CN105457908B - Fast sorting based positioning method and system for small-size glass panel of monocular ccd - Google Patents

Fast sorting based positioning method and system for small-size glass panel of monocular ccd Download PDF

Info

Publication number
CN105457908B
CN105457908B CN201510771640.7A CN201510771640A CN105457908B CN 105457908 B CN105457908 B CN 105457908B CN 201510771640 A CN201510771640 A CN 201510771640A CN 105457908 B CN105457908 B CN 105457908B
Authority
CN
China
Prior art keywords
row
image
coordinates
glass panel
gray
Prior art date
Application number
CN201510771640.7A
Other languages
Chinese (zh)
Other versions
CN105457908A (en
Inventor
孙高磊
程涛
冯平
彭涛
刘新辉
Original Assignee
孙高磊
程涛
冯平
彭涛
刘新辉
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 孙高磊, 程涛, 冯平, 彭涛, 刘新辉 filed Critical 孙高磊
Priority to CN201510771640.7A priority Critical patent/CN105457908B/en
Publication of CN105457908A publication Critical patent/CN105457908A/en
Application granted granted Critical
Publication of CN105457908B publication Critical patent/CN105457908B/en

Links

Abstract

本发明适用于玻璃面板定位,提供了一种基于单目CCD小尺寸玻璃面板的分拣快速定位方法,步骤包括:A,利用单目CCD采集放置有玻璃面板的卡槽图像,将采集到的图像进行灰度转换然后进行预处理得到灰度图像;B,计算灰度图像的行像素灰度均值,确定玻璃面板所在行坐标;C,对灰度图像进行二值化分割,根据玻璃面板提取感兴趣区域边缘坐标;D,根据行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标,得到吸附位置,控制机械手达到吸附位置进行吸附。 The present invention is applicable to the glass panel is positioned, there is provided a sorting method based on rapid positioning monocular CCD small-size glass panel, comprising the step of: A, using the image acquisition slot glass panel placed monocular CCD, the collected gradation conversion image is then preprocessed to obtain a grayscale image; B, the gray image is calculated gray value pixel rows, the row coordinate of the glass panel is determined; C, the gray image is binarized segmentation extraction according to the glass panel coordinate edge region of interest; D, is determined according to the slot center coordinates and row coordinates of the region of interest edge coordinates, to give the suction position, the suction position to control the robot reaches adsorption. 本发明基于灰度转换和边缘检测,能够迅速找到视野内位于卡槽的每块玻璃面板的中心位置,同时结合相机视野正中心位置,可以快速寻找到当前需要抓取的玻璃面板,实现小尺寸面板的分拣快速定位。 The present invention is based on the gradation conversion and edge detection, it is possible to quickly find the center position of each of glass panels located within the visual field of the slot, while the center of the field of view in conjunction with the camera position, the glass panel can quickly find the current needs to crawl, to realize a small size sorting panel to quickly locate.

Description

基于单目CCD的小尺寸玻璃面板的分拣快速定位方法及系统 Fast sorting based positioning method and system for small-size glass panel of monocular CCD

技术领域 FIELD

[0001] 本发明属于图像定位领域,尤其涉及一种基于单目CCD的小尺寸玻璃面板的分拣快速定位方法及系统。 [0001] The present invention belongs to the field of image localization, in particular, it relates to sorting based on monocular CCD small size glass panel quick positioning method and system.

背景技术 Background technique

[0002] 目前,定位方法大致可分为机械定位和机器视觉定位两大类,机械定位比较简单, 但是自适应性不高,尤其是对于尺寸大小不一的玻璃面板,而机器视觉方法定位精度高,速度快,自适应高,而且非接触性,能满足实时检测,因而应用越来越广。 [0002] Currently, the positioning method can be divided into mechanical positioning machine vision and positioning two categories, simple mechanical positioning, but not high adaptability, especially for large and small size of the glass panel, and the positioning precision machine vision methods high speed, high adaptive and non-contact, to meet the real-time detection, and thus more and more widely. 根据CCD数量,机器视觉定位方法可分为单目视觉定位和多目视觉定位方法;根据目标物空间维数,又可分为二维定位方法和三维空间定位方法。 The number of the CCD, the positioning can be divided into machine vision monocular vision monocular vision positioning and multi-location method; The spatial dimension of the target, the method can be divided into two-dimensional and three-dimensional positioning locating method. 多目视觉定位方法常用于比较复杂的空间多维定位。 Multi-purpose visual positioning method commonly used in more complex multi-dimensional spatial positioning. 但是目前在使用单目CCD进行分拣快速定位时,机械手的机械定位自适应性不高,容易接触玻璃面板造成面板划伤。 However, currently in use monocular CCD sorting fast positioning, mechanical positioning of the robot is not high adaptability, easy access panel glass panel caused by scratches.

发明内容 SUMMARY

[0003] 本发明所要解决的技术问题在于提供一种基于单目CCD的小尺寸玻璃面板的分拣快速定位方法及系统,旨在使用单目CCD进行分拣快速定位时,机械手的机械定位自适应性不高,容易接触玻璃面板造成面板划伤的问题。 [0003] The present invention solves the technical problem is to provide a positioning method and system for fast sorting based on small-size glass panel monocular CCD aimed monocular CCD sorting fast positioning, since mechanical positioning manipulator adaptability is not high, easily accessible panel glass panel scratches cause problems.

[0004] 本发明是这样实现的,一种基于单目CCD的小尺寸玻璃面板的分拣快速定位方法, 步骤包括: [0004] The present invention is achieved, rapid positioning method of sorting small-size glass panel based on monocular CCD, the steps comprising:

[0005] 步骤A,利用单目CCD采集放置有玻璃面板的卡槽的图像,并将采集到的图像进行灰度转换,然后将灰度转换得到的图像进行预处理得到灰度图像; [0005] Step A, is placed monocular CCD image acquisition card slot of the glass panel, and the collected image gradation conversion, gradation conversion and the image obtained by preprocessing the image to obtain a gray;

[0006] 步骤B,计算所述灰度图像的行像素灰度均值,然后根据所述行灰度均值确定所述玻璃面板所在行坐标; [0006] Procedure B, calculating said gray pixel image row gray value, and then determining the mean gray-scale glass panel according to the row coordinate;

[0007] 步骤C,对所述灰度图像进行二值化分割,根据所述玻璃面板提取感兴趣区域边缘坐标; [0007] Procedure C, the gray image is binarized divided, extracting the region of interest according to the coordinates of the edges of the glass panel;

[0008] 步骤D,根据所述行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标,以所述卡槽中心坐标作为所述玻璃面板的吸附位置,然后控制机械手达到所述玻璃面板的吸附位置进行吸附。 [0008] Procedure D, based on the coordinates and row coordinates of the edge region of interest to determine the coordinates of the center slot, said slot to a center coordinate position of suction of the glass panel, and then controls the robot reaches the glass panel adsorption sites for adsorption.

[0009] 进一步地,步骤A具体包括: [0009] Further, the step A comprises:

[0010] 步骤A1,控制机械手移动至料架卡槽上方,并控制固定于所述机械手上的单目CCD 相机采集放置有玻璃面板的卡槽的图像; [0010] Step A1, the robot control rack slot to move upward, and fixed to said mechanical hand control monocular CCD camera with an image acquisition card slot is placed in the glass panel;

[0011] 步骤A2,对步骤Al采集的图像进行灰度转换,然后进行预处理后得到灰度图像;所述预处理包括滤波、去噪。 [0011] Step A2, step Al captured image gradation conversion, and then, the gradation image preprocessing; treatment comprising filtering, denoising.

[0012] 进一步地,步骤B具体包括: [0012] Further, the step B comprises:

[0013] 步骤BI,计算所述灰度图像的每一行像素的灰度值的总和; [0013] Step BI, the sum of gradation values ​​of the grayscale image is calculated for each row of pixels;

[00M]以I (i,j)表示所述灰度图像第i行第j列,r表示所述灰度图像的高,c表示所述灰度图像的宽,Row (i)表示所述灰度图像第i行像素的灰度值的总和,则: [00M] represented by I (i, j) of the gray image column i-th row j, r represents the high gradation image, c denotes the width of the gray scale image, Row (i) represents the the sum of the gradation values ​​of the i-th row pixel grayscale image, then:

Figure CN105457908BD00061

[0015] 步骤B2,根据每一行像素的灰度值的总和,计算每一行像素灰度均值; [0015] Step B2, the sum of the pixel gray values ​​for each row, each row of pixels is calculated gray value;

[0016] 以RowAve⑴表示第i行像素灰度值均值,则: [0016] In RowAve⑴ row represents the mean gray value of pixel i, then:

[0017] RowAve ⑴=Row (i) /c; [0017] RowAve ⑴ = Row (i) / c;

[0018] 步骤B3,根据每一行像素灰度均值寻找行像素灰度极大值,以该行像素灰度极大值确认所述玻璃面板所在行坐标。 [0018] Step B3, each row of pixels row of pixel gray value to find the maximum value according to the gradation, the maximum gradation to the pixel row to confirm the row coordinate of the glass panel.

