WO2017107564A1 - 板卡图像获取方法和系统 - Google Patents

板卡图像获取方法和系统 Download PDF

Info

Publication number
WO2017107564A1
WO2017107564A1 PCT/CN2016/098232 CN2016098232W WO2017107564A1 WO 2017107564 A1 WO2017107564 A1 WO 2017107564A1 CN 2016098232 W CN2016098232 W CN 2016098232W WO 2017107564 A1 WO2017107564 A1 WO 2017107564A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
mark2
region
board
standard
Prior art date
Application number
PCT/CN2016/098232
Other languages
English (en)
French (fr)
Inventor
罗汉杰
Original Assignee
广州视源电子科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州视源电子科技股份有限公司 filed Critical 广州视源电子科技股份有限公司
Publication of WO2017107564A1 publication Critical patent/WO2017107564A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/22Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0004Industrial image inspection
    • G06T7/001Industrial image inspection using an image reference approach
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/10Image acquisition
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10004Still image; Photographic image
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10024Color image
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30108Industrial image inspection
    • G06T2207/30141Printed circuit board [PCB]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00Indexing scheme relating to image or video recognition or understanding
    • G06V2201/06Recognition of objects for industrial automation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07Target detection

Definitions

  • the present invention relates to the field of automatic optical detection technology, and in particular, to a board image acquisition method and system.
  • the board when the factory is assembling the board (for example, PCB board), the board is usually placed on the assembly line, and the image of the board is acquired on the assembly line by using the AOI (Automatic Optic Inspection) system. Then, the background information of the pipeline is filtered out from the image to obtain an image of the board.
  • AOI Automatic Optic Inspection
  • a common practice is to place the board on a conveyor belt or tray of the assembly line, mount the camera directly above the conveyor belt, and use sensors to detect the position of the board. When the sensor detects that the board enters the shooting area of the camera, the camera is used to capture the image of the area where the board is located.
  • the prior art has the following disadvantages: an external sensor is required to detect the position of the board, which increases the complexity and cost of the system, and can only acquire an image of one board at a time, and the image acquisition efficiency is low.
  • a method for acquiring a board image includes the following steps:
  • the board image acquisition method further includes the following steps:
  • a fourth image of the card is obtained from the third image based on the coordinates.
  • a board image acquisition system comprising:
  • a first acquiring module configured to acquire a first image of an area where the board is located at a preset time interval; wherein the first image includes at least one complete board;
  • a first calculation module configured to calculate a first matching matrix of the pre-stored standard map in the first image; wherein the standard map is a reference map of the board;
  • a second calculating module configured to calculate a first target element having the smallest value in the first matching matrix, and acquire a first coordinate of the first target element
  • An intercepting module configured to set the first coordinate to an upper left coordinate of a second image of the card if the value of the first target element is less than a preset threshold, and according to a size and a size of the standard image The upper left corner coordinate intercepts a second image of the card from the first image.
  • the board image acquisition system further includes:
  • a setting module configured to set a pixel value of an element of the second image corresponding region in the first image to 0, to obtain a third image
  • a fourth calculating module configured to calculate a second matching matrix of the pre-stored standard map in the third image
  • a third acquiring module configured to calculate a second target element with the smallest value in the second matching matrix, and acquire coordinates of the second target element
  • a fourth acquiring module configured to: if the value of the second target element is less than a preset threshold, according to The coordinates obtain a fourth image of the card from the third image.
  • the method and system for acquiring a card image by acquiring a first image of a region where the card is located at a preset rate, calculating a first matching matrix of the pre-stored standard image in the first image, and calculating a first matching matrix a first target element having a smallest value, and acquiring a first coordinate of the first target element, and setting the first coordinate to a second of the board when a value of the first target element is less than a preset threshold
  • the coordinates of the upper left corner of the image, and the second image of the board is intercepted from the first image according to the size of the standard map and the upper left corner coordinate, and no additional sensor is needed to detect the position of the board, and the complexity is low. ,low cost.
  • the third image is obtained by setting the pixel value of the detected board to 0, and calculating the second matching of the pre-stored standard map in the third image. a matrix, calculating a second target element having the smallest value in the second matching matrix, and acquiring coordinates of the second target element, if the value of the second target element is less than a preset threshold, according to the coordinate from the first.
  • the fourth image of the board is obtained in the three images until the board is not detected in the third image, and images of the plurality of boards can be acquired each time, and the image acquisition efficiency is high.
  • 1 is a flow chart of a method for acquiring a board image of an embodiment
  • Figure 2 is a schematic diagram of marked points and corresponding areas in a standard diagram of an embodiment
  • FIG. 3 is a schematic structural diagram of a board image acquisition system according to an embodiment.
  • FIG. 1 is a flow chart of a method for acquiring a board image of an embodiment. As shown in FIG. 1, the board image acquisition method may include the following steps:
  • S1 Acquire a first image of a region where the card is located at a preset time interval; wherein the first image includes at least one complete card;
  • the first image I obj of the card can be acquired by the camera.
  • One or more boards may be included in the first image I obj .
  • the first image I obj may be a grayscale image.
  • the image acquired by the camera may be a multi-channel color image, and the color image may be first converted into a grayscale image.
  • the camera may acquire a first image of a region where the card is located at a preset time interval. For example, when the card is located on the conveyor belt, the first image of the conveyor belt where the card is located may be acquired.
  • the preset time interval should satisfy the following conditions:
  • v is the moving speed of the board
  • t is the preset time interval
