CN112284258A - System for measuring cable structure size parameters based on machine vision algorithm - Google Patents

Abstract

The invention provides a system for measuring structural dimension parameters of a cable based on a machine vision algorithm, which adopts an illumination unit to illuminate a cable sample to be measured, a position control unit controls the sample to rotate and jump, an optical imaging unit acquires moving image information of the cable sample to be measured and transmits the image information to a data processing unit, the data processing unit calculates the structural dimension parameters of the sample according to the received image information, the method adopts a sub-pixel edge extraction and fitting algorithm to calculate the diameter parameters of a cable profile, a matching algorithm is used to calculate the thickness parameters of the cable profile, and an area filling method is used to calculate the number of conductor cores of the cable. The system and the method can accurately, quickly and comprehensively measure the structural dimension parameters of the cable, the maximum thickness of a tested cable sample can reach 10cm, the measurement time of the structural dimension parameters of the cable is greatly shortened, the measurement efficiency of the structural dimension parameters of the cable is improved, the measurement procedures are saved, and the labor intensity is reduced.

Description

A system for measuring the size parameters of cable structure based on machine vision algorithm

technical field

The invention relates to the field of cable parameter measurement, and more particularly, to a system for measuring cable structure dimension parameters based on a machine vision algorithm.

Background technique

The measurement of cable structure and size parameters is a mandatory inspection item for cable arrival inspection and random inspection, and it is also one of the most basic test items for discovering cable manufacturing quality defects. Carrying out the measurement of cable structure and size parameters, it is easy to find defects in cable manufacturing quality, which can effectively improve the quality of network cables and ensure long-term safe operation of cables. The measurement parameters of the cable structure size mainly include: the diameter parameter of the cable structure outline, the thickness parameter of the cable structure and the number of metal conductors of the cable. At present, when measuring the parameters of the cable structure and size, the cable is opened layer by layer and sampled for parameter measurement mainly with the help of instruments or manual methods, which is heavily dependent on the experience of the test personnel. Therefore, there is an urgent need for a non-contact, high-precision, high-speed, integrated intelligent measurement system for cable structure and size parameters in all aspects of cable manufacturing, testing, and operation.

With the increasing improvement of image acquisition-related technologies such as high-definition lenses, optical imaging, and motion control, and the increasing improvement of machine vision algorithms such as image denoising, edge detection, image segmentation and recognition, images are used as the carrier of information transmission. The principle and digital image processing technology to measure the structure size of the imaging image of the object plays an important role in the field of modern industrial inspection. Image measurement technology based on machine vision has been practically applied in application fields such as machine manufacturing, communication, national defense and aerospace. Due to the large span of cable structure size and high requirements for the measurement accuracy of structural parameters such as insulation thickness and shielding thickness, the measurement methods disclosed in the existing technical documents have low measurement accuracy, few measurement parameters, slow measurement speed, and high measurement costs. It is suitable for measuring the structural size parameters of cables with many structural layers, large parameter spans and high precision requirements.

SUMMARY OF THE INVENTION

In order to solve the technical problems of measuring the parameters of the cable in the prior art by means of instruments or manual methods, the cable is opened layer by layer and sampled for parameter measurement, which is heavily dependent on the experience of the test personnel, the measurement work is cumbersome, the measurement time is long, the measurement error is large, and the accuracy is difficult to guarantee. , the present invention provides a system for measuring parameters of cable structure size based on machine vision algorithm, the system includes:

The optical imaging unit, one end of which is connected to the data processing unit and the other end is connected to the position control unit, is used for collecting the information of the cable sample to be measured placed in the position control unit when the lighting unit illuminates the cable sample to be measured. image information, and transmit the image information to the data processing unit;

a position control unit, one end of which is connected to the data processing unit, and the other end is connected to the optical imaging unit, which is used for jumping and rotating the cable sample to be measured according to the second instruction of the data processing unit;

The data processing unit is connected with the optical imaging unit, the lighting unit and the position control unit respectively, and is used for outputting a first command to control the lighting unit, and outputting a second command to control the beating and rotating of the cable sample to be measured placed on the position control unit , and calculate the structural dimension parameters of the cable sample to be measured according to the image information transmitted by the optical imaging unit;

An illumination unit, which is connected to the data processing unit, and is used for illuminating the cable sample to be measured placed on the position control unit according to the first instruction of the data processing unit, wherein the illumination unit includes:

a ring light source, which is connected to the light source controller and used for illuminating the cable sample to be measured according to the instructions of the light source controller;

The light source controller, one end of which is connected with the ring light source, and the other end is connected with the data processing unit, is used for controlling the working state of the ring light source according to the first instruction of the data processing unit.

Further, the system also includes:

The result output unit is used for generating a detection report of the structural dimension parameters of the cable sample to be measured according to the calculation result of the data processing unit, the parameters including diameter parameters, thickness parameters and the number of conductor cores.