[0019] 进一步地,步骤B3具体包括: [0019] Further, B3 comprises the step of:

[0020] 步骤B31,计算所述灰度图像的行像素灰度均值; [0020] Step B31, calculating said gray pixel image row gray value;

[0021] 以RowAverage表示所述灰度图像的行像素灰度均值,则: [0021] In RowAverage represents the row mean gray image pixel gray, then:

Figure CN105457908BD00062

[0023] 步骤B32,计算每一行像素的灰度差值; [0023] Step B32, the gradation difference is calculated for each row of pixels;

[0024] 以Delta⑴表示第i行像素的灰度差值,则该行像素偏离所述行像素灰度均值的大小为:Delta (i) =RowAve (i) -RowAverage; [0024] In Delta⑴ grayscale difference represents the i-th row of pixels, the pixel rows offset from the row of pixel gray Mean size: Delta (i) = RowAve (i) -RowAverage;

[0025] 步骤B33,遍历每一行像素灰度均值,得到最大值,以所述最大值进行阈值设置; [0025] Step B33, through each row of the pixel gray value, maximum value is obtained to set the maximum threshold;

[0026] 以De I ta表示所述阈值,MaxRowAve表示所述最大值,贝丨J: [0026] In De I ta denotes the threshold value, MaxRowAve represents the maximum value, Tony Shu J:

[0027] Delta= (MaxRowAve-RowAverage) *0.8; [0027] Delta = (MaxRowAve-RowAverage) * 0.8;

[0028] 步骤B34,判断所述灰度差值是否满足所述阈值,根据判断结果确定灰度极大值, 从而确定所述玻璃面板所在行坐标; [0028] Step B34, the gradation is determined whether the difference meets the threshold, maximum gradation is determined according to the determination result, so as to determine the row coordinate of the glass panel;

[0029] 若Delta<Delta(i),则确定第i行为灰度极大值所在的行,即所述玻璃面板所在的行,并获得所述玻璃面板所在行坐标。 [0029] If Delta <Delta (i), it is determined that the i-th row where the maximum gradation behavior, i.e. the line where the glass panel, the glass panel is obtained and the row coordinate.

[0030] 进一步地,所述步骤C具体包括: [0030] Further, the step C comprises:

[0031] 步骤Cl,对所述灰度图像进行二值化分割处理,获取灰度二值化图像; [0031] Step Cl, the gray image is binarized division processing, gradation obtain a binarized image;

[0032] 步骤C2,对所述灰度二值化图像进行BLOB分析,得到感兴趣区域图像; [0032] Step C2, the grayscale image binarized BLOB analysis, regions of interest;

[0033] 步骤C3,对所述感兴趣区域图像进行边缘提取,根据提取的边缘得到感兴趣区域边缘坐标; [0033] Step C3, an image of the region of interest for edge extraction, the edge coordinate to obtain the region of interest based on the extracted edge;

[0034] 以CoIGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标, ColGrayVa2 (i)表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝Ij: [0034] In CoIGrayVal⑴ represents the coordinates of endpoints of the left edge of the region of interest of the image i-th row, ColGrayVa2 (i) represents the line region of interest of the image coordinates of the i-th edge point on the right, Tony Ij:

[0035] 对第i行从左开始遍历所述感兴趣区域图像,当满足ColGrayVal (i) =255时跳出循环,记录该点坐标并从右开始遍历;当满足ColGrayVa2⑴=255,记录该点坐标,然后对第i+Ι从左开始遍历所述感兴趣区域图像。 [0035] i-th row of the left image traversing the region of interest, when satisfied ColGrayVal (i) = 255 when the out of the loop, the right to record the coordinates and traversing; satisfied when ColGrayVa2⑴ = 255, recording the coordinates of the point and the first i + Ι traversing the region of interest from the left image.

[0036] 进一步地,步骤D具体包括: [0036] Further, the step D comprises:

[0037] 步骤Dl,根据所述行坐标和所述感兴趣区域边缘坐标,确定该行的卡槽中心; [0037] Step Dl, based on the coordinates and row coordinates edge region of interest, determining the center of the slot line;

[0038] 以RowCenter (i)表示第i行的卡槽中心,以ColCenter⑴表示第i行的行中心,以ColGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2⑴表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝1J: [0038] In RowCenter (i) represents the i-th row of slot center to center row ColCenter⑴ represents i-th row, to the left edge of interest ColGrayVal⑴ represents coordinates of endpoints of the i-th row area of ​​the image, ColGrayVa2⑴ representing the endpoint coordinates of the right edge of the i-th row area of ​​the image of interest, shellfish 1J:

[0039] RowCenter ⑴=i; [0039] RowCenter ⑴ = i;

[0040] ColCenter ⑴=(ColGrayVal (i)+ColGrayVa2 ⑴)/2; [0040] ColCenter ⑴ = (ColGrayVal (i) + ColGrayVa2 ⑴) / 2;

[0041] 步骤D2,根据相机视野中心坐标和所述卡槽中心坐标确定所述玻璃面板的吸附位置; [0041]. Step D2, the suction position of the glass panel is determined according to the camera field of view center coordinates and the coordinates of the center slot;

[0042] 以CameraRowCenter表不相机视野中心的行坐标,CameraColCenter表不相机视野中心的列坐标,r表示所述灰度图像的高,c表示所述灰度图像的宽,则: [0042] In the table does not CameraRowCenter row coordinate of the center of camera views, CameraColCenter coordinates are not listed center of visual field of the camera, r represents a high-gradation image, c denotes the width of the gray scale image, then:

[0043] CameraRowCenter = r/2 ; CameraCo lCenter = c/2; [0043] CameraRowCenter = r / 2; CameraCo lCenter = c / 2;

[0044] 当且仅当满足abs (CameraRowCenter ⑴-RowCenter (i))和 [0044] if and only if the abs (CameraRowCenter ⑴-RowCenter (i)) and

[0045] abs (CameraColCenter (i) -ColCenter⑴)为最小时,确定该行为所要吸附的玻璃面板的所在行,以该行所在行坐标,结合相机的三维坐标,得到所述玻璃面板的吸附位置; [0045] abs (CameraColCenter (i) -ColCenter⑴) is the minimum, it is determined that the row to be adsorbed behavior of the glass panel, to the rows where the row coordinate, in conjunction with three-dimensional coordinates of the camera, to obtain suction position of the glass panel;

[0046] 步骤D3,控制机械手到达所述吸附位置进行吸附。 [0046] Step D3, to control the robot reaches the pickup position adsorption.

[0047] 本发明还提供了一种基于单目C⑶的小尺寸玻璃面板的分拣快速定位系统,包括: [0047] The present invention further provides a method to quickly locate a small-sized sorting based monocular C⑶ glass panel, comprising:

[0048] 采集处理单元,用于利用单目CCD采集放置有玻璃面板的卡槽的图像,并将采集到的图像进行灰度转换,然后将灰度转换得到的图像进行预处理得到灰度图像; [0048] acquisition and processing unit, for monocular CCD image acquisition card slot is placed in the glass panel, and the collected image gradation conversion, gradation conversion and the image obtained by preprocessing the image, the gradation ;

[0049] 计算单元,用于计算所述灰度图像的行像素灰度均值,然后根据所述行灰度均值确定所述玻璃面板所在行坐标; [0049] calculation means for calculating gradation of the pixel row mean gray image, and then determining the mean of the glass panel according to the row coordinate of gray-scale;

[0050] 边缘提取单元,用于对所述灰度图像进行二值化分割,根据所述玻璃面板提取感兴趣区域边缘坐标; [0050] The edge extraction unit, for the gray image is binarized divided, extracting the region of interest according to the coordinates of the edges of the glass panel;

[0051] 定位吸附单元,用于根据所述行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标,以所述卡槽中心坐标作为所述玻璃面板的吸附位置,然后控制机械手达到所述玻璃面板的吸附位置进行吸附。 [0051] positioned adsorption unit, means for determining the center coordinate of slot according to the region of interest and the coordinates of the line edge coordinates to the coordinates as suction slot center position of the glass panel, and then controls the robot reaches the adsorption sites for adsorption of the glass panel.

[0052] 进一步地,所述采集处理单元具体用于: [0052] Furthermore, the acquisition and processing unit is configured to:

[0053] 首先,控制机械手移动至料架卡槽上方,并控制固定于所述机械手上的单目CCD相机采集放置有玻璃面板的卡槽的图像; [0053] First, the robot control rack slot to move upward, and fixed to said mechanical hand control monocular CCD camera with an image acquisition card slot is placed in the glass panel;

[0054] 最后,对采集的图像进行灰度转换,然后进行预处理后得到灰度图像;所述预处理包括滤波、去噪。 [0054] Finally, the image acquisition gradation conversion is performed, and then pre-gradation image obtained; processing comprises filtering, denoising.