  • S is the length of the shooting area of the camera, wherein the length of the shooting area of the camera should be greater than the width of the board.
  • an image of the card may be first acquired and pre-stored in the memory as a standard I model required for reference.
  • a first matching matrix of the pre-stored standard map I model in the first image I obj can be calculated. Assuming that the direction of the card in the first image I obj is substantially the same as the direction of the card in the standard figure I model , the first matching matrix may be calculated according to the following formula:
  • R(x,y) is the first matching matrix
  • w model and h model are the width and height of the standard Figure I model
  • I model (i,j) and I obj (x+i,y+ j) is the pixel value of the element of the i-th row and the j-th column in the standard figure I model
  • the pixel value of the element of the x+ith row and the y+jth column in the first image Iobj respectively.
  • the first matching matrix may also be calculated according to other manners, and the specific calculation manner will not affect the implementation manner of the subsequent board image acquisition method, which is described here.
  • the first target element with the smallest value in the first matching matrix can be calculated, namely:
  • R(x, y) represents the first matching matrix and r represents the first target element.
  • the second image I tar of the card may be obtained from the first image I obj ; otherwise, if r is greater than or equal to a preset threshold r threshold , the first image I may be considered as The second image I tar of the board is not found in obj .
  • the threshold r threshold can be set according to experience, and can be calculated by using a corresponding algorithm. The method for obtaining the threshold r threshold will not affect the implementation manner of the subsequent board image acquisition method.
  • the board in the position of the second image I tar displacement may occur.
  • the second image I tar can be registered .
  • the second image I tar can be registered in the following manner:
  • the first marker point image and the second marker point image may be acquired in the standard diagram I model .
  • two patterns can be taken in the standard figure I model as the first point image I mark1 and the second point image I mark2, respectively .
  • first marker point image I mark1 and the second marker point image I mark2 may be enlarged in a certain proportion (for example, the ratio may be 0.2 times), and the enlarged first marker point image I mark1 and The second marker point image is obtained to obtain a first region P mark1 and a second region P mark2 ; wherein the coordinates of the first region can be recorded as (x mark1 , y mark1 , w mark1 , h mark1 ), the second region
  • the coordinates can be written as (x mark2 , y mark2 , w mark2 , h mark2 ), (x mark1 , y mark1 ) and (x mark2 , y mark2 ) respectively represent the upper left corner coordinate of the first region P mark1 and the second region
  • the first marker point image I mark1 and the second marker point image I mark2 may be a standard exposed copper circle on the card, or may be two patterns arbitrarily set by the user.
  • the first marker point image I mark1 and the second marker point image I mark2 may be sharp, the position is relatively fixed, and there is no image of confusing marker points around.
  • the first marker point and the second marker point may take two points with a large linear distance, for example, two points on the diagonal of the card may be taken.
  • the marked points and corresponding areas in the standard Figure I model are shown in FIG. 2.
  • the third region I search1 and the fourth region I search2 corresponding to the first region P mark1 and the second region P mark2 may be intercepted in the second image I tar , and may be respectively the upper left coordinate of the first region P mark1 and
  • the upper left corner coordinate of the second region P mark2 is that the template finds the upper left corner coordinate of the corresponding third region P' mark1 and the upper left corner coordinate of the fourth region P' mark2 in the third region I search1 and the fourth region I search2 , Its coordinates are recorded as (x' mark1 , y' mark1 ) and (x' mark2 , y' mark2 );
  • the coordinates of the upper left corner may be P mark1 first region, the second upper-left coordinates P mark2 region, the third region P 'mark1 left corner and the fourth region P' left corner of calculating the standard FIG mark2
  • a transformation matrix H between the I model and the second image I tar can be calculated according to the following formula:
  • x 0 x mark2 -(x' mark2 cos ⁇ -y' mark2 sin ⁇ )scalar
  • y 0 y mark2 -(x' mark2 sin ⁇ +y' mark2 cos ⁇ )scalar
  • H is the transformation matrix
  • scalar is the scaling between the second image I tar and the standard image I model
  • is the rotation angle between the second image I tar and the standard image I model
  • [x 0 , y 0 ] is the translation vector between the second image I tar and the standard image I model
  • (x mark1 , y mark1 ) and (x mark2 , y mark2 ) are the first region P mark1 and the second in the standard figure I model , respectively.
  • the upper left corner coordinates of the region P mark2 , (x' mark1 , y' mark1 ) and (x' mark2 , y' mark2 ) are the upper left corners of the third region P' mark1 and the fourth region P' mark2 in the second image I tar coordinate.
  • Matrix H can be registered to the position of the second image based on the conversion of I tar. For example, registration can be performed using the function warpAffine in the OpenCV (Open Source Computer Vision Library) library.
  • OpenCV Open Source Computer Vision Library
  • the pixel value of the element I obj (x, y) corresponding to the second image I tar may be set to 0 to obtain a third image I′ obj ; , x ⁇ [x best , x best +w model ), y ⁇ [y best , y best +h model ), (x best , y best ) is the smallest value among the first matching matrix R(x, y)
  • the coordinates of the first target element, w model and h model are the width and height of the standard figure I model , respectively.
  • a pre-stored standard map I model is obtained in the second image I' obj as a second matching matrix R'(x, y). Then, the second target element having the smallest value among the second matching matrix R'(x, y) can be calculated, namely:
  • R'(x, y) is the second matching matrix and r' is the second target element.
  • FIG. 3 is a schematic structural diagram of a board image acquisition system according to an embodiment. As shown in FIG. 3, the board image acquisition system may include:
  • the first acquiring module 10 is configured to acquire a first image of a region where the card is located every preset time interval; wherein the first image includes at least one complete card;
  • the first acquisition module 10 can acquire the first image I obj of the card.
  • the first acquisition module 10 can be a camera.
  • One or more boards may be included in the first image I obj .
  • the first image I obj may be a grayscale image.
  • the image acquired by the first acquisition module 10 may be a multi-channel color image, and the color image may be first converted into a grayscale image.
  • the camera may acquire a first image of a region where the card is located at a preset time interval. For example, when the card is located on the conveyor belt, the first image of the conveyor belt where the card is located may be acquired. In order to ensure that an image of each board can be acquired, the preset time interval should satisfy the condition given by the formula (1).