Further, the optical imaging unit includes:

A telecentric lens, one end of which is connected with the industrial camera, and the other end is connected with the position control unit, for magnifying the image of the cable sample to be measured;

An industrial camera, one end of which is connected to a telecentric lens and the other end is connected to a data processing unit, is used to collect image information of the image of the cable sample to be measured after being enlarged by the telecentric lens, and transmit the image information to the data processing unit.

Further, the position control unit includes:

a motion control unit, one end of which is connected to the data processing unit, and the other end is connected to the object driving unit, for receiving the second instruction of the data processing unit and outputting a signal to the object driving unit to make the object driving unit act;

The object drive unit, one end of which is connected with the motion control unit, and the other end is connected with the position measurement unit, is used for receiving the output signal of the motion control unit to move, and drives the position measurement unit connected to it to jump and rotate;

The position measurement unit, one end of which is connected with the object driving unit and the other end is connected with the optical imaging unit, is used for carrying the cable sample to be measured, and drives the cable sample to be measured to jump and rotate through its own motion.

According to another aspect of the present invention, the present invention provides a method for measuring parameters of a cable structure based on a machine vision algorithm, the method comprising:

When the cable sample to be measured is illuminated by the lighting unit, when the position control unit drives the cable sample to be measured to move, the optical imaging unit enlarges the cable sample to be measured and generates image information, and obtains the image information according to the image information. A cross-sectional small field of view image of the cable sample to be measured, and the cross-sectional small field of view image is transmitted to the data processing unit;

The data processing unit selects any one of the cross-sectional small field of view images as a reference image, extracts the feature points of the reference image and the image to be spliced, and matches the feature points of the reference image and the image to be spliced, so as to realize the small field of view of the cross-section. image splicing to obtain a first cable cross-sectional profile image, wherein the image to be spliced is an image other than the reference image in the cross-sectional small field of view image;

Filtering the first cable cross-sectional profile image while retaining contour edge details, and performing a binarization process on the filtered image to generate a second cable cross-sectional profile image;

performing sub-pixel edge extraction on the second cable cross-sectional profile image to generate a third cable cross-sectional profile image;

The structural dimension parameter of the cable sample to be measured is calculated according to the data of the third cable cross-sectional profile image, and the structural dimension parameter includes a diameter parameter, a thickness parameter and the number of conductor cores.

Further, the data processing unit selects any one of the cross-sectional small field-of-view images as the reference image, extracts the feature points of the reference image and the image to be spliced, and matches the feature points of the reference image and the image to be spliced to realize the matching. The splicing of the cross-sectional small field of view images to obtain the first cable cross-sectional profile image includes:

The Hessian matrix is used to construct a scale space for the cross-sectional small field of view image, and image-stabilized edge points are generated to achieve feature extraction of multiple scales, wherein the Hessian matrix expression of each pixel in the cross-sectional small field of view image is:

表示图像函数，

、

、

为图像函数

关于

、

方向的二阶导数；In the formula,

represents the image function,

,

,

is the image function

about

,

the second derivative of the direction;

The Hessian matrix discriminant is used to determine the local maximum value, and the brightest or darkest point in the local area is obtained, which is determined as the feature point position, wherein the Hessian matrix discriminant is:

为Hessian矩阵的两个特征值，

为Hessian矩阵的行列式，

为Hessian矩阵经行列变化为对角矩阵M的行列式；in,

are the two eigenvalues of the Hessian matrix,

is the determinant of the Hessian matrix,

is the determinant of the Hessian matrix transformed into the diagonal matrix M by rows and columns;

Count the Haar wavelets around the feature points, and select the direction with the largest sum of wavelet features as the main direction;

与基准图像坐标系下对应特征点像素坐标

有如下对应关系：The RANSAC algorithm is used to solve the perspective transformation matrix of the feature points to be paired, and each image to be spliced is transformed into the coordinate system of the reference image through the perspective transformation matrix to achieve image registration, wherein the pixel coordinates of the feature points in the coordinate system of each image to be spliced

The pixel coordinates of the corresponding feature points in the reference image coordinate system

There are the following correspondences:

为所求透视变换矩阵；In the formula,

is the desired perspective transformation matrix;

All the images to be spliced after transformation by the perspective transformation matrix are fused with the reference image to generate a first cable cross-sectional profile image, wherein, for the discontinuous splicing edge that exists during image splicing, the method of weighted fusion is used to analyze the overlapping area of the images. The pixel value of the reference image and the pixel value in the image to be spliced are respectively combined according to a certain weight. The calculation formula is:

为拼接后图像的像素值，

为基准图像的像素值，

为待拼接图像的像素值，

和

为对应的权值，所述权值可根据像素点距离两幅图像边界的距离进行分配。In the formula,

is the pixel value of the stitched image,

is the pixel value of the reference image,

is the pixel value of the image to be stitched,

and

is the corresponding weight value, and the weight value can be assigned according to the distance between the pixel point and the boundary of the two images.