[0055] 进一步地,所述计算单元具体用于: [0055] Furthermore, the computing unit is configured to:

[0056] 首先,计算所述灰度图像的每一行像素的灰度值的总和; [0056] First, the sum of gray values ​​is calculated for each row of the gray image pixel;

[0057] 以I (i,j)表示所述灰度图像第i行第j列,r表示所述灰度图像的高,c表示所述灰度图像的宽,Row (i)表示所述灰度图像第i行像素的灰度值的总和,则: [0057] represented by I (i, j) of the gray image column i-th row j, r represents the high gradation image, c denotes the width of the gray scale image, Row (i) represents the the sum of the gradation values ​​of the i-th row pixel grayscale image, then:

Figure CN105457908BD00071

[0058] 其次,根据每一行像素的灰度值的总和,计算每一行像素灰度均值; [0058] Next, the sum of gradation values ​​of each row of pixels, the gray value calculated for each row of pixels;

[0059] 以RowAve⑴表示第i行像素灰度值均值,则: [0059] In RowAve⑴ row represents the mean gray value of pixel i, then:

[0060] RowAve ⑴=Row (i) /c; [0060] RowAve ⑴ = Row (i) / c;

[0061] 最后,根据每一行像素灰度均值寻找行像素灰度极大值,以该行像素灰度极大值确认所述玻璃面板所在行坐标。 [0061] Finally, each row of pixels row of pixel gray value to find the maximum value according to the gradation, the maximum gradation to the pixel row to confirm the row coordinate of the glass panel.

[0062] 进一步地,定位吸附单元具体用于: [0062] Further, the suction unit is positioned to:

[0063] 首先,根据所述行坐标和所述感兴趣区域边缘坐标,确定该行的卡槽中心; [0063] First, based on the coordinates and row coordinates edge region of interest, determining the center of the slot line;

[0064] 以RowCenter (i)表示第i行的卡槽中心,以ColCenter⑴表示第i行的行中心,以ColGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2⑴表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝1J: [0064] In RowCenter (i) represents the i-th row of slot center to center row ColCenter⑴ represents i-th row, to the left edge of interest ColGrayVal⑴ represents coordinates of endpoints of the i-th row area of ​​the image, ColGrayVa2⑴ representing the endpoint coordinates of the right edge of the i-th row area of ​​the image of interest, shellfish 1J:

[0065] RowCenter ⑴=i; [0065] RowCenter ⑴ = i;

[0066] ColCenter ⑴=(ColGrayVal (i) +ColGrayVa2 (i)) /2; [0066] ColCenter ⑴ = (ColGrayVal (i) + ColGrayVa2 (i)) / 2;

[0067] 其次,根据相机视野中心坐标和所述卡槽中心坐标确定所述玻璃面板的吸附位置; [0067] Next, the suction position of the glass panel is determined according to the camera field of view center coordinates and the center coordinates of the slot;

[0068] 以CameraRowCenter表不相机视野中心的行坐标,CameraColCenter表不相机视野中心的列坐标,r表示所述灰度图像的高,c表示所述灰度图像的宽,则: [0068] In the table does not CameraRowCenter row coordinate of the center of camera views, CameraColCenter coordinates are not listed center of visual field of the camera, r represents a high-gradation image, c denotes the width of the gray scale image, then:

[0069] CameraRowCenter = r/2 ; CameraCo lCenter = c/2; [0069] CameraRowCenter = r / 2; CameraCo lCenter = c / 2;

[0070] 当且仅当满足abs (CameraRowCenter ⑴-RowCenter (i))和 [0070] if and only if the abs (CameraRowCenter ⑴-RowCenter (i)) and

[0071] abs (CameraColCenter (i) -ColCenter⑴)为最小时,确定该行为所要吸附的玻璃面板的所在行,以该行所在行坐标,结合相机的三维坐标,得到所述玻璃面板的吸附位置; [0071] abs (CameraColCenter (i) -ColCenter⑴) is the minimum, it is determined that the row to be adsorbed behavior of the glass panel, to the rows where the row coordinate, in conjunction with three-dimensional coordinates of the camera, to obtain suction position of the glass panel;

[0072] 最后,控制机械手到达所述吸附位置进行吸附。 [0072] Finally, the control of the robot reaches the pickup position adsorption.

[0073] 本发明与现有技术相比,有益效果在于:本发明建立在单目视觉定位和二维定位方法的基础上,基于灰度转换和边缘检测,能够迅速找到视野内位于卡槽的每块玻璃面板的中心位置,同时结合相机视野正中心位置,可以快速寻找到当前需要抓取的玻璃面板,实现小尺寸面板的分拣快速定位。 [0073] Compared with the prior art, the beneficial effects that: the present invention is based on monocular vision and positioning the two-dimensional positioning method, based on the gradation conversion and edge detection, the slot is located within the field of view can be found quickly the center position of each of glass panels, the camera field of view combined with n-center position, can quickly find the glass panel needs to capture the current to achieve fast sorting of the small-size panel positioning. 进一步地,本发明利用机器视觉,避免机械定位因接触玻璃面板而造成的二次划伤,同时能够根据卡槽误差进行调整,为玻璃面板检测实现自动化,快速分拣,本发明能自适应多种面板型号。 Further, the present invention is the use of machine vision, avoid secondary mechanical positioning scratches caused by the contact of the glass panel, while the slot can be adjusted according to the error, the glass panel is detected for the automated, rapid sorting, the present invention can be adaptive multi kind panel models.

附图说明 BRIEF DESCRIPTION

[0074] 图1是本发明实施例提供的一种基于单目CCD小尺寸玻璃面板的分拣快速定位方法的流程图。 [0074] FIG. 1 is a flow chart of a method provided a rapid positioning monocular CCD-based small-sized sorting the glass panel of the present invention.

[0075] 图2是本发明实施例提供的进料料架的灰度示意图。 [0075] FIG. 2 is a schematic diagram of the gradation provided in the feed rack of the present invention.

[0076] 图3是本发明实施例提供的进料料架的行灰度值示意图。 [0076] FIG. 3 is the feed frame of gray-scale values ​​provided by the schematic embodiment of the present invention.

[0077] 图4是本发明实施例提供的感兴趣区域边缘示意图。 [0077] FIG. 4 is a schematic view of an edge region of interest according to an embodiment of the present invention.

[0078] 图5是本发明实施例提供的面板吸附位置示意图。 [0078] FIG. 5 is a schematic view of the panel suction position according to an embodiment of the present invention.

[0079] 图6是本发明实施例提供的一种基于单目CCD小尺寸玻璃面板的分拣快速定位系统的结构示意图。 [0079] FIG. 6 is a block diagram of a system of rapid sorting location based on monocular CCD a small-size glass panel according to an embodiment of the present invention.

具体实施方式 Detailed ways

[0080] 为了使本发明的目的、技术方案及优点更加清楚明白,以下结合附图及实施例,对本发明进行进一步详细说明。 [0080] To make the objectives, technical solutions and advantages of the present invention will become more apparent hereinafter in conjunction with the accompanying drawings and embodiments of the present invention will be further described in detail. 应当理解,此处所描述的具体实施例仅仅用以解释本发明,并不用于限定本发明。 It should be understood that the specific embodiments described herein are only intended to illustrate the present invention and are not intended to limit the present invention.

[0081] 基于单目CCD小尺寸玻璃面板分拣快速定位方法是建立在单目视觉定位和二维定位方法的基础上,基于灰度转换和边缘检测,找到视野内位于卡槽的每块玻璃面板的中心位置,并结合相机视野正中心位置寻找当前抓取的玻璃面板。 [0081] Based on the small size of the monocular CCD quick positioning of the glass panel sorting method is based on a two-dimensional visual positioning method and positioning of the monocular, gradation conversion and edge detection based on each piece of glass is located in the slot found in the field of view center of the panel, combined with the center of the camera field of view to find the current location of the captured glass panel. 其整体思路是对机械手上相机抓取的灰度图像进行行投影,计算行灰度均值,根据极大值提取玻璃面板所在行,对原图进行二值化,提取感兴趣区域,对该区域进行边缘提取,进而得到边缘坐标,结合面板所在行坐标,得到面板所在的行和列,结合相机视野和坐标,得到当前分拣面板所在卡槽中心坐标,进行分拣。 Overall idea is the gray level image the camera on the robot gripping line is projected gray value calculation line, extraction of the glass panel according to the row maximum value, the binarization of the original image, extracting a region of interest, the region edge extraction, and further obtain the coordinates of the edge, combined panel row coordinates, where the resulting rows and columns of the panel, in conjunction with the camera field of view and the coordinates, to obtain the current slot sorting center panel location coordinates sorted.