  • a first calculation module 20 configured to calculate a first matching matrix of the pre-stored standard map in the first image; wherein the standard map is a reference map of the board;
  • the acquisition unit of the first calculation module 20 can acquire an image I model of the board and pre-stored in the memory as a standard diagram required for reference.
  • the computing unit of the first computing module 20 may calculate a first matching matrix of the pre-stored standard map I model in the first image I obj . Assuming that the direction of the card in the first image I obj is substantially the same as the direction of the card in the standard FIG. 1 model , the first matching matrix can be calculated according to formula (2).
  • the first matching matrix may also be calculated according to other manners, and the specific calculation manner will not affect the implementation manner of the subsequent board image acquisition method, which is described here.
  • a second calculation module 30 configured to calculate a first target element having the smallest value in the first matching matrix, and acquire a first coordinate of the first target element
  • the second calculating module 30 may calculate the first target element having the smallest value among the first matching matrices according to formula (3).
  • the intercepting module 40 is configured to set the first coordinate to an upper left coordinate of the second image of the card if the value of the first target element is less than a preset threshold, and according to the size of the standard image And capturing the second image of the card from the first image with the upper left coordinate.
  • the intercepting module 40 may obtain the second image I tar of the card from the first image I obj ; otherwise, if r is greater than or equal to a preset threshold r threshold , the first The second image I tar of the card is not searched in an image I obj .
  • the threshold r threshold can be set according to experience, and can be calculated by using a corresponding algorithm. The method for obtaining the threshold r threshold will not affect the implementation manner of the subsequent board image acquisition method.
  • the board position in the second image I tar may be displaced compared to the board in the standard figure I model .
  • the second image I tar can be registered .
  • the board image acquisition system may further include:
  • a second acquiring module configured to acquire the first point image and the second point image in the standard figure I model .
  • two patterns can be taken as the point image in the standard figure I model , which are respectively recorded as the first point image I mark1 and the second point image I mark2 .
  • Generating module for generating a first region and a second region P mark1 P mark2 image according to the first marker and the second marker I mark1 image I mark2, and acquires the top left corner of the first region and the second region P mark1 of P mark2 coordinate.
  • the first marker point image I mark1 and the second marker point image I mark2 may be enlarged in a certain proportion (for example, the ratio may be 0.2 times), and the enlarged first marker point image I mark1 and The second mark point image is obtained to obtain a first area P mark1 and a second area P mark2 ; wherein the coordinates of the first area can be recorded as (x mark1 , y mark1 , w mark1 , h mark1 ), the second area
  • the coordinates can be written as (x mark2 , y mark2 , w mark2 , h mark2 ), (x mark1 , y mark1 ) and (x mark2 , y mark2 ) respectively represent the upper left corner coordinate of the first region P mark1 and the second region P
  • the upper left corner coordinates of mark2 , w mark1 and w mark2 respectively indicate the widths of the first region P mark1 and the second region P mark2
  • h mark1 and h mark2 indicate the heights of the first region P
  • the first marker point image I mark1 and the second marker point image I mark2 may be a standard exposed copper circle on the card, or may be two patterns arbitrarily set by the user.
  • the first marker point image I mark1 and the second marker point image I mark2 may be sharp, the position is relatively fixed, and there is no image of confusing marker points around.
  • the first marker point and the second marker point may take two points with a large linear distance, for example, two points on the diagonal of the card may be taken.
  • a search module configured to intercept, in the second image I tar , a third region I search1 and a fourth region I search2 corresponding to the first region P mark1 and the second region P mark2 , and may respectively use the first region P mark1
  • the upper left corner coordinate and the upper left corner coordinate of the second area P mark2 are templates for finding the upper left corner coordinate of the corresponding third area P' mark1 and the fourth area P' mark2 in the third area I search1 and the fourth area I search2
  • Upper left coordinates, the coordinates are denoted by (x 'mark1, y' mark1 ) and (x 'mark2, y' mark2 );
  • a third calculation module configured to: coordinate an upper left corner of the first region P mark1, an upper left coordinate of the second region P mark2, an upper left coordinate of the third region P′ mark1, and an upper left corner of the fourth region P′ mark2
  • the transformation calculates a transformation matrix H between the standard image I model and the second image I tar .
  • the transformation matrix can be calculated according to formula (4).
  • a registration module for registering the position of the second image I tar according to the transformation matrix H.
  • registration can be performed using the function warpAffine in the OpenCV (Open Source Computer Vision Library) library.
  • the pixel value of the element I obj (x, y) corresponding to the second image I tar may be set to 0 by the setting module to obtain the third image I′.
  • Obj x ⁇ [x best , x best +w model ), y ⁇ [y best , y best +h model ), (x best , y best ) is taken from the first matching matrix R(x, y)
  • the coordinates of the first target element with the smallest value, w model and h model are the width and height of the standard figure I model , respectively.
  • FIG calculation standard may be pre-stored by the fourth calculation module I model in the third image I 'obj second matching matrix R' (x, y). Then, the second target element having the smallest value among the second matching matrix R'(x, y) is calculated by the third obtaining module, as shown in the formula (5), and the coordinates of the second target element are obtained, and finally And if the value of the second target element is less than a preset threshold, the fourth acquiring module may obtain the fourth image I′ tar of the board from the third image I′ obj according to the coordinate;
  • the functions of the setting module, the fourth calculating module, the third obtaining module, and the fourth obtaining module may be executed cyclically until the board is not found in the third image I'obj .
  • the board image acquisition system of the present invention has a one-to-one correspondence with the board image acquisition method of the present invention, and the technical features and the beneficial effects described in the embodiments of the board image acquisition method are applicable to the embodiment of the board image acquisition system. In this regard, hereby declare.