Further, the filtering of the first cable cross-sectional profile image while retaining the contour edge details, and performing binarization processing on the filtered image to generate a second cable cross-sectional profile image includes:

与滤波输出

之间的局部线性模型，其代价函数为The first cable cross-sectional profile image is filtered using a guided filtering method, and the edge details of the first cable cross-sectional profile image are preserved during filtering, wherein the guided image is determined by least squares fitting

with filtered output

The local linear model between , whose cost function is

分别是引导图像和输入图像，

是像素索引，

是正则化参数，

和

分别是引导图像

在窗口

内的均值和方差，

是窗口

内元素总数，

为输入图像

在窗口

内的均值；in the formula

are the guide image and the input image, respectively,

is the pixel index,

is the regularization parameter,

and

bootstrap images

in the window

mean and variance within ,

is the window

the total number of inner elements,

for the input image

in the window

mean within;

After filtering the first cable cross-sectional profile image based on the local linear model, the filtered image is binarized, and only the black and white color information in the image is retained to generate a second cable cross-sectional profile image.

Further, performing sub-pixel edge extraction on the second cable cross-sectional profile image to generate a third cable cross-sectional profile image includes:

The Sobel edge detection method is used to calculate the local gradient modulus value of the second cable cross-sectional profile image point by point, take the maximum value of the modulus values as the gradient value of this point, and record the gradient direction of the template corresponding to the maximum value, As the gradient direction of the point to generate a gradient image with direction information, the formula for calculating the local gradient modulus value is:

为图像局部

区域按照从左到右、从上到下依次递增编号的9个点像素值；in the formula

local to the image

The area is incremented by 9 pixel values of 9 points from left to right and from top to bottom;

、背景灰度级

和圆心到边缘的垂直距离

，并将所述参数

和

的计算结果分别与设定的阈值

和

进行比较以生成比较结果，当所述比较结果满足

时，根据所述参数

和

的值确定每个边缘点的精确亚像素位置。The Zernike moment subpixel edge detection method is used to determine the subpixel position of each edge pixel in the gradient image to generate a third cable cross-sectional profile image, wherein, for each edge pixel, the Zernike moment rotation invariance of the image is used , calculate the grayscale step height

, background grayscale

and the vertical distance from the center to the edge

, and set the parameter

and

The calculation results are respectively related to the set threshold

and

make a comparison to generate a comparison result, when the comparison result satisfies

, according to the parameter

and

The value of determines the exact subpixel location of each edge point.

Further, calculating the structural dimension parameters of the cable sample to be measured according to the data of the third cable cross-sectional profile image includes:

The diameter parameter of the cable sample to be measured is calculated based on the data of the third cable cross-sectional profile image, wherein:

个边缘点计算得到粗糙轮廓拟合圆的圆心

和直径

，其中，

，

，

为自然数，

和

的初始值均为1；Step 1. Based on the third cable cross-sectional profile image

The center of the rough contour fitting circle is obtained by calculating the edge points

and diameter

,in,

,

,

is a natural number,

and

The initial value of is 1;

到所述粗糙轮廓的拟合圆心

的距离

，确定

与的差值

，并生成差值集

；Step 2. Calculate all edge points in the third cable cross-sectional profile image

to the fitted center of the rough contour

the distance

,Sure

difference with

, and generate a difference set

;

中差值元素大于设定的距离差阈值的点，生成差值集

，并根据所述差值集

计算轮廓算数平均偏差

，

，

为自然数；Step 3. Filter out the difference set

The point whose difference element is greater than the set distance difference threshold will generate a difference set

, and according to the difference set

Calculate the arithmetic mean deviation of contours

,

,

is a natural number;

对圆心坐标

、

和直径

的偏导函数确定新的圆心

和直径

；Step 4. According to the average deviation

coordinates of the center of the circle

,

and diameter

The partial derivative of , determines the new center of the circle

and diameter

;

和

计算

，根据圆心坐标

和计算

，根据直径

和

计算

，其计算公式为：Step 5. According to the coordinates of the center of the circle

and

calculate

, according to the coordinates of the center of the circle

and calculation

, according to the diameter

and

calculate

, its calculation formula is:

、

和

均小于设定的误差阈值时，所述直径

为待测量的电缆样品直径参数；当

、

和

中任意一个大于设定的误差阈值，且

时，令

，返回步骤2；当

、

和

中任意一个大于设定的误差阈值，且

时，令

，

，返回步骤1，其中，所述第三电缆截面轮廓图像的

个边缘点包括上一次迭代时选择的第三电缆截面轮廓图像的边缘点；Step 6, when

,

and

are less than the set error threshold, the diameter

is the diameter parameter of the cable sample to be measured; when

,

and

Any one of them is greater than the set error threshold, and

season

, return to step 2; when

,

and

Any one of them is greater than the set error threshold, and

season

,

, return to step 1, where the third cable cross-sectional profile image is

The edge points include the edge points of the third cable cross-section profile image selected in the last iteration;

For the profile of the sample to be measured in the third cable cross-sectional profile image, set the outer layer profile as N and the inner layer profile as M, and calculate the thickness parameter of the to-be-measured cable sample based on search matching, where:

Select a point from N, start from any point, find the matching point with the shortest distance on M within a certain area of the point, traverse a points on N, record the matching point of each point, and save it to the set In P, a is a natural number;

Select a point from M, start from any point, find the matching point with the shortest distance on N in a certain area of the point, traverse a points on M, record the matching point of each point, and save it to the set In Q, a is a natural number;

Compare the matching point pairs in the set P and the set Q, select the two identical point pairs in each point pair as the matching point pair, and save all the successfully matched point pairs into the set;

中；For the successfully matched point pairs in the set, calculate the distance between each matching point pair and save it to the set

middle;

中所有匹配点对间距离的最大值、最小值以及平均值作为厚度的最大值、最小值和平均值；Computational Collection

The maximum value, minimum value and average value of the distances between all matching point pairs in the

Calculate the number of conductor cores of the cable sample to be measured, and the calculation formula is:

为待测量的电缆样品的金属导体线芯概数，

为所述第三电缆截面轮廓图像中的导体区域的像素总面积，

为所述第三电缆截面轮廓图像中单根导体对应的像素面积。In the formula,

is the approximate number of metal conductor cores of the cable sample to be measured,

is the total pixel area of the conductor region in the third cable cross-sectional profile image,

is the pixel area corresponding to a single conductor in the third cable cross-sectional profile image.

The system for measuring the parameters of the cable structure based on the machine vision algorithm provided by the technical solution of the present invention, the system uses an illumination unit to illuminate the cable sample to be measured, the position control unit controls the rotation and beating of the sample, and the optical imaging unit collects the moving The measured image information of the cable sample is transmitted to the data processing unit, and the data processing unit calculates the structural size parameter of the sample according to the received image information. Calculate the thickness parameter of the cable outline, and use the area filling method to calculate the number of conductor cores of the cable. The system and method can accurately, quickly and comprehensively measure the structural size parameters of the cable, and the maximum thickness of the tested cable sample can reach 10cm. Improve the measurement accuracy; through the image stitching of the small field of view of the cable, the resolution of a single pixel of the image can be improved, and the measurement accuracy can be further improved; it can automatically complete the measurement of the cable outline diameter parameters, thickness parameters and the number of conductor cores, compared with the existing technology, The measurement error is small, the measured cable structure size parameters are more comprehensive, the measurement process is automatically completed by the system, avoiding manual participation, and the cable structure size parameter measurement can be completed within 1 minute, the measurement process is faster, and the cable structure size parameter measurement time is greatly reduced , to avoid layer-by-layer sampling and measurement, improve the measurement efficiency of cable structure size parameters, save the measurement process, and reduce labor intensity.

Description of drawings

Exemplary embodiments of the present invention may be more fully understood by reference to the following drawings:

1 is a schematic structural diagram of a system for measuring parameters of cable structure dimensions based on a machine vision algorithm according to a preferred embodiment of the present invention;

FIG. 2 is a flowchart of a method for measuring a cable structure dimension parameter based on a machine vision algorithm according to a preferred embodiment of the present invention;

3 is a cross-sectional small field of view image structure diagram of a cable sample to be measured according to a preferred embodiment of the present invention;

FIG. 4 is a structural diagram of splicing a small field of view image of a cable cross-section to generate a complete cable outline image according to a preferred embodiment of the present invention.

Detailed ways

Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, however, the present invention may be embodied in many different forms and is not limited to the embodiments described herein, which are provided for the purpose of this thorough and complete disclosure invention, and fully convey the scope of the invention to those skilled in the art. The terms used in the exemplary embodiments shown in the drawings are not intended to limit the invention. In the drawings, the same elements/elements are given the same reference numerals.

Unless otherwise defined, terms (including scientific and technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is to be understood that terms defined in commonly used dictionaries should be construed as having meanings consistent with the context in the related art, and should not be construed as idealized or overly formal meanings.

FIG. 1 is a schematic structural diagram of a system for measuring parameters of cable structure dimensions based on a machine vision algorithm according to a preferred embodiment of the present invention. As shown in FIG. 1 , the system 100 for measuring the parameters of the cable structure based on the machine vision algorithm according to the present preferred embodiment includes:

One end of the optical imaging unit 101 is connected to the data processing unit 104, and the other end is connected to the position control unit 103, and is used for collecting the to-be-measured cable samples placed in the position control unit 103 when the lighting unit 102 illuminates the cable sample to be measured. The measured image information of the cable sample is transmitted to the data processing unit 104 .