[0082] 基于上述理论,本发明提出了如图1所示的一种基于单目CXD小尺寸玻璃面板的分拣快速定位方法,步骤包括: [0082] Based on the above theory, the present invention provides a sorter shown in Figure 1 based on monocular CXD small size of the glass panel rapid positioning method, comprising the step of:

[0083] S1,利用单目CCD采集放置有玻璃面板的卡槽的图像,并将采集到的图像进行灰度转换,然后将灰度转换得到的图像进行预处理得到灰度图像; [0083] S1, monocular CCD image acquisition card slot is placed in the glass panel, and the collected image gradation conversion, gradation conversion and the image obtained by preprocessing the image to obtain a gray;

[0084] S2,计算所述灰度图像的行像素灰度均值,然后根据所述行灰度均值确定所述玻璃面板所在行坐标。 [0084] S2, calculating said gray pixel image row gray value, and then determining the mean of the glass panel according to the row coordinate of gray-scale. 在本步骤中,将初始的灰度图像的二维图像信息转化为一维信息。 In this step, the two-dimensional image information of the original gray scale image is transformed into one-dimensional information. 同时,本步骤虽然获得了面板所在的行,但是对应的卡槽不一定位于图片正中心,不能通过图像的中心来确定,因此还需要获取卡槽左右两个端点来进一步获取面板行所对应的卡槽正中心,因此还需要进行步骤S3。 Meanwhile, although the present step is obtained where the panel row, but the corresponding images located in the center of the slot is not necessarily, the center of the image can not be determined, it is also necessary to obtain approximately two endpoints slot panel further acquires corresponding row slot center, it is also the need for a step S3.

[0085] S3,对所述灰度图像进行二值化分割,根据所述玻璃面板提取感兴趣区域边缘坐标; [0085] S3, the gray image is binarized divided, extracting the region of interest according to the coordinates of the edges of the glass panel;

[0086] S4,根据所述行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标,以所述卡槽中心坐标作为所述玻璃面板的吸附位置,然后控制机械手达到所述玻璃面板的吸附位置进行吸附。 [0086] S4, is determined based on the coordinates of the center line of the slot and the region of the edge coordinates of the coordinates of interest, to the suction position as the slot center coordinates of the glass panel, and then controls the robot reaches the adsorption of the glass panel position adsorption.

[0087] 具体的,步骤Sl具体包括: [0087] Specifically, the step Sl comprises:

[0088] Sll,控制机械手移动至料架卡槽上方,并控制固定于所述机械手上的单目CCD相机采集放置有玻璃面板的卡槽的图像。 [0088] Sll, controlling the robot to move over the rack slot, and controls the monocular CCD camera is fixed to said mechanical hand image acquisition is placed in the slot of the glass panel. 本步骤中,采集的图像如图2所示。 In this step, the captured images as shown in FIG.

[0089] S12,对步骤Sll采集的图像进行灰度转换,然后进行预处理后得到灰度图像;所述预处理包括滤波、去噪等。 [0089] S12, step Sll captured image gradation conversion, and then pretreated to obtain a gray image; processing comprises filtering, denoising.

[0090] 具体的,步骤S2具体包括: [0090] Specifically, the step S2 comprises:

[0091] S21,计算所述灰度图像的每一行像素的灰度值的总和; [0091] S21, the sum of gray values ​​is calculated for each row of the gray image pixel;

[0092] 以I (i,j)表示所述灰度图像第i行第j列,r表示所述灰度图像的高,c表示所述灰度图像的宽,Row (i)表示所述灰度图像第i行像素的灰度值的总和,则: [0092] represented by I (i, j) of the gray image column i-th row j, r represents the high gradation image, c denotes the width of the gray scale image, Row (i) represents the the sum of the gradation values ​​of the i-th row pixel grayscale image, then:

Figure CN105457908BD00091

[0093] S22,根据每一行像素的灰度值的总和,计算每一行像素灰度均值; [0093] S22, the sum of gradation values ​​of each row of pixels, the gray value calculated for each row of pixels;

[0094] 以RowAve⑴表示第i行像素灰度值均值,则: [0094] In RowAve⑴ row represents the mean gray value of pixel i, then:

[0095] RowAve ⑴=Row (i) /c; [0095] RowAve ⑴ = Row (i) / c;

[0096] S23,根据每一行像素灰度均值寻找行像素灰度极大值,以该行像素灰度极大值确认所述玻璃面板所在行坐标。 [0096] S23, each row of pixels row of pixel gray value to find the maximum value according to the gradation, the maximum gradation to the pixel row to confirm the row coordinate of the glass panel.

[0097] 具体的,上述步骤S23具体包括: [0097] Specifically, the step S23 described above comprises:

[0098] S231,计算所述灰度图像的行像素灰度均值; [0098] S231, calculates the row pixel grayscale image gray value;

[0099] 以RowAverage表示所述灰度图像的行像素灰度均值,则: [0099] In RowAverage represents the row mean gray image pixel gray, then:

Figure CN105457908BD00092

[0101] S232,计算每一行像素的灰度差值; [0101] S232, the gradation difference is calculated for each row of pixels;

[0102] 以Delta⑴表示第i行像素的灰度差值,则该行像素偏离所述行像素灰度均值的大小为:Delta (i) =RowAve (i) -RowAverage; [0102] In Delta⑴ grayscale difference represents the i-th row of pixels, the pixel rows offset from the row of pixel gray Mean size: Delta (i) = RowAve (i) -RowAverage;

[0103] S233,遍历每一行像素灰度均值,得到最大值,以所述最大值进行阈值设置; [0103] S233, through each row of the pixel gray value, maximum value is obtained to set the maximum threshold;

[01 04] 以De I ta表示所述阈值,MaxRowAve表示所述最大值,贝丨J: [0104] In De I ta denotes the threshold value, MaxRowAve represents the maximum value, Tony Shu J:

[0105] Delta= (MaxRowAve-RowAverage) *0 · 8。 [0105] Delta = (MaxRowAve-RowAverage) * 0 · 8. 本步骤中的0 ·8是根据实验数据获得,针对不同卡槽,可相应做改变。 In this step 0.8 are based on experimental data obtained for different slot, it can be changed correspondingly.

[0106] S234,判断所述灰度差值是否满足所述阈值,根据判断结果确定灰度极大值,从而确定所述玻璃面板所在行坐标; [0106] S234, determines whether the gradation difference value meets the threshold, maximum gradation is determined according to the determination result, so as to determine the row coordinate of the glass panel;

[0107] 若Delta<Delta(i),则确定第i行为灰度极大值所在的行,即所述玻璃面板所在的行,并获得所述玻璃面板所在行坐标。 [0107] If Delta <Delta (i), it is determined that the i-th row where the maximum gradation behavior, i.e. the line where the glass panel, the glass panel is obtained and the row coordinate.

[0108] 具体地,上述步骤S3具体包括: [0108] Specifically, in step S3 comprises:

[0109] S31,对所述灰度图像进行二值化分割处理,获取灰度二值化图像; [0109] S31, the gray image is binarized division processing, gradation obtain a binarized image;

[0110] S32,对所述灰度二值化图像进行BLOB分析,得到感兴趣区域图像; [0110] S32, the binarized image gradation BLOB analysis, regions of interest;

[0111] S33,对所述感兴趣区域图像进行边缘提取,根据提取的边缘得到感兴趣区域边缘坐标; [0111] S33, the edge extraction image region of interest, the region of interest to obtain the coordinates of the edge based on the extracted edge;

[0112] 以CoIGrayVal (i)表示所述感兴趣区域图像的第i行的左边边缘端点坐标, ColGrayVa2 (i)表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝Ij: [0112] In CoIGrayVal (i) represents the coordinates of endpoints of the left edge of the i-th row area of ​​the image of interest, ColGrayVa2 (i) represents the line region of interest of the image coordinates of the i-th edge point on the right, Tony Ij:

[0113] 对第i行从左开始遍历所述感兴趣区域图像,当满足ColGrayVal (i) =255时跳出循环,记录该点坐标并从右开始遍历;当满足ColGrayVa2⑴=255,记录该点坐标,然后对第i+Ι从左开始遍历所述感兴趣区域图像。 [0113] i-th row of the left image traversing the region of interest, when satisfied ColGrayVal (i) = 255 when the out of the loop, the right to record the coordinates and traversing; satisfied when ColGrayVa2⑴ = 255, recording the coordinates of the point and the first i + Ι traversing the region of interest from the left image. 在本步骤中,在遍历第i行结束后,开始第i+Ι行, 重复上述遍历获得左右边缘端点坐标,直至遍历完感兴趣区域图像。 In this step, after traversing the i-th row, i + Ι start of the row, repeating the above left and right edge points obtained coordinate traversal, traversing the region of interest until the complete image.