Abstract

本发明涉及一种板卡图像获取方法和系统,其中,方法包括以下步骤:每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。所述板卡图像获取方法和系统复杂度低、成本低、效率高。

Description

板卡图像获取方法和系统 技术领域
本发明涉及自动光学检测技术领域,特别是涉及一种板卡图像获取方法和系统。
背景技术
现阶段工厂在对板卡(例如,PCB板)进行组装等处理时,一般是先将板卡放置在流水线上,采用AOI(Automatic Optic Inspection,自动光学检测)系统获取板卡在流水线上的图像,再从图像中过滤掉流水线的背景信息,得到板卡的图像。在AOI系统中,一种常用的做法是将板卡放置在流水线的传送带或托盘之上,将摄像头架设在传送带正上方,并采用传感器对板卡的位置进行检测。当传感器检测到板卡进入摄像头的拍摄区域时,采用摄像头对板卡所在区域的图像进行拍摄。
现有技术具有以下缺点:需要外部传感器检测板卡的位置,增加了系统的复杂度和成本,且每次只能获取一张板卡的图像,图像获取效率低。
发明内容
基于此,有必要针对现有技术复杂度高、成本高和效率低的问题,提供一种板卡图像获取方法和系统。
一种板卡图像获取方法,包括以下步骤:
每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;
计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;
计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;
如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所述板 卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。
所述板卡图像获取方法还包括以下步骤:
将所述第一图像中所述第二图像对应区域的元素的像素值设为0,得到第三图像;
计算预存的标准图在所述第三图像中的第二匹配矩阵;
计算第二匹配矩阵中取值最小的第二目标元素,并获取所述第二目标元素的坐标;
如果所述第二目标元素的值小于所述阈值,根据所述坐标从第三图像中获得所述板卡的第四图像。
一种板卡图像获取系统,包括:
第一获取模块,用于每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;
第一计算模块,用于计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;
第二计算模块,用于计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;
截取模块,用于如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。
所述板卡图像获取系统还包括:
设置模块,用于将所述第一图像中所述第二图像对应区域的元素的像素值设为0,得到第三图像;
第四计算模块,用于计算预存的标准图在所述第三图像中的第二匹配矩阵;
第三获取模块,用于计算第二匹配矩阵中取值最小的第二目标元素,并获取所述第二目标元素的坐标;
第四获取模块,用于如果所述第二目标元素的取值小于预设的阈值,根据 所述坐标从第三图像中获得所述板卡的第四图像。
上述板卡图像获取方法和系统,通过以预设的速率获取板卡所在区域的第一图像,计算预存的标准图在所述第一图像中的第一匹配矩阵,计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标,当第一目标元素的值小于预设的阈值时,将所述第一坐标设为所述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像,无需设置额外的传感器来检测板卡的位置,复杂度低,成本低。
另外,当第一图像中包含多个板卡时,通过将已检测到的板卡的像素值设为0,得到第三图像,计算预存的标准图在所述第三图像中的第二匹配矩阵,计算第二匹配矩阵中取值最小的第二目标元素,并获取所述第二目标元素的坐标,如果所述第二目标元素的取值小于预设的阈值,根据所述坐标从第三图像中获得所述板卡的第四图像,直到所述第三图像中检测不到板卡为止,每次可以获取多张板卡的图像,图像获取效率高。
附图说明
图1为一个实施例的板卡图像获取方法流程图;
图2为一个实施例的标准图中的标记点和对应区域的示意图;
图3为一个实施例的板卡图像获取系统结构示意图。
具体实施方式
下面结合附图对本发明的板卡图像获取方法的实施例进行描述。
图1为一个实施例的板卡图像获取方法流程图。如图1所示,所述板卡图像获取方法可包括以下步骤:
S1,每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;
在本步骤中,可通过摄像头获取板卡的第一图像Iobj。所述第一图像Iobj中可包含一个或多个板卡。所述第一图像Iobj可以是灰度图像。在实际情况下, 通过摄像头获取到的图像可能是多通道彩色图像,可首先将彩色图像转换为灰度图。
所述摄像头可以每隔预设的时间间隔获取板卡所在区域的第一图像,例如,当板卡位于传送带上时,可获取板卡所在传送带的第一图像。为了保证能够获取到每张板卡的图像,所述预设的时间间隔应满足以下条件:
vt≤S                      (1)
式中,v为板卡的移动速度,t为所述预设的时间间隔,S为摄像头的拍摄区域的长度,其中,摄像头的拍摄区域的长度应大于所述板卡的宽度。
S2,计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;
在本步骤中,可首先获取所述板卡的一张图像,并预存在存储器中,作为参照所需的标准图Imodel。可计算预存的标准图Imodel在所述第一图像Iobj中的第一匹配矩阵。假设所述第一图像Iobj中板卡的方向与所述标准图Imodel中板卡的方向基本一致,可根据如下公式计算所述第一匹配矩阵:
Figure PCTCN2016098232-appb-000001
式中,R(x,y)为第一匹配矩阵,wmodel和hmodel分别为所述标准图Imodel的宽和高,Imodel(i,j)和Iobj(x+i,y+j)分别为标准图Imodel中第i行第j列的元素的像素值和第一图像Iobj中第x+i行第y+j列的元素的像素值。
在实际情况下,还可根据其他方式计算所述第一匹配矩阵,具体的计算方式将不会影响后续板卡图像获取方法的实施方式,特此说明。
S3,计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;
在本步骤中,可计算第一匹配矩阵中取值最小的第一目标元素,即:
r=min(R(x,y))                 (3)