Preferably, the optical imaging unit 101 includes:

A telecentric lens 111, one end of which is connected to the industrial camera 112, and the other end is connected to the position control unit 103, for magnifying the image of the cable sample to be measured;

The industrial camera 112 has one end connected to the telecentric lens 111 and the other end connected to the data processing unit 104 for collecting the image information of the image of the cable sample to be measured after being enlarged by the telecentric lens, and transmitting it to the data processing unit 104 .

The optical imaging unit uses a telecentric lens to magnify the cable sample to be measured and provide superior image quality, while the industrial camera can collect image information stably and efficiently, improving the accuracy of the image information source of the cable sample to be measured. Calculating the parameters of the cable structure size provides accurate image information.

The lighting unit 102, which is connected with the data processing unit 104, is used for lighting the cable sample to be measured placed in the position control unit 103 according to the first instruction of the data processing unit 104.

Preferably, the lighting unit 102 includes:

The ring light source 121, which is connected with the light source controller 122, is used for illuminating the cable sample to be measured according to the instruction of the light source controller 122;

The light source controller 122 has one end connected to the ring light source 121 and the other end connected to the data processing unit 104 for controlling the working state of the ring light source according to the first instruction of the data processing unit.

The ring light source can provide illumination with high brightness and high directivity, and the working state of the ring light source is controlled by the light source controller to further improve the illumination accuracy of the cable sample to be measured by the ring light source.

One end of the position control unit 103 is connected to the data processing unit 104 , and the other end is connected to the optical imaging unit 101 , which is used for jumping and rotating the cable sample to be measured according to the second instruction of the data processing unit 104 .

Preferably, the position control unit 103 includes:

The motion control unit 131, one end is connected with the data processing unit 104, and the other end is connected with the object driving unit 132, for receiving the second instruction of the data processing unit 104 and outputting a signal to the object driving unit 132 to make the object driving unit 132 act;

The object driving unit 132, one end is connected with the motion control unit 131, and the other end is connected with the position measurement unit 133, is used for receiving the output signal of the motion control unit 132 to move, and drives the position measurement unit 133 connected to it to jump and rotate;

The position measurement unit 133 has one end connected to the object driving unit 132 and the other end connected to the optical imaging unit 101, for carrying the cable sample to be measured, and drives the cable sample to be measured to bounce and rotate through its own motion.

The object drive unit is composed of a motor and a driver. After the driver receives the control signal from the motion control unit, it controls the movement of the motor, thereby driving the position measurement unit connected to it to move. The position measurement unit is a working platform for carrying the cable sample to be measured, and drives the cable sample to be measured to jump and rotate through its own motion. The table surface is marked with a scale size, which can roughly estimate the cable structure size.

The data processing unit 104, which is connected with the optical imaging unit 101, the lighting unit 102 and the position control unit 103 respectively, is used for outputting a first instruction to control the lighting unit 102, and outputting a second instruction to control the object to be measured placed in the position control unit 103. The cable sample jumps and rotates, and the structural dimension parameters of the cable sample to be measured are calculated according to the image information transmitted by the optical imaging unit 101 .

The result output unit 105 is configured to generate a detection report of the structural dimension parameters of the cable sample to be measured according to the calculation result of the data processing unit, the parameters including diameter parameters, thickness parameters and the number of conductor cores.

FIG. 2 is a flow chart of a method for measuring the dimension parameters of a cable structure based on a machine vision algorithm according to a preferred embodiment of the present invention. As shown in FIG. 2 , the method for measuring the parameter of the cable structure size based on the machine vision algorithm according to the present preferred embodiment starts from step 200 .

In step 201, when the cable sample to be measured is illuminated by the lighting unit, when the position control unit drives the cable sample to be measured to move, the optical imaging unit enlarges the cable sample to be measured and generates image information, and according to the The image information is used to obtain a cross-sectional small field of view image of the cable sample to be measured, and the cross-sectional small field of view image is transmitted to the data processing unit.

3 is a cross-sectional small field of view image structure diagram of a cable sample to be measured according to a preferred embodiment of the present invention. As shown in Figure 3, when the cable sample to be measured is moved by the position control unit under the illumination of the lighting unit, the optical imaging unit amplifies the sample and generates a small cross-sectional field of view of several samples image, the structure is not the same in each cross-sectional small field of view image.

