定多个训练图像的仿射姿态。 进行滤波以去除高频内容后，将训练图像及其对应的仿射姿态应用于仿射变换。 累积所得到的变换图像以计算用于运行时检查的模板和阈值图像。 在运行时检查期间，确定运行时图像相对于对准模型图像的仿射姿态。 在运行时图像过滤之后，运行时图像通过其仿射姿态进行仿射变换。 将所得到的变换图像与在统计训练期间计算的模板和阈值图像进行比较，以确定对象状态。 以这种方式，自动化培训和检查对系统存储的要求相对较低，并且导致系统速度和精度的提高。

公开(公告)号：US6141033A

公开(公告)日：2000-10-31

申请号：US856920

申请日：1997-05-15

Applicant: David J. Michael ,  Philip B. Romanik

Inventor： David J. Michael ,  Philip B. Romanik

IPC: G06K9/38 ,  H04N7/26 ,  H04N5/14

CPC classification number: G06K9/38 ,  H04N19/85 ,  H04N19/30

Abstract: The invention provides methods and apparatus for reducing the bandwidth of a multichannel image. In one aspect, the methods and apparatus call for acquiring a multichannel training image representing a training scene. Weighting factors for the respective channels are determined based on the contrast at corresponding locations in that multichannel training image. A reduced bandwidth runtime image is generated from the multichannel runtime images as a function of (i) the weighting factors determined from the training image and (ii) a multichannel image representing the runtime scene.

Abstract translation: 本发明提供了减少多通道图像的带宽的方法和装置。 一方面，所述方法和装置要求获取表示训练场景的多通道训练图像。 基于该多通道训练图像中相应位置处的对比度来确定相应通道的加权系数。 根据（i）从训练图像确定的加权因子和（ii）表示运行时场景的多通道图像，从多通道运行时映像生成减少带宽的运行时间映像。