式中,R(x,y)表示所述第一匹配矩阵,r表示所述第一目标元素。
S4,如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所 述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。
如果r小于预设的阈值rthreshold,可从第一图像Iobj中获取所述板卡的第二图像Itar;否则,如果r大于或等于预设的阈值rthreshold,可认为第一图像Iobj中没有搜索到所述板卡的第二图像Itar。其中,所述阈值rthreshold可根据经验设定,也可采用相应的算法来计算,阈值rthreshold的获取方式将不会影响后续板卡图像获取方法的实施方式,特此说明。
由于摆放的原因,与标准图Imodel中的板卡相比,第二图像Itar中的板卡位置可能发生位移。为了消除所述位移引起的误差,可对第二图像Itar进行配准操作。在一个实施例中,可采用以下方式对第二图像Itar进行配准操作:
可在所述标准图Imodel中获取第一标记点图像和第二标记点图像。例如,可在所述标准图Imodel中截取两个图案分别设为第一标记点图像Imark1和第二标记点图像Imark2
然后,可根据第一标记点图像Imark1和第二标记点图像Imark2生成第一区域Pmark1和第二区域Pmark2,并获取第一区域Pmark1和第二区域Pmark2的左上角坐标。例如,可对第一标记点图像Imark1和第二标记点图像Imark2进行一定比例的扩大(例如,所述比例可以是0.2倍),并框取扩大后的第一标记点图像Imark1和第二标记点图像,得到第一区域Pmark1和第二区域Pmark2;其中,所述第一区域的坐标可记为(xmark1,ymark1,wmark1,hmark1),所述第二区域的坐标可记为(xmark2,ymark2,wmark2,hmark2),(xmark1,ymark1)和(xmark2,ymark2)分别表示第一区域Pmark1的左上角坐标和第二区域Pmark2的左上角坐标,wmark1和wmark2分别表示第一区域Pmark1和第二区域Pmark2的宽,hmark1和hmark2分别表示第一区域Pmark1和第二区域Pmark2的高。所述第一标记点图像Imark1和第二标记点图像Imark2可以是板卡上的标准露铜圆形,也可以是用户任意设定的两个图案。所述第一标记点图像Imark1和第二标记点图像Imark2可以是清晰,位置相对固定,并且周围没有易混淆的标记点的图像。第一标记点和第二标记点可以取直线距离较大的两个点,例如,可以取板卡的对角线上的两个点。所述标准图Imodel中的标记点和对应区域如图2所示。
可在所述第二图像Itar中截取第一区域Pmark1和第二区域Pmark2对应的第三区 域Isearch1和第四区域Isearch2,并可分别以第一区域Pmark1的左上角坐标和第二区域Pmark2的左上角坐标为模板在第三区域Isearch1和第四区域Isearch2中找出对应的第三区域P′mark1的左上角坐标和第四区域P′mark2的左上角坐标,其坐标分别记为(x'mark1,y'mark1)和(x'mark2,y'mark2);
可根据所述第一区域Pmark1的左上角坐标、第二区域Pmark2的左上角坐标、第三区域P′mark1的左上角坐标和第四区域P′mark2的左上角坐标计算所述标准图Imodel与所述第二图像Itar之间的变换矩阵H。可根据以下公式计算所述变换矩阵:
Figure PCTCN2016098232-appb-000002
其中,
Figure PCTCN2016098232-appb-000003
x0=xmark2-(x'mark2cosθ-y'mark2sinθ)scalar,
y0=ymark2-(x'mark2sinθ+y'mark2cosθ)scalar,
θ=arctan(ymark2-ymark1,xmark2-xmark1)-arctan(y'mark2-y'mark1,x'mark2-x'mark1)
式中,H为所述变换矩阵,scalar为第二图像Itar与标准图Imodel之间的缩放比例,θ为第二图像Itar与标准图Imodel之间的旋转角度,[x0,y0]为第二图像Itar与标准图Imodel之间的平移向量,(xmark1,ymark1)和(xmark2,ymark2)分别为标准图Imodel中第一区域Pmark1和第二区域Pmark2的左上角坐标,(x'mark1,y'mark1)和(x'mark2,y'mark2)为第二图像Itar中第三区域P′mark1和第四区域P′mark2的左上角坐标。
可根据所述变换矩阵H对所述第二图像Itar的位置进行配准。例如,可使用OpenCV(Open Source Computer Vision Library,基于开源的跨平台计算机视觉库)库中的函数warpAffine进行配准。
由于第一图像Iobj中可能有多张板卡,可将所述第二图像Itar对应区域的元素Iobj(x,y)的像素值设为0,得到第三图像I'obj;其中,x∈[xbest,xbest+wmodel),y∈[ybest,ybest+hmodel),(xbest,ybest)为第一匹配矩阵R(x,y)中取值最小的第一目标元素的坐标,wmodel和hmodel分别为所述标准图Imodel的宽和高。可获得预存的标准图 Imodel在所述第三图像I'obj中的第二匹配矩阵R'(x,y)。然后,可计算第二匹配矩阵R'(x,y)中取值最小的第二目标元素,即:
r'=min(R'(x,y))                     (5)
式中,R'(x,y)为第二匹配矩阵,r'为第二目标元素。
可获取所述第二目标元素的坐标,如果所述第二目标元素的值小于预设的阈值,根据所述坐标从第三图像I'obj中获得所述板卡的第四图像I′tar;可循环执行本步骤,直到第三图像I'obj中找不到板卡为止。
本发明的板卡图像获取方法具有以下优点:
(1)通过算法实现板卡检测,无需外部设备配合,使用方便,复杂度低,能有效地降低成本。
(2)可基于标记点信息对板卡为止进行精确配准,精度高,速度快。
(3)每次可获取多张板卡的图像,图像获取效率高。
下面结合附图对本发明的板卡图像获取系统的实施例进行描述。
图3为一个实施例的板卡图像获取系统结构示意图。如图3所示,所述板卡图像获取系统可包括:
第一获取模块10,用于每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;
第一获取模块10可获取板卡的第一图像Iobj。所述第一获取模块10可以是摄像头。所述第一图像Iobj中可包含一个或多个板卡。所述第一图像Iobj可以是灰度图像。在实际情况下,通过第一获取模块10获取到的图像可能是多通道彩色图像,可首先将彩色图像转换为灰度图。