In step 202, the data processing unit selects any one of the cross-sectional small field of view images as the reference image, extracts the feature points of the reference image and the image to be spliced, and matches the feature points of the reference image and the image to be spliced, so as to realize the matching of all the images. A first cable cross-sectional profile image is obtained by splicing the cross-sectional small field of view images, wherein the to-be-spliced image is an image other than the reference image in the cross-sectional small field of view image.

Preferably, the data processing unit selects any one of the cross-sectional small field-of-view images as the reference image, extracts the feature points of the reference image and the image to be spliced, and matches the feature points of the reference image and the image to be spliced to realize the matching The splicing of the cross-sectional small field of view images to obtain the first cable cross-sectional profile image includes:

The Hessian matrix is used to construct a scale space for the cross-sectional small field of view image, and image-stabilized edge points are generated to achieve feature extraction of multiple scales, wherein the Hessian matrix expression of each pixel in the cross-sectional small field of view image is:

表示图像函数，

、

、

为图像函数

关于

、

方向的二阶导数；In the formula,

represents the image function,

,

,

is the image function

about

,

the second derivative of the direction;

The Hessian matrix discriminant is used to determine the local maximum value, and the brightest or darkest point in the local area is obtained, which is determined as the feature point position, wherein the Hessian matrix discriminant is:

为Hessian矩阵的两个特征值，

为Hessian矩阵的行列式，

为Hessian矩阵经行列变化为对角矩阵M的行列式；in,

are the two eigenvalues of the Hessian matrix,

is the determinant of the Hessian matrix,

is the determinant of the Hessian matrix transformed into the diagonal matrix M by rows and columns;

Count the Haar wavelets around the feature points, and select the direction with the largest sum of wavelet features as the main direction;

与基准图像坐标系下对应特征点像素坐标

有如下对应关系：The RANSAC algorithm is used to solve the perspective transformation matrix of the feature points to be paired, and each image to be spliced is transformed into the coordinate system of the reference image through the perspective transformation matrix to achieve image registration, wherein the pixel coordinates of the feature points in the coordinate system of each image to be spliced

The pixel coordinates of the corresponding feature points in the reference image coordinate system

There are the following correspondences:

为所求透视变换矩阵；In the formula,

is the desired perspective transformation matrix;

All the images to be spliced after transformation by the perspective transformation matrix are fused with the reference image to generate a first cable cross-sectional profile image, wherein, for the discontinuous splicing edge that exists during image splicing, the method of weighted fusion is used to analyze the overlapping area of the images. The pixel value of the reference image and the pixel value in the image to be spliced are respectively combined according to a certain weight. The calculation formula is:

为拼接后图像的像素值，

为基准图像的像素值，

为待拼接图像的像素值，

和

为对应的权值，所述权值可根据像素点距离两幅图像边界的距离进行分配。In the formula,

is the pixel value of the stitched image,

is the pixel value of the reference image,

is the pixel value of the image to be stitched,

and

is the corresponding weight value, and the weight value can be assigned according to the distance between the pixel point and the boundary of the two images.

FIG. 4 is a structural diagram of splicing a small field of view image of a cable cross-section to generate a complete cable outline image according to a preferred embodiment of the present invention. As shown in FIG. 4 , in the cross-sectional small field of view images of the multiple cable samples to be measured in FIG. 3 , a complete cable is obtained by selecting the reference image and the spliced image, and extracting the feature points of the reference image and the image to be spliced for matching. Cross-sectional profile image.

In step 203, the first cable cross-sectional profile image is filtered while retaining contour edge details, and the filtered image is binarized to generate a second cable cross-sectional profile image.

Preferably, the filtering of the first cable cross-sectional profile image while retaining the contour edge details, and performing binarization processing on the filtered image to generate the second cable cross-sectional profile image includes:

与滤波输出

之间的局部线性模型，其代价函数为The first cable cross-sectional profile image is filtered using a guided filtering method, and the edge details of the first cable cross-sectional profile image are preserved during filtering, wherein the guided image is determined by least squares fitting

with filtered output

The local linear model between , whose cost function is

分别是引导图像和输入图像，

是像素索引，

是正则化参数，

和

分别是引导图像

在窗口

内的均值和方差，

是窗口

内元素总数，

为输入图像

在窗口

内的均值；in the formula

are the guide image and the input image, respectively,

is the pixel index,

is the regularization parameter,

and

bootstrap images

in the window

mean and variance within ,

is the window

the total number of inner elements,

for the input image

in the window

mean within;

After filtering the first cable cross-sectional profile image based on the local linear model, the filtered image is binarized, and only the black and white color information in the image is retained to generate a second cable cross-sectional profile image.

The guided filtering method retains the edge features of the complete cable cross-sectional profile image well, and can retain the edge details of the contour while filtering the spliced image, which can restore the contour image of the cable sample to the greatest extent, and has a good image restoration effect. The binarization of the image is convenient to extract the information in the image, thereby increasing the recognition efficiency during computer recognition.