[0114] 具体地,步骤S4进一步包括: [0114] Specifically, step S4 further comprises:

[0115] S41,根据所述行坐标和所述感兴趣区域边缘坐标,确定该行的卡槽中心; [0115] S41, based on the coordinates and row coordinates edge region of interest, determining the center of the slot line;

[0116] 以RowCenter⑴表示第i行的卡槽中心,以ColCenter⑴表示第i行的行中心,以ColGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2⑴表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝1J: [0116] In the center of the slot RowCenter⑴ represents i-th row to the center line ColCenter⑴ represents i-th row, to the left edge of interest ColGrayVal⑴ represents coordinates of endpoints of the i-th row area of ​​the image, ColGrayVa2⑴ representing the region of interest endpoint coordinates of the right edge of the i-th row of the image, shellfish 1J:

[0117] RowCenter ⑴=i; [0117] RowCenter ⑴ = i;

[0118] ColCenter⑴=(ColGrayVal (i)+ColGrayVa2⑴)/2; [0118] ColCenter⑴ = (ColGrayVal (i) + ColGrayVa2⑴) / 2;

[0119] S42,根据相机视野中心坐标和所述卡槽中心坐标确定所述玻璃面板的吸附位置; [0119] S42, the suction position of the glass panel is determined according to the camera field of view center coordinates and the coordinates of the center slot;

[0120] 以CameraRowCenter表不相机视野中心的行坐标,CameraColCenter表不相机视野中心的列坐标,r表示所述灰度图像的高,c表示所述灰度图像的宽,则: [0120] In the table does not CameraRowCenter row coordinate of the center of camera views, CameraColCenter coordinates are not listed center of visual field of the camera, r represents a high-gradation image, c denotes the width of the gray scale image, then:

[0121] CameraRowCenter = r/2 ; CameraCo lCenter = c/2; [0121] CameraRowCenter = r / 2; CameraCo lCenter = c / 2;

[0122] 当且仅当满足abs (CameraRowCenter ⑴-RowCenter (i))和 [0122] if and only if the abs (CameraRowCenter ⑴-RowCenter (i)) and

[0123] abs (CameraColCenter⑴-ColCenter⑴)为最小时,确定该行为所要吸附的玻璃面板的所在行,以该行所在行坐标,结合相机的三维坐标,得到所述玻璃面板的吸附位置; [0123] abs (CameraColCenter⑴-ColCenter⑴) is the minimum, it is determined that the row to be adsorbed behavior of the glass panel, to the rows where the row coordinate, in conjunction with three-dimensional coordinates of the camera, to obtain suction position of the glass panel;

[0124] S43,控制机械手到达所述玻璃面板的吸附位置进行吸附。 [0124] S43, the robot reaches the pickup position control of the glass panel for adsorption.

[0125] 本发明中,提取卡槽两个端点位置,进而得到卡槽中心位置,结合相机所在坐标和视野,得到当前吸附玻璃面板中心坐标,驱动机械手到达该位置吸附,完成定位。 [0125] In the present invention, two end positions extraction slot, the slot and thus obtain the center position, where the coordinates and the camera field of view in conjunction with, the current center coordinates obtained adsorption glass panel, the driving robot reaches the suction position, positioning is completed.

[0126] 本发明还提供了如图6所示的一种基于单目CCD小尺寸玻璃面板的分拣快速定位系统,包括: [0126] The present invention also provides a method as shown in FIG. 6 quickly locate a monocular CCD-based sorting small size glass panel, comprising:

[0127] 采集处理单元1,用于利用单目CCD采集放置有玻璃面板的卡槽的图像,并将采集到的图像进行灰度转换,然后将灰度转换得到的图像进行预处理得到灰度图像; [0127] acquisition and processing unit 1, a monocular CCD image acquisition card slot is placed in the glass panel, and the collected image gradation conversion, gradation conversion and the image obtained by preprocessing, the gradation image;

[0128] 计算单元2,用于计算所述灰度图像的行像素灰度均值,然后根据所述行灰度均值确定所述玻璃面板所在行坐标; [0128] calculation means 2, gradation pixel row for calculating the mean of the gray image, and then determining the mean gray-scale glass panel according to the row coordinate;

[0129] 边缘提取单元3,用于对所述灰度图像进行二值化分割,根据所述玻璃面板提取感兴趣区域边缘坐标; [0129] edge extraction unit 3 for the gray image is binarized divided, extracting the region of interest according to the coordinates of the edges of the glass panel;

[0130] 定位吸附单元4,用于根据所述行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标,以所述卡槽中心坐标作为所述玻璃面板的吸附位置,然后控制机械手达到所述玻璃面板的吸附位置进行吸附。 [0130] positioned adsorption unit 4, for determining the slot according to the center coordinates of the region of interest and the row coordinate edge coordinates to the coordinates as suction slot center position of the glass panel, and then controls the robot reaches a said adsorption sites for adsorption of the glass panel.

[0131] 进一步地,采集处理单元1具体用于: [0131] Further, particularly for acquisition and processing unit 1:

[0132] 首先,控制机械手移动至料架卡槽上方,并控制固定于所述机械手上的单目CCD相机采集放置有玻璃面板的卡槽的图像; [0132] First, the robot control rack slot to move upward, and fixed to said mechanical hand control monocular CCD camera with an image acquisition card slot is placed in the glass panel;

[0133] 最后,对采集的图像进行灰度转换,然后进行预处理后得到灰度图像;所述预处理包括滤波、去噪等。 [0133] Finally, the image acquisition gradation conversion is performed, and then pre-gradation image obtained; processing comprises filtering, denoising.

[0134] 进一步地,计算单元2具体用于: [0134] Furthermore, the computing unit 2 is configured to:

[0135] 首先,计算所述灰度图像的每一行像素的灰度值的总和; [0135] First, the sum of gray values ​​is calculated for each row of the gray image pixel;

[0136] 以I (i,j)表示所述灰度图像第i行第j列,r表示所述灰度图像的高,c表示所述灰度图像的宽,Row (i)表示所述灰度图像第i行像素的灰度值的总和,则: [0136] represented by I (i, j) of the gray image column i-th row j, r represents the high gradation image, c denotes the width of the gray scale image, Row (i) represents the the sum of the gradation values ​​of the i-th row pixel grayscale image, then:

Figure CN105457908BD00111

[0137] 其次,根据每一行像素的灰度值的总和,计算每一行像素灰度均值; [0137] Next, the sum of gradation values ​​of each row of pixels, the gray value calculated for each row of pixels;

[0138] 以RowAve⑴表示第i行像素灰度值均值,则: [0138] In RowAve⑴ row represents the mean gray value of pixel i, then:

[0139] RowAve ⑴=Row (i) /c; [0139] RowAve ⑴ = Row (i) / c;

[0140] 最后,根据每一行像素灰度均值寻找行像素灰度极大值,以该行像素灰度极大值确认所述玻璃面板所在行坐标。 [0140] Finally, each row of pixels row of pixel gray value to find the maximum value according to the gradation, the maximum gradation to the pixel row to confirm the row coordinate of the glass panel. 其中,在本实施过程中还进一部包括, Wherein, in the process of the present embodiment further comprises an inlet,

[0141] 计算所述灰度图像的行像素灰度均值; [0141] Calculation of the row pixel grayscale image gray value;

[0142] 以RowAverage表示所述灰度图像的行像素灰度均值,则: [0142] In RowAverage represents the row mean gray image pixel gray, then:

Figure CN105457908BD00112

[0144] 计算每一行像素的灰度差值; [0144] grayscale difference is calculated for each row of pixels;

[0145] 以Delta⑴表示第i行像素的灰度差值,则该行像素偏离所述行像素灰度均值的大小为:Delta (i) =RowAve (i) -RowAverage; [0145] In Delta⑴ grayscale difference represents the i-th row of pixels, the pixel rows offset from the row of pixel gray Mean size: Delta (i) = RowAve (i) -RowAverage;

[0146] 遍历每一行像素灰度均值,得到最大值,以所述最大值进行阈值设置; [0146] through each row of the pixel gray value, maximum value is obtained to set the maximum threshold;

[0147] 以De I ta表示所述阈值,MaxRowAve表示所述最大值,贝Ij: [0147] In De I ta denotes the threshold value, MaxRowAve represents the maximum value, Tony Ij:

[0148] Delta= (MaxRowAve-RowAverage) *0.8; [0148] Delta = (MaxRowAve-RowAverage) * 0.8;

[0M9]判断所述灰度差值是否满足所述阈值,根据判断结果确定灰度极大值,从而确定玻璃面板所在行坐标; [0M9] determines whether the gradation difference value meets the threshold, maximum gradation is determined according to the determination result, so as to determine coordinates of the line where the glass panel;

[0150] 若Delta<Delta(i),则确定第i行为灰度极大值所在的行,即玻璃面板所在的行, 并获得玻璃面板所在行坐标。 [0150] If Delta <Delta (i), it is determined that the i-th row where the maximum gradation behavior, i.e. the line where the glass panel, the glass panel is obtained and the row coordinate.