所述摄像头可以每隔预设的时间间隔获取板卡所在区域的第一图像,例如,当板卡位于传送带上时,可获取板卡所在传送带的第一图像。为了保证能够获取到每张板卡的图像,所述预设的时间间隔应满足公式(1)给出的条件。
第一计算模块20,用于计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;
第一计算模块20的获取单元可获取所述板卡的一张图像Imodel,并预存在存储器中,作为参照所需的标准图。第一计算模块20的计算单元可计算预存 的标准图Imodel在所述第一图像Iobj中的第一匹配矩阵。假设所述第一图像Iobj中板卡的方向与所述标准图Imodel中板卡的方向基本一致,可根据公式(2)计算所述第一匹配矩阵。
在实际情况下,还可根据其他方式计算所述第一匹配矩阵,具体的计算方式将不会影响后续板卡图像获取方法的实施方式,特此说明。
第二计算模块30,用于计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;
第二计算模块30可根据公式(3)计算第一匹配矩阵中取值最小的第一目标元素。
截取模块40,用于如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。
如果r小于预设的阈值rthreshold,截取模块40可从第一图像Iobj中获取所述板卡的第二图像Itar;否则,如果r大于或等于预设的阈值rthreshold,可认为第一图像Iobj中没有搜索到所述板卡的第二图像Itar。其中,所述阈值rthreshold可根据经验设定,也可采用相应的算法来计算,阈值rthreshold的获取方式将不会影响后续板卡图像获取方法的实施方式,特此说明。
由于摆放的原因,与标准图Imodel中的板卡相比,第二图像Itar中的板卡位置可能发生位移。为了消除所述位移引起的误差,可对第二图像Itar进行配准操作。在一个实施例中,所述板卡图像获取系统还可包括:
第二获取模块,用于在所述标准图Imodel中获取第一标记点图像和第二标记点图像。例如,可在所述标准图Imodel中截取两个图案作为标记点图像,分别记为第一标记点图像Imark1和第二标记点图像Imark2
生成模块,用于根据第一标记点图像Imark1和第二标记点图像Imark2生成第一区域Pmark1和第二区域Pmark2,并获取第一区域Pmark1和第二区域Pmark2的左上角坐标。例如,可对第一标记点图像Imark1和第二标记点图像Imark2进行一定比例的扩大(例如,所述比例可以是0.2倍),并框取扩大后的第一标记点图像Imark1和第二标记点图像,得到第一区域Pmark1和第二区域Pmark2;其中,所述第一区域的坐 标可记为(xmark1,ymark1,wmark1,hmark1),所述第二区域的坐标可记为(xmark2,ymark2,wmark2,hmark2),(xmark1,ymark1)和(xmark2,ymark2)分别表示第一区域Pmark1的左上角坐标和第二区域Pmark2的左上角坐标,wmark1和wmark2分别表示第一区域Pmark1和第二区域Pmark2的宽,hmark1和hmark2分别表示第一区域Pmark1和第二区域Pmark2的高。所述第一标记点图像Imark1和第二标记点图像Imark2可以是板卡上的标准露铜圆形,也可以是用户任意设定的两个图案。所述第一标记点图像Imark1和第二标记点图像Imark2可以是清晰,位置相对固定,并且周围没有易混淆的标记点的图像。第一标记点和第二标记点可以取直线距离较大的两个点,例如,可以取板卡的对角线上的两个点。
查找模块,用于在所述第二图像Itar中截取第一区域Pmark1和第二区域Pmark2对应的第三区域Isearch1和第四区域Isearch2,并可分别以第一区域Pmark1的左上角坐标和第二区域Pmark2的左上角坐标为模板在第三区域Isearch1和第四区域Isearch2中找出对应的第三区域P′mark1的左上角坐标和第四区域P′mark2的左上角坐标,其坐标分别记为(x'mark1,y'mark1)和(x'mark2,y'mark2);
第三计算模块,用于根据所述第一区域Pmark1的左上角坐标、第二区域Pmark2的左上角坐标、第三区域P′mark1的左上角坐标和第四区域P′mark2的左上角坐标计算所述标准图Imodel与所述第二图像Itar之间的变换矩阵H。可根据公式(4)计算所述变换矩阵。
配准模块,用于根据所述变换矩阵H对所述第二图像Itar的位置进行配准。例如,可使用OpenCV(Open Source Computer Vision Library,基于开源的跨平台计算机视觉库)库中的函数warpAffine进行配准。
由于第一图像Iobj中可能有多张板卡,可通过设置模块将所述第二图像Itar对应区域的元素Iobj(x,y)的像素值设为0,得到第三图像I'obj;其中,x∈[xbest,xbest+wmodel),y∈[ybest,ybest+hmodel),(xbest,ybest)为第一匹配矩阵R(x,y)中取值最小的第一目标元素的坐标,wmodel和hmodel分别为所述标准图Imodel的宽和高。可通过第四计算模块计算预存的标准图Imodel在所述第三图像I'obj中的第二匹配矩阵R'(x,y)。然后,可通过第三获取模块计算第二匹配矩阵R'(x,y)中取值最小的第二目标元素,如公式(5)所示,并获取所述第二目标元素的坐标,最后, 如果所述第二目标元素的取值小于预设的阈值,第四获取模块可根据所述坐标从第三图像I'obj中获得所述板卡的第四图像I′tar
可循环执行设置模块、第四计算模块、第三获取模块和第四获取模块的功能,直到第三图像I'obj中找不到板卡为止。
本发明的板卡图像获取系统具有以下优点:
(1)通过算法实现板卡检测,无需外部设备配合,使用方便,能有效地降低成本。
(2)先用模版匹配找出板卡的大概位置,再用基于标记点信息的精确配准,精度高,速度快。
(3)每次可获取多张板卡的图像,图像获取效率高。
本发明的板卡图像获取系统与本发明的板卡图像获取方法一一对应,在上述板卡图像获取方法的实施例阐述的技术特征及其有益效果均适用于板卡图像获取系统的实施例中,特此声明。
以上所述实施例的各技术特征可以进行任意的组合,为使描述简洁,未对上述实施例中的各个技术特征所有可能的组合都进行描述,然而,只要这些技术特征的组合不存在矛盾,都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本发明的几种实施方式,其描述较为具体和详细,但并不能因此而理解为对发明专利范围的限制。应当指出的是,对于本领域的普通技术人员来说,在不脱离本发明构思的前提下,还可以做出若干变形和改进,这些都属于本发明的保护范围。因此,本发明专利的保护范围应以所附权利要求为准。