In step 204, sub-pixel edge extraction is performed on the second cable cross-sectional profile image to generate a third cable cross-sectional profile image.

Preferably, performing sub-pixel edge extraction on the second cable cross-sectional profile image to generate a third cable cross-sectional profile image includes:

The Sobel edge detection method is used to calculate the local gradient modulus value of the second cable cross-sectional profile image point by point, take the maximum value of the modulus values as the gradient value of this point, and record the gradient direction of the template corresponding to the maximum value, As the gradient direction of the point, to generate a gradient image with direction information, where the formula for calculating the local gradient modulus value is:

为图像局部

区域按照从左到右、从上到下依次递增编号的9个点像素值；in the formula

local to the image

The area is incremented by 9 pixel values of 9 points from left to right and from top to bottom;

、背景灰度级

和圆心到边缘的垂直距离

，并将所述参数

和

的计算结果分别与设定的阈值

和

进行比较以生成比较结果，当所述比较结果满足

时，根据所述参数

和

的值确定每个边缘点的精确亚像素位置。The Zernike moment subpixel edge detection method is used to determine the subpixel position of each edge pixel in the gradient image to generate a third cable cross-sectional profile image, wherein, for each edge pixel, the Zernike moment rotation invariance of the image is used , calculate the grayscale step height

, background grayscale

and the vertical distance from the center to the edge

, and set the parameter

and

The calculation results are respectively related to the set threshold

and

make a comparison to generate a comparison result, when the comparison result satisfies

, according to the parameter

and

The value of determines the exact subpixel location of each edge point.

In step 205, a structural dimension parameter of the cable sample to be measured is calculated according to the data of the third cable cross-sectional profile image, where the structural dimension parameter includes a diameter parameter, a thickness parameter and the number of conductor cores.

Preferably, calculating the structural dimension parameters of the cable sample to be measured according to the data of the third cable cross-sectional profile image includes:

The diameter parameter of the cable sample to be measured is calculated based on the data of the third cable cross-sectional profile image, wherein:

个边缘点计算得到粗糙轮廓拟合圆的圆心

和直径

，其中，

，

，

为自然数，

和

的初始值均为1；Step 1. Based on the third cable cross-sectional profile image

The center of the rough contour fitting circle is obtained by calculating the edge points

and diameter

,in,

,

,

is a natural number,

and

The initial value of is 1;

到所述粗糙轮廓的拟合圆心

的距离

，确定

与的差值

，并生成差值集

；Step 2. Calculate all edge points in the third cable cross-sectional profile image

to the fitted center of the rough contour

the distance

,Sure

difference with

, and generate a difference set

;

中差值元素大于设定的距离差阈值的点，生成差值集

，并根据所述差值集

计算轮廓算数平均偏差

，

，

为自然数；Step 3. Filter out the difference set

The point whose difference element is greater than the set distance difference threshold will generate a difference set

, and according to the difference set

Calculate the arithmetic mean deviation of contours

,

,

is a natural number;

对圆心坐标

、

和直径

的偏导函数确定新的圆心

和直径

；Step 4. According to the average deviation

coordinates of the center of the circle

,

and diameter

The partial derivative of , determines the new center of the circle

and diameter

;

和

计算

，根据圆心坐标

和计算

，根据直径

和

计算

，其计算公式为：Step 5. According to the coordinates of the center of the circle

and

calculate

, according to the coordinates of the center of the circle

and calculation

, according to the diameter

and

calculate

, its calculation formula is:

、

和

均小于设定的误差阈值时，所述直径

为待测量的电缆样品直径参数；当

、

和

中任意一个大于设定的误差阈值，且

时，令

，返回步骤2；当

、

和

中任意一个大于设定的误差阈值，且

时，令

，

，返回步骤1，其中，所述第三电缆截面轮廓图像的

个边缘点包括上一次迭代时选择的第三电缆截面轮廓图像的边缘点；Step 6, when

,

and

are less than the set error threshold, the diameter

is the diameter parameter of the cable sample to be measured; when

,

and

Any one of them is greater than the set error threshold, and

season

, return to step 2; when

,

and

Any one of them is greater than the set error threshold, and

season

,

, return to step 1, where the third cable cross-sectional profile image is

The edge points include the edge points of the third cable cross-section profile image selected at the last iteration;

For the profile of the sample to be measured in the third cable cross-sectional profile image, set the outer layer profile as N and the inner layer profile as M, and calculate the thickness parameter of the to-be-measured cable sample based on search matching, where:

Select a point from N, start from any point, find the matching point with the shortest distance on M within a certain area of the point, traverse a points on N, record the matching point of each point, and save it to the set In P, a is a natural number;

Select a point from M, start from any point, find the matching point with the shortest distance on N in a certain area of the point, traverse a points on M, record the matching point of each point, and save it to the set In Q, a is a natural number;