[0151] 进一步地,边缘提取单元3具体用于: [0151] Further, the edge extraction unit 3 is configured to:

[0152] 首先,对所述灰度图像进行二值化分割处理,获取灰度二值化图像; [0152] First, the gray image is binarized division processing, gradation obtain a binarized image;

[0153] 然后,对所述灰度二值化图像进行BLOB分析,得到感兴趣区域图像; [0153] Then, the gray scale image is binarized BLOB analysis, regions of interest;

[0154] 最后,对所述感兴趣区域图像进行边缘提取,根据提取的边缘得到感兴趣区域边缘坐标; [0154] Finally, an image of the region of interest for edge extraction, the edge coordinate to obtain the region of interest based on the extracted edge;

[0155] 以CoIGrayVal (i)表示所述感兴趣区域图像的第i行的左边边缘端点坐标, ColGrayVa2 (i)表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝Ij: [0155] In CoIGrayVal (i) represents the coordinates of endpoints of the left edge of the i-th row area of ​​the image of interest, ColGrayVa2 (i) represents the line region of interest of the image coordinates of the i-th edge point on the right, Tony Ij:

[0156] 对第i行从左开始遍历所述感兴趣区域图像,当满足ColGrayVal (i) =255时跳出循环,记录该点坐标并从右开始遍历;当满足ColGrayVa2⑴=255,记录该点坐标,然后对第i+Ι从左开始遍历所述感兴趣区域图像。 [0156] i-th row of the left image traversing the region of interest, when satisfied ColGrayVal (i) = 255 when the out of the loop, the right to record the coordinates and traversing; satisfied when ColGrayVa2⑴ = 255, recording the coordinates of the point and the first i + Ι traversing the region of interest from the left image.

[0157] 进一步地,定位吸附单元具体用于: [0157] Further, the suction unit is positioned to:

[0158] 首先,根据所述行坐标和所述感兴趣区域边缘坐标,确定该行的卡槽中心,并得到其坐标; [0158] First, based on the coordinates and row coordinates edge region of interest, determining the center of the slot line, and get the coordinates;

[0159] 以RowCenter⑴表示第i行的卡槽中心,以ColCenter⑴表示第i行的行中心,以ColGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2⑴表示所述感兴趣区域图像的第i行的右边边缘端点坐标,贝1J: [0159] In the center of the slot RowCenter⑴ represents i-th row to the center line ColCenter⑴ represents i-th row, to the left edge of interest ColGrayVal⑴ represents coordinates of endpoints of the i-th row area of ​​the image, ColGrayVa2⑴ representing the region of interest endpoint coordinates of the right edge of the i-th row of the image, shellfish 1J:

[0160] RowCenter ⑴=i; [0160] RowCenter ⑴ = i;

[0161] ColCenter⑴=(ColGrayVal (i)+ColGrayVa2⑴)/2; [0161] ColCenter⑴ = (ColGrayVal (i) + ColGrayVa2⑴) / 2;

[0162] 其次,根据相机视野中心坐标和所述卡槽中心坐标确定所述玻璃面板的吸附位置; [0162] Next, the suction position of the glass panel is determined according to the camera field of view center coordinates and the center coordinates of the slot;

[0163] 以CameraRowCenter表不相机视野中心的行坐标,CameraColCenter表不相机视野中心的列坐标,r表示所述灰度图像的高,c表示所述灰度图像的宽,则: [0163] In the table does not CameraRowCenter row coordinate of the center of camera views, CameraColCenter coordinates are not listed center of visual field of the camera, r represents a high-gradation image, c denotes the width of the gray scale image, then:

[0164] CameraRowCenter = r/2 ; CameraCo lCenter = c/2; [0164] CameraRowCenter = r / 2; CameraCo lCenter = c / 2;

[0165] 当且仅当满足abs (CameraRowCenter ⑴-RowCenter (i))和 [0165] if and only if the abs (CameraRowCenter ⑴-RowCenter (i)) and

[0166] abs (CameraColCenter⑴-ColCenter⑴)为最小时,确定该行为所要吸附的玻璃面板的所在行,以该行所在行坐标,结合相机的三维坐标,得到所述玻璃面板的吸附位置; [0166] abs (CameraColCenter⑴-ColCenter⑴) is the minimum, it is determined that the row to be adsorbed behavior of the glass panel, to the rows where the row coordinate, in conjunction with three-dimensional coordinates of the camera, to obtain suction position of the glass panel;

[0167] 最后,控制机械手到达所述吸附位置进行吸附。 [0167] Finally, the control of the robot reaches the pickup position adsorption.

[0168] 以上所述仅为本发明的较佳实施例而已,并不用以限制本发明,凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等,均应包含在本发明的保护范围之内。 [0168] The foregoing is only preferred embodiments of the present invention but are not intended to limit the present invention, any modifications within the spirit and principle of the present invention, equivalent substitutions and improvements should be included in the present within the scope of the invention.

Claims (10)