Claims (10)

  1. 一种板卡图像获取方法,其特征在于,包括以下步骤:
    每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;
    计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;
    计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;
    如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。
  2. 根据权利要求1所述的板卡图像获取方法,其特征在于,获取所述板卡的第二图像之后,还包括以下步骤:
    在所述标准图中获取第一标记点图像和第二标记点图像;
    根据第一标记点图像和第二标记点图像生成第一区域和第二区域,并获取所述第一区域和第二区域的左上角坐标;
    在所述第二图像中截取第一区域和第二区域对应的第三区域和第四区域,并可分别以第一区域的左上角坐标和第二区域的左上角坐标为模板在第三区域和第四区域中找出第三区域和第四区域的左上角坐标;
    根据所述第一区域、第二区域、第三区域和第四区域的左上角坐标计算所述标准图与所述第二图像之间的变换矩阵;
    根据所述变换矩阵对所述第二图像的位置进行配准。
  3. 根据权利要求2所述的板卡图像获取方法,其特征在于,根据所述第一标记点、第二标记点、第三标记点和第四标记点计算所述标准图与所述第二图像之间的变换矩阵的步骤包括:
    根据如下公式计算所述标准图与所述第二图像之间的变换矩阵:
    Figure PCTCN2016098232-appb-100001
    其中,
    Figure PCTCN2016098232-appb-100002
    x0=xmark2-(x'mark2cosθ-y'mark2sinθ)scalar,
    y0=ymark2-(x'mark2sinθ+y'mark2cosθ)scalar,
    θ=arctan(ymark2-ymark1,xmark2-xmark1)-arctan(y'mark2-y'mark1,x'mark2-x'mark1),
    式中,H为所述变换矩阵,scalar为第二图像与标准图之间的缩放比例,θ为第二图像与标准图之间的旋转角度,[x0,y0]为第二图像与标准图之间的平移向量,[xmark1,ymark1]和[xmark2,ymark2]分别为标准图中两个标记点所在区域的左上角坐标,[x'mark1,y'mark1]和[x'mark2,y'mark2]为第二图像中两个标记点所在区域的左上角坐标。
  4. 根据权利要求1所述的板卡图像获取方法,其特征在于,获取所述板卡的第二图像之后,还包括以下步骤:
    将所述第一图像中所述第二图像对应区域的元素的像素值设为0,得到第三图像;
    计算预存的标准图在所述第三图像中的第二匹配矩阵;
    计算第二匹配矩阵中取值最小的第二目标元素,并获取所述第二目标元素的坐标;
    如果所述第二目标元素的值小于所述阈值,根据所述坐标从第三图像中获得所述板卡的第四图像。
  5. 根据权利要求1所述的板卡图像获取方法,其特征在于,计算预存的标准图在所述第一图像中的第一匹配矩阵的步骤包括:
    根据如下公式计算预存的标准图在所述第一图像中的第一匹配矩阵:
    Figure PCTCN2016098232-appb-100003
    式中,R(x,y)为第一匹配矩阵,wmodel和hmodel分别为所述标准图的宽和高,Imodel(i,j)和Iobj(x+i,y+j)分别为标准图中第i行第j列的元素的像素值和第一图像中第x+i行第y+j列的元素的像素值。
  6. 一种板卡图像获取系统,其特征在于,包括:
    第一获取模块,用于每隔预设的时间间隔获取板卡所在区域的第一图像;其中,所述第一图像包括至少一张完整的板卡;
    第一计算模块,用于计算预存的标准图在所述第一图像中的第一匹配矩阵;其中,所述标准图为所述板卡的参照图;
    第二计算模块,用于计算第一匹配矩阵中取值最小的第一目标元素,并获取所述第一目标元素的第一坐标;
    截取模块,用于如果所述第一目标元素的值小于预设的阈值,将所述第一坐标设为所述板卡的第二图像的左上角坐标,并根据所述标准图的尺寸和所述左上角坐标从第一图像中截取所述板卡的第二图像。
  7. 根据权利要求6所述的板卡图像获取系统,其特征在于,还包括:
    第二获取模块,用于在所述标准图中获取第一标记点图像和第二标记点图像;
    生成模块,用于根据第一标记点图像和第二标记点图像生成第一区域和第二区域,并获取所述第一区域和第二区域的左上角坐标;
    查找模块,用于在所述第二图像中截取第一区域和第二区域对应的第三区域和第四区域,并可分别以第一区域的左上角坐标和第二区域的左上角坐标为模板在第三区域和第四区域中找出第三区域和第四区域的左上角坐标;
    第三计算模块,用于根据所述第一区域、第二区域、第三区域和第四区域的左上角坐标计算所述标准图与所述第二图像之间的变换矩阵;
    配准模块,用于根据所述变换矩阵对所述第二图像的位置进行配准。
  8. 根据权利要求7所述的板卡图像获取系统,其特征在于,所述第二计算模块根据以下公式计算所述变换矩阵:
    Figure PCTCN2016098232-appb-100004
    其中,
    Figure PCTCN2016098232-appb-100005
    x0=xmark2-(x'mark2cosθ-y'mark2sinθ)scalar,
    y0=ymark2-(x'mark2sinθ+y'mark2cosθ)scalar,
    θ=arctan(ymark2-ymark1,xmark2-xmark1)-arctan(y'mark2-y'mark1,x'mark2-x'mark1),
    式中,H为所述变换矩阵,scalar为第二图像与标准图之间的缩放比例,θ为第二图像与标准图之间的旋转角度,[x0,y0]为第二图像与标准图之间的平移向量,[xmark1,ymark1]和[xmark2,ymark2]分别为标准图中两个标记点所在区域的左上角坐标,[x'mark1,y'mark1]和[x'mark2,y'mark2]为第二图像中两个标记点所在区域的左上角坐标。
  9. 根据权利要求6所述的板卡图像获取系统,其特征在于,还包括:
    设置模块,用于将所述第一图像中所述第二图像对应区域的元素的像素值设为0,得到第三图像;
    第四计算模块,用于计算预存的标准图在所述第三图像中的第二匹配矩阵;
    第三获取模块,用于计算第二匹配矩阵中取值最小的第二目标元素,并获取所述第二目标元素的坐标;
    第四获取模块,用于如果所述第二目标元素的取值小于预设的阈值,根据所述坐标从第三图像中获得所述板卡的第四图像。
  10. 根据权利要求1所述的板卡图像获取系统,其特征在于,第一计算模块根据如下公式计算所述第一匹配矩阵:
    Figure PCTCN2016098232-appb-100006
    式中,R(x,y)为第一匹配矩阵,wmodel和hmodel分别为所述标准图的宽和高,Imodel(i,j)和Iobj(x+i,y+j)分别为标准图中第i行第j列的元素的像素值和第一图像中第x+i行第y+j列的元素的像素值。
PCT/CN2016/098232 2015-12-23 2016-09-06 板卡图像获取方法和系统 WO2017107564A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510990864.7A CN105469085B (zh) 2015-12-23 2015-12-23 板卡图像获取方法和系统
CN201510990864.7 2015-12-23