Compare the matching point pairs in the set P and the set Q, select the two identical point pairs in each point pair as the matching point pair, and save all the successfully matched point pairs into the set;

中；For the successfully matched point pairs in the set, calculate the distance between each matching point pair and save it to the set

middle;

中所有匹配点对间距离的最大值、最小值以及平均值作为厚度的最大值、最小值和平均值；Computational Collection

The maximum value, minimum value and average value of the distances between all matching point pairs in the

Calculate the number of conductor cores of the cable sample to be measured, and the calculation formula is:

为待测量的电缆样品的金属导体线芯概数，

为所述第三电缆截面轮廓图像中的导体区域的像素总面积，

为所述第三电缆截面轮廓图像中单根导体对应的像素面积。In the formula,

is the approximate number of metal conductor cores of the cable sample to be measured,

is the total pixel area of the conductor region in the third cable cross-sectional profile image,

is the pixel area corresponding to a single conductor in the third cable cross-sectional profile image.

The present invention has been described with reference to a few embodiments. However, as is known to those skilled in the art, other embodiments than the above disclosed invention are equally within the scope of the invention, as defined by the appended patent claims.

Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise herein. All references to "a/the/the [means, component, etc.]" are open to interpretation as at least one instance of said means, component, etc., unless expressly stated otherwise. The steps of any method disclosed herein do not have to be performed in the exact order disclosed, unless explicitly stated.

As will be appreciated by those skilled in the art, the embodiments of the present application may be provided as a method, a system, or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It will be understood that each flow and/or block in the flowcharts and/or block diagrams, and combinations of flows and/or blocks in the flowcharts and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in one or more of the flowcharts and/or one or more blocks of the block diagrams.

These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory result in an article of manufacture comprising instruction means, the instructions An apparatus implements the functions specified in a flow or flows of the flowcharts and/or a block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the present invention can still be Modifications or equivalent replacements are made to the specific embodiments of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention shall be included within the protection scope of the claims of the present invention.

Claims (4)

1. A system for measuring dimensional parameters of a cable structure based on machine vision algorithms, the system comprising:

the optical imaging unit is connected with the data processing unit at one end and the position control unit at the other end, and is used for collecting image information of the cable sample to be measured placed in the position control unit and transmitting the image information to the data processing unit when the illumination unit illuminates the cable sample to be measured, wherein the thickness of the cable sample to be measured is not more than 10 cm;

the position control unit is connected with the data processing unit at one end and the optical imaging unit at the other end and used for jumping and rotating the cable sample to be measured according to a second instruction of the data processing unit;

the data processing unit is respectively connected with the optical imaging unit, the illuminating unit and the position control unit and is used for outputting a first instruction to control the illuminating unit, outputting a second instruction to control the jumping and rotation of a cable sample to be measured placed in the position control unit and calculating the structural size parameter of the cable sample to be measured according to image information transmitted by the optical imaging unit, wherein the parameter comprises a diameter parameter, a thickness parameter and the number of conductor cores; and

an illumination unit connected with the data processing unit and used for illuminating the cable sample to be measured placed on the position control unit according to a first instruction of the data processing unit, wherein the illumination unit comprises:

the annular light source is connected with the light source controller and used for illuminating a cable sample to be measured according to the instruction of the light source controller;

and one end of the light source controller is connected with the annular light source, and the other end of the light source controller is connected with the data processing unit and used for controlling the working state of the annular light source according to the first instruction of the data processing unit.

2. The system of claim 1, further comprising:

and the result output unit is used for generating a detection report of the structural dimension parameters of the cable sample to be measured according to the calculation result of the data processing unit.

3. The system of claim 2, wherein the optical imaging unit comprises:

one end of the telecentric lens is connected with the industrial camera, and the other end of the telecentric lens is connected with the position control unit and is used for amplifying the image of the cable sample to be measured;

and one end of the industrial camera is connected with the telecentric lens, and the other end of the industrial camera is connected with the data processing unit and is used for acquiring the image information of the cable sample image to be measured, which is amplified by the telecentric lens, and transmitting the image information to the data processing unit.

4. The system of claim 2, wherein the position control unit comprises:

the motion control unit is connected with the data processing unit at one end and connected with the object driving unit at the other end, and is used for receiving a second instruction of the data processing unit and outputting a signal to the object driving unit to enable the object driving unit to act;

the object driving unit is connected with the motion control unit at one end and connected with the position measuring unit at the other end, and is used for receiving the output signal of the motion control unit to move and driving the position measuring unit connected with the object driving unit to jump and rotate;

and one end of the position measuring unit is connected with the object driving unit, and the other end of the position measuring unit is connected with the optical imaging unit and is used for bearing the cable sample to be measured and driving the cable sample to be measured to bounce and rotate through the self-movement.

Images