1. 一种基于单目CCD的小尺寸玻璃面板的分拣快速定位方法,其特征在于,所述分拣快速定位方法步骤包括: 步骤A,利用单目CCD采集放置有玻璃面板的卡槽的图像,并将采集到的图像进行灰度转换,然后将灰度转换得到的图像进行预处理得到灰度图像; 步骤B,计算所述灰度图像的行像素灰度均值,然后根据所述行灰度均值确定所述玻璃面板所在行坐标; 步骤C,对所述灰度图像进行二值化分割,根据所述玻璃面板提取感兴趣区域边缘坐标; 步骤D,根据所述行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标,以所述卡槽中心坐标作为所述玻璃面板的吸附位置,然后控制机械手达到所述玻璃面板的吸附位置进行吸附。 A sorting based on monocular CCD quickly locate the small size of the glass panel, wherein said step of sorting rapid positioning method comprising: step A, is placed monocular CCD acquisition card slot glass panel images, and images acquired gradation conversion, gradation conversion and the image obtained by preprocessing the image to obtain a gray; step B, and calculating said gray pixel image row gray value, and then according to the line determining the mean gray row coordinate of the glass panel; step C, and the gray image is binarized divided, extracting the region of interest according to the coordinates of the edges of the glass panel; Procedure D, based on the row coordinate and the determining the coordinates of the edge region of the slot center coordinates of interest, to the suction position as the slot center coordinates of the glass panel, and then controls the robot reaches the suction position of the glass panel for adsorption.
2. 如权利要求1所述的分拣快速定位方法,其特征在于,步骤A具体包括: 步骤A1,控制机械手移动至料架卡槽上方,并控制固定于所述机械手上的单目CCD相机采集放置有玻璃面板的卡槽的图像; 步骤A2,对步骤Al采集的图像进行灰度转换,然后进行预处理后得到灰度图像;所述预处理包括滤波、去噪。 2. The method of sorting the quick positioning of claim 1, wherein step A comprises: step A1, control of the robot to move over the rack slot and fixed to said mechanical hand control monocular CCD camera an image acquisition card slot is placed a glass panel; step A2, step Al captured image gradation conversion, and then pretreated to obtain a gray image; processing comprises filtering, denoising.
3. 如权利要求1所述的分拣快速定位方法,其特征在于,步骤B具体包括: 步骤B1,计算所述灰度图像的每一行像素的灰度值的总和; 以I (i,j)表示所述灰度图像第i行第j列,r表示所述灰度图像的高,c表示所述灰度图像的宽,Row (i)表示所述灰度图像第i行像素的灰度值的总和,则: 3. The method of sorting the quick positioning of claim 1, wherein step B comprises: the sum of the step B1, each row of calculating the gray scale image pixel gray value; to I (i, j ) representing the grayscale image column i-th row j, r represents the high gradation image, c denotes the width of the gray scale image, row (i) denotes the i-th row ash gray image pixels the sum of the value, then:
Figure CN105457908BC00021
其中0彡i彡r,0彡j彡c; 步骤B2,根据每一行像素的灰度值的总和,计算每一行像素灰度均值; 以RowAve⑴表示第i行像素灰度值均值,则: RowAve (i) =Row (i) /c ; 步骤B3,根据每一行像素灰度均值寻找行像素灰度极大值,以该行像素灰度极大值确认所述玻璃面板所在行坐标。 0 wherein R & lt San San i, j San San 0 C; step B2, the sum of gradation values ​​of each row of pixels, the gray value calculated for each row of pixels; RowAve⑴ represents i-th row in the pixel gray value mean is: RowAve (i) = row (i) / c; step B3, each row of pixels row of pixel gray value to find the maximum value according to the gradation, the maximum gradation to the pixel row to confirm the row coordinate of the glass panel.
4. 如权利要求3所述的分拣快速定位方法,其特征在于,步骤B3具体包括: 步骤B31,计算所述灰度图像的行像素灰度均值; 以RowAverage表示所述灰度图像的行像素灰度均值,贝Ij: 4. The method of sorting the quick positioning of claim 3, wherein the step of B3 specifically includes: Step B31, calculating said gray pixel image row gray value; RowAverage to represent the gray-scale image pixel gray mean, Tony Ij:
Figure CN105457908BC00022
步骤B32,计算每一行像素的灰度差值; 以Delta⑴表示第i行像素的灰度差值,则该行像素偏离所述行像素灰度均值的大小为:Delta (i) =RowAve (i) -RowAverage; 步骤B33,遍历每一行像素灰度均值,得到最大值,以所述最大值进行阈值设置; 以De I ta表示所述阈值,MaxRowAve表示所述最大值,贝Ij: Delta= (MaxRowAve-RowAverage) *0.8; 步骤B34,判断所述灰度差值是否满足所述阈值,根据判断结果确定灰度极大值,从而确定所述玻璃面板所在行坐标; 若Delta<Delta (i),则确定第i行为灰度极大值所在的行,即所述玻璃面板所在的行, 并获得所述玻璃面板所在行坐标。 B32, calculated for each row of pixels grayscale difference step; Delta⑴ to grayscale difference represents the i-th row of pixels, the pixel rows offset from the row of pixel gray Mean size: Delta (i) = RowAve (i ) -RowAverage; step B33, through each row of the pixel gray value, maximum value is obtained to set the maximum threshold; to De I ta represents the threshold value, MaxRowAve represents the maximum value, Tony Ij: Delta = ( MaxRowAve-RowAverage) * 0.8; step B34, the gradation is determined whether the difference meets the threshold, maximum gradation is determined according to the determination result, so as to determine the row coordinate of the glass panel; if Delta <Delta (i) , it is determined that the behavior of the i-th row maximum gray value is located, i.e. the line where the glass panel, the glass panel is obtained and the row coordinate.
5. 如权利要求1所述的分拣快速定位方法,其特征在于,所述步骤C具体包括: 步骤C1,对所述灰度图像进行二值化分割处理,获取灰度二值化图像; 步骤C2,对所述灰度二值化图像进行BLOB分析,得到感兴趣区域图像; 步骤C3,对所述感兴趣区域图像进行边缘提取,根据提取的边缘得到感兴趣区域边缘坐标; 以ColGrayVal (i)表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2 (i)表示所述感兴趣区域图像的第i行的右边边缘端点坐标,则: 对第i行从左开始遍历所述感兴趣区域图像,当满足ColGrayVal (i) =255时跳出循环, 记录该点坐标并从右开始遍历;当满足ColGrayVa2 (i) =255,记录该点坐标,然后对第i+1 从左开始遍历所述感兴趣区域图像。 5. The sorting of the quick positioning of claim 1, characterized in that said step C comprises: step C1, the gray image is binarized division processing, gradation obtain a binarized image; step C2, the grayscale image binarized BLOB analysis, regions of interest; step C3, the region of interest for extracting the edge image, to obtain the coordinates of the region of interest based on the extracted edge edge; in ColGrayVal ( i) represents the coordinates of endpoints of the left edge of the i-th row area of ​​the image of interest, ColGrayVa2 (i) represents the coordinates of endpoints of the right edge of the i-th row area of ​​the image of interest, then: for i-th row from left to traverse image the region of interest, when satisfied ColGrayVal (i) = 255 when the out of the loop, the right to record the coordinates and traversing; satisfied when ColGrayVa2 (i) = 255, recording the coordinates of the point, and then from the first i + 1 left start traversing the region of interest image.
6. 如权利要求1所述的分拣快速定位方法,其特征在于,步骤D具体包括: 步骤Dl,根据所述行坐标和所述感兴趣区域边缘坐标,确定该行的卡槽中心,并得到其坐标; 以RowCenter⑴表示第i行的卡槽中心,以ColCenter⑴表示第i行的行中心,以ColGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2⑴表示所述感兴趣区域图像的第i行的右边边缘端点坐标,RowCenter⑴=i表示第i行卡槽中心的行坐标,贝Ij: RowCenter (i) = i ; ColCenter ⑴=(ColGrayVal (i)+ColGrayVa2 ⑴)/2; 步骤D2,根据相机视野中心坐标和所述卡槽中心坐标确定所述玻璃面板的吸附位置; 以CameraRowCenter表示相机视野中心的行坐标,CameraCo I Center表示相机视野中心的列坐标,r表示所述灰度图像的高,c表示所述灰度图像的宽,则: CameraRowCenter = r/2;CameraCoICenter = c/2; 当且仅当满足abs (CameraRowCenter (i 6. The method of sorting the quick positioning of claim 1, wherein the step D comprises: the step of Dl, based on the coordinates and row coordinates edge region of interest, determining the center of the slot line, and to obtain their coordinates; represents RowCenter⑴ slot in the center of the i-th row to the center line ColCenter⑴ represents i-th row, to the left edge of interest ColGrayVal⑴ represents coordinates of endpoints of the i-th row area of ​​the image, ColGrayVa2⑴ expressed interest endpoint coordinates of the right edge of the i-th row area of ​​the image, RowCenter⑴ = i represents the row coordinate of the i-th row center slot, Tony Ij: RowCenter (i) = i; ColCenter ⑴ = (ColGrayVal (i) + ColGrayVa2 ⑴) / 2 ; step D2, the camera field of view is determined according to the center coordinates and the center coordinates of the suction slot position of the glass panel; CameraRowCenter represents a row coordinate in the center of the field of view camera, CameraCo I Center represents the column coordinate of the center of the camera field of view, r represents the high-gradation image, c denotes the width of the gray scale image, then: CameraRowCenter = r / 2; CameraCoICenter = c / 2; if and only if the abs (CameraRowCenter (i ) -RowCenter (i))和abs (CameraColCenter⑴-ColCenter⑴)为最小时,确定该行为所要吸附的玻璃面板的所在行,以该行所在行坐标,结合相机的三维坐标,得到所述玻璃面板的吸附位置; 步骤D3,控制机械手到达所述吸附位置进行吸附。 ) -RowCenter (i)) and abs (CameraColCenter⑴-ColCenter⑴) is the minimum, it is determined that the behavior of the glass panel to be adsorbed row to row coordinates of the line, in conjunction with three-dimensional coordinates of the camera, to obtain adsorption of the glass panel position; step D3, to control the robot reaches the pickup position adsorption.
7. —种基于单目CCD的小尺寸玻璃面板的分拣快速定位系统,其特征在于,所述分拣快速定位系统包括: 采集处理单元,用于利用单目CCD采集放置有玻璃面板的卡槽的图像,并将采集到的图像进行灰度转换,然后将灰度转换得到的图像进行预处理得到灰度图像; 计算单元,用于计算所述灰度图像的行像素灰度均值,然后根据所述行灰度均值确定所述玻璃面板所在行坐标; 边缘提取单元,用于对所述灰度图像进行二值化分割,根据所述玻璃面板提取感兴趣区域边缘坐标; 定位吸附单元,用于根据所述行坐标和所述感兴趣区域边缘坐标确定卡槽中心坐标, 以所述卡槽中心坐标作为所述玻璃面板的吸附位置,然后控制机械手达到所述玻璃面板的吸附位置进行吸附。 7. - speedy and small-sized sorting based positioning system monocular CCD glass panel, characterized in that said sorting rapid positioning system comprising: acquisition and processing means for monocular CCD acquisition card of the glass panel is placed image tank, and the collected image gradation conversion, gradation conversion and the image obtained by preprocessing the image to obtain a gray; calculation unit for calculating the gray value pixel row grayscale image, then determining the mean gray-scale glass panel according to the row coordinate; edge extraction unit, for the gray image is binarized divided, extracting the region of interest according to the coordinates of the edges of the glass panel; positioning adsorption unit, the means for determining the coordinates and row coordinates of the edge region of the slot center coordinates of interest, to the suction position as the slot center coordinates of the glass panel, and then controls the robot reaches the suction position of the glass panel adsorbs .
8. 如权利要求7所述的分拣快速定位系统,其特征在于,所述采集处理单元具体用于: 首先,控制机械手移动至料架卡槽上方,并控制固定于所述机械手上的单目CCD相机采集放置有玻璃面板的卡槽的图像; 最后,对采集的图像进行灰度转换,然后进行预处理后得到灰度图像;所述预处理包括滤波、去噪。 8. The sorting system according to quickly locate claim 7, characterized in that said acquisition and processing unit is configured to: First, the robot control rack slot to move upward, and fixed to said mechanical hand control single CCD camera head is placed an image acquisition card slot of the glass panel; Finally, the image acquisition gradation conversion is performed, and then, the gradation image preprocessing; treatment comprising filtering, denoising.
9. 如权利要求7所述的分拣快速定位系统,其特征在于,所述计算单元具体用于: 首先,计算所述灰度图像的每一行像素的灰度值的总和; 以I (i,j)表示所述灰度图像第i行第j列,r表示所述灰度图像的高,c表示所述灰度图像的宽,Row (i)表示所述灰度图像第i行像素的灰度值的总和,则: In I (i; First, calculate the sum of the gradation value of the gray scale image for each pixel row: 9. Sorting rapid positioning system according to claim 7, wherein said calculating unit is configured to , j) denotes the gray scale image column i-th row j, r represents the high gradation image, c denotes the width of the gray scale image, row (i) represents the i-th row of the pixel gray scale image the sum of the gray values, then:
Figure CN105457908BC00041
其中0彡i彡r,0彡j彡c; 其次,根据每一行像素的灰度值的总和,计算每一行像素灰度均值; 以RowAve⑴表示第i行像素灰度值均值,则: RowAve (i) =Row (i) /c ; 最后,根据每一行像素灰度均值寻找行像素灰度极大值,以该行像素灰度极大值确认所述玻璃面板所在行坐标。 0 wherein R & lt San San i, j San San 0 C; secondly, the sum of gradation values ​​of each row of pixels, the gray value calculated for each row of pixels; RowAve⑴ represents i-th row in the pixel gray value mean it is: RowAve ( i) = row (i) / c; Finally, each row of pixels according to the row of pixel gray value gray Looking maximum value to the maximum value of the pixel gray glass panel for confirmation of the row coordinates.
10. 如权利要求7所述的分拣快速定位系统,其特征在于,定位吸附单元具体用于: 首先,根据所述行坐标和所述感兴趣区域边缘坐标,确定该行的卡槽中心,并得到其坐标; 以RowCenter⑴表示第i行的卡槽中心,以ColCenter⑴表示第i行的行中心,以ColGrayVal⑴表示所述感兴趣区域图像的第i行的左边边缘端点坐标,ColGrayVa2⑴表示所述感兴趣区域图像的第i行的右边边缘端点坐标,RowCenter⑴=i表示第i行卡槽中心的行坐标,贝Ij: RowCenter (i) = i ; ColCenter ⑴=(ColGrayVal (i)+ColGrayVa2 ⑴)/2; 其次,根据相机视野中心坐标和所述卡槽中心坐标确定所述玻璃面板的吸附位置; 以CameraRowCenter表示相机视野中心的行坐标,CameraCo I Center表示相机视野中心的列坐标,r表示所述灰度图像的高,c表示所述灰度图像的宽,则: CameraRowCenter = r/2;CameraCoICenter = c/2; 当且仅当满足abs (CameraRo 10. The sorting according to claim 7 rapid positioning system, characterized in that the positioning unit is configured to adsorption: First, based on the coordinates and row coordinates edge region of interest, determining the center of the slot line, and with their coordinates; represents RowCenter⑴ slot in the center of the i-th row to the center line ColCenter⑴ represents i-th row to the sensing end ColGrayVal⑴ represents the left edge coordinate of the i-th row area of ​​the image of interest, ColGrayVa2⑴ representing the endpoint coordinates of the right edge of the i-th row of the image region of interest, RowCenter⑴ = i represents the row coordinate of the i-th row center slot, Tony Ij: RowCenter (i) = i; ColCenter ⑴ = (ColGrayVal (i) + ColGrayVa2 ⑴) / 2; secondly, the camera field of view is determined according to the center coordinates and the center coordinates of the suction slot position of the glass panel; CameraRowCenter represents a row coordinate in the center of the field of view camera, CameraCo I Center represents the column coordinate of the center of the camera field of view, r represents the high-gradation image, c denotes the width of the gray scale image, then: CameraRowCenter = r / 2; CameraCoICenter = c / 2; if and only if the abs (CameraRo wCenter (i) -RowCenter (i))和abs (CameraColCenter-ColCenter⑴)为最小时,确定该行为所要吸附的玻璃面板的所在行,以该行所在行坐标,结合相机的三维坐标,得到所述玻璃面板的吸附位置; 最后,控制机械手到达所述吸附位置进行吸附。 wCenter (i) -RowCenter (i)) and abs (CameraColCenter-ColCenter⑴) is the minimum, it is determined that the row to be adsorbed act glass panel to the row coordinate of the line, in conjunction with three-dimensional coordinates of the camera, to obtain a glass suction position panel; finally, controlling the robot reaches the pickup position adsorption.
CN201510771640.7A 2015-11-12 2015-11-12 Fast sorting based positioning method and system for small-size glass panel of monocular ccd CN105457908B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510771640.7A CN105457908B (en) 2015-11-12 2015-11-12 Fast sorting based positioning method and system for small-size glass panel of monocular ccd