Publications (1)

Publication Number Publication Date
WO2017107564A1 true WO2017107564A1 (zh) 2017-06-29

Family

ID=55606756

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/098232 WO2017107564A1 (zh) 2015-12-23 2016-09-06 板卡图像获取方法和系统

Country Status (2)

Country Link
CN (1) CN105469085B (zh)
WO (1) WO2017107564A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105469085B (zh) * 2015-12-23 2019-03-26 广州视源电子科技股份有限公司 板卡图像获取方法和系统
CN106846389A (zh) * 2016-12-08 2017-06-13 国网北京市电力公司 机柜内液态汇聚部位的观测方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101738401A (zh) * 2008-11-11 2010-06-16 奥林巴斯株式会社 缺陷检查装置和缺陷检查方法
CN104504375A (zh) * 2014-12-18 2015-04-08 广州视源电子科技股份有限公司 一种pcb元件的识别方法及装置
CN104568985A (zh) * 2014-12-30 2015-04-29 东莞市合易自动化科技有限公司 一种aoi光学检测设备
CN105118086A (zh) * 2015-06-30 2015-12-02 东莞市盟拓光电科技有限公司 3d-aoi设备中的3d点云数据配准方法及系统
CN105469085A (zh) * 2015-12-23 2016-04-06 广州视源电子科技股份有限公司 板卡图像获取方法和系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1865950A (zh) * 2005-04-25 2006-11-22 王�琦 一种印刷电路板的自动对位方法
KR101595547B1 (ko) * 2010-11-20 2016-02-18 주식회사 고영테크놀러지 검사방법
CN102486829B (zh) * 2010-12-01 2016-05-04 罗普特(厦门)科技集团有限公司 图像分析系统及方法
CN102937595B (zh) * 2012-11-13 2015-05-20 浙江省电力公司电力科学研究院 一种pcb板检测方法、装置及系统
CN105118021B (zh) * 2015-07-28 2019-04-19 深圳市汇川技术股份有限公司 基于特征点的图像配准方法及系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101738401A (zh) * 2008-11-11 2010-06-16 奥林巴斯株式会社 缺陷检查装置和缺陷检查方法
CN104504375A (zh) * 2014-12-18 2015-04-08 广州视源电子科技股份有限公司 一种pcb元件的识别方法及装置
CN104568985A (zh) * 2014-12-30 2015-04-29 东莞市合易自动化科技有限公司 一种aoi光学检测设备
CN105118086A (zh) * 2015-06-30 2015-12-02 东莞市盟拓光电科技有限公司 3d-aoi设备中的3d点云数据配准方法及系统
CN105469085A (zh) * 2015-12-23 2016-04-06 广州视源电子科技股份有限公司 板卡图像获取方法和系统

Also Published As

Publication number Publication date
CN105469085B (zh) 2019-03-26
CN105469085A (zh) 2016-04-06

Similar Documents

Publication Publication Date Title
JP6934026B2 (ja) ビジョンシステムでラインを検出するためのシステム及び方法
US9124873B2 (en) System and method for finding correspondence between cameras in a three-dimensional vision system
KR102198352B1 (ko) Tiled Display에서 3D 영상을 보정하는 방법 및 장치
US20160371855A1 (en) Image based measurement system
CN110163912B (zh) 二维码位姿标定方法、装置及系统
CN106441138A (zh) 基于视觉测量的变形监测方法
CN110189322B (zh) 平整度检测方法、装置、设备、存储介质及系统
US11488322B2 (en) System and method for training a model in a plurality of non-perspective cameras and determining 3D pose of an object at runtime with the same
KR102649038B1 (ko) 비전 시스템으로 이미지에서 라인을 찾기 위한 시스템 및 방법
CA2507174A1 (en) Method of registering and aligning multiple images
CN108596867A (zh) 一种基于orb算法的图片校正方法及系统
JP2006148745A (ja) カメラの校正方法及びカメラの校正装置
CN110619662A (zh) 一种基于单目视觉的多行人目标空间连续定位方法及系统
CN113538583A (zh) 一种工件在机床上位置的精准定位方法及视觉系统
KR20120054689A (ko) 검사방법
WO2018142533A1 (ja) 位置姿勢推定装置および位置姿勢推定方法
CN111488775A (zh) 注视度判断装置及方法
CN108844462A (zh) 一种尺寸检测方法、装置、设备和系统
JP2009276233A (ja) パラメータ計算装置、パラメータ計算システムおよびプログラム
WO2017107564A1 (zh) 板卡图像获取方法和系统
CN103607558A (zh) 一种视频监控系统及其目标匹配方法和装置
KR102023087B1 (ko) 카메라 캘리브레이션 방법
JP2024016287A (ja) ビジョンシステムでラインを検出するためのシステム及び方法
WO2016141984A1 (en) Image processing device and method for geometric calibration of images
JP6304815B2 (ja) 画像処理装置ならびにその画像特徴検出方法、プログラムおよび装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16877375

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 16/11/2018)

122 Ep: pct application non-entry in european phase

Ref document number: 16877375

Country of ref document: EP

Kind code of ref document: A1