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510771640.7A CN105457908B (en) 2015-11-12 2015-11-12 Fast sorting based positioning method and system for small-size glass panel of monocular ccd

Publications (2)

Publication Number Publication Date
CN105457908A CN105457908A (en) 2016-04-06
CN105457908B true CN105457908B (en) 2018-04-13

Family

ID=55596316

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510771640.7A CN105457908B (en) 2015-11-12 2015-11-12 Fast sorting based positioning method and system for small-size glass panel of monocular ccd

Country Status (1)

Country Link
CN (1) CN105457908B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106493086B (en) * 2016-10-21 2019-03-05 北京源著智能科技有限公司 The sorting processing method and system of plate
CN106378663A (en) * 2016-11-28 2017-02-08 重庆大学 Machine tool auxiliary tool setting system based on machine vision

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101334263A (en) * 2008-07-22 2008-12-31 东南大学 Circular target circular center positioning method
US8224078B2 (en) * 2000-11-06 2012-07-17 Nant Holdings Ip, Llc Image capture and identification system and process
EP2639745A1 (en) * 2012-03-16 2013-09-18 Thomson Licensing Object identification in images or image sequences
CN104751147A (en) * 2015-04-16 2015-07-01 成都汇智远景科技有限公司 Image recognition method

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4493679B2 (en) * 2007-03-29 2010-06-30 富士フイルム株式会社 Target area extraction method, apparatus, and program
GB2499799A (en) * 2012-02-28 2013-09-04 Snell Ltd Indentifying points of interest in an image using maxima and minima

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8224078B2 (en) * 2000-11-06 2012-07-17 Nant Holdings Ip, Llc Image capture and identification system and process
CN101334263A (en) * 2008-07-22 2008-12-31 东南大学 Circular target circular center positioning method
EP2639745A1 (en) * 2012-03-16 2013-09-18 Thomson Licensing Object identification in images or image sequences
CN104751147A (en) * 2015-04-16 2015-07-01 成都汇智远景科技有限公司 Image recognition method

Also Published As

Publication number Publication date
CN105457908A (en) 2016-04-06

Similar Documents

Publication Publication Date Title
Singh et al. Bigbird: A large-scale 3d database of object instances
US8265425B2 (en) Rectangular table detection using hybrid RGB and depth camera sensors
CN103052968B (en) Object detection apparatus and method for detecting an object
CN101271514B (en) Image detection method and device for fast object detection and objective output
US8269722B2 (en) Gesture recognition system and method thereof
JP5213486B2 (en) Object tracking apparatus and the object tracking method
CN101794515B (en) Target detection system and method based on covariance and binary-tree support vector machine
CN102999939B (en) Coordinate acquiring device, system and method for real-time three-dimensional reconstruction, stereo interaction device
CN101334837A (en) Multi-method integrated license plate image positioning method
CN102006425A (en) Method for splicing video in real time based on multiple cameras
US10234957B2 (en) Information processing device and method, program and recording medium for identifying a gesture of a person from captured image data
CN102722731A (en) Efficient image matching method based on improved scale invariant feature transform (SIFT) algorithm
CN103991449B (en) A vehicle traveling control method and system
CN102184551A (en) Automatic target tracking method and system by combining multi-characteristic matching and particle filtering
CN104156704A (en) Novel license plate identification method and system
CN104517111B (en) Lane line detection method, system, and method for lane departure warning system
JP2008046903A (en) Apparatus and method for detecting number of objects
JP2013065304A (en) High-speed obstacle detection
CN105354568A (en) Convolutional neural network based vehicle logo identification method
US20130193211A1 (en) System and method for robust real-time 1d barcode detection
CN102629326A (en) Lane line detection method based on monocular vision
US9041834B2 (en) Systems and methods for reducing noise in video streams
CN105046196B (en) Concatenated convolutional neural network based on the preceding vehicle vehicle information output method Structured
CN102354363A (en) Identification method of two-dimensional barcode image on high-reflect light cylindrical metal
CN103927016B (en) A real-time three-dimensional hand gesture recognition method and system based on binocular vision

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
GR01