CN-121981983-A - On-line detection method and device for plane glass

CN121981983ACN 121981983 ACN121981983 ACN 121981983ACN-121981983-A

Abstract

The invention discloses a method and a device for online detection of plane glass, wherein the method comprises the steps of sensing glass arrival through a photoelectric sensor, providing accurate position feedback by utilizing a rotary encoder, and carrying out collaborative image acquisition by adopting three linear array cameras, wherein the middle camera directly measures the full width of the glass to eliminate Abbe errors, the encoder hardware triggers to ensure that the image acquisition is strictly synchronous with glass movement, and image distortion is avoided.

Inventors

ZHOU RUIHONG
DOU LITONG
LI SHUJUAN
MA TIANBING
HUANG KAIFENG
SHEN XIAOBO
SHI QINGGUI
ZHANG YUDI
FAN CUNCHAO
DENG QINGQING
Shao Changxiong

Assignees

淮南师范学院

Dates

Publication Date: 20260505
Application Date: 20260119

Claims (8)

1. The on-line detection method for the planar glass is characterized by comprising the following steps of: the method comprises the steps that 1, an image acquisition module is used for acquiring an image of planar glass to be detected, the image acquisition module comprises a first linear array camera, a second linear array camera and a third linear array camera which are sequentially arranged along the glass conveying direction, and the visual field of the second linear array camera completely covers the width direction of the planar glass to be detected; step 2, splicing the images acquired by the first linear array camera, the second linear array camera and the third linear array camera to obtain a complete image; step 3, preprocessing the complete image to obtain coordinate parameters acquired by each linear array camera; step 4, calculating the actual width of the planar glass to be detected by using a formula (1): (1); Where W is the actual width of the strip, And For edge coordinates in the image acquired by the second line camera, Pixel equivalent coefficients for the second present camera; step 5, calculating the actual length of the planar glass to be detected by using a formula (2): (2); wherein L is the actual length of the steel wire, And delta Respectively the pixel spans of the glass in the length direction in the images acquired by the first linear array camera and the third linear array camera, And Is the pixel equivalent coefficient of the corresponding linear array camera.
2. The method for on-line detection of flat glass according to claim 1, wherein the step 1 comprises: Step 11, when the photoelectric sensor detects the front edge of the planar glass to be detected, generating a trigger signal and sending the trigger signal to the controller; step 12, after receiving the trigger signal, the controller immediately latches the count value C 1 of the current rotary encoder; Step 13, calculating a target trigger position according to a preset calibration distance S by using a formula (3): (3); wherein P is the resolution of the encoder, S is the distance from the photoelectric sensor to the first linear array camera, and delta is the triggering frequency division coefficient; step 14, when the count value reaches or exceeds the target trigger position When the acquisition start command is sent to the three cameras; And 15, synchronously acquiring image data of the plane glass to be detected by each camera under the pulse triggering of the encoder.
3. The method for on-line detection of flat glass according to claim 1, wherein the step 2 comprises: Step 21, during image acquisition, each acquisition time triggered by the encoder pulse, the system records the accurate count value of the encoder at the time and takes the value as a common position label of three lines of image data acquired by three cameras at the same time; Step 22, in the splicing stage, three rows of image data acquired under the same encoder count value are arranged according to the corresponding spatial sequence of the first, second and third linear cameras by taking the encoder count value as a sequence reference, and a complete transverse stripe image is formed; step 23, arranging all the transverse stripe images in sequence according to the sequence from the small value to the large value of the encoder, namely reconstructing a complete two-dimensional digital image corresponding to the glass physical area one by one; And step 24, obtaining a pixel-level mapping relation between the images of each camera through the calibration of the cameras in advance, and performing fine adjustment on the image rows during splicing so as to ensure that smooth and accurate butt joint is realized in the overlapping area of the camera fields of view.
4. The method for on-line detection of flat glass according to claim 1, wherein the step 3 comprises: A unified world coordinate system is established through a coordinate transformation formula (4): (4); Wherein, the 、、 Respectively representing coordinate points in a world coordinate system; 、、 respectively representing coordinate points under a camera coordinate system, R representing rotation parameters, subscript numbers representing serial numbers of the linear array camera, R representing a rotation matrix, and t representing a translation vector; perspective projection of the camera coordinate system to the image coordinate system according to equation (5): (5); Wherein, the 、 Respectively representing coordinates in an image coordinate system; 、 representing the focal length of the camera in the x and y directions, respectively; respectively representing the coordinates of the principal point of the camera (i.e. the intersection point of the optical axis and the image plane) in the image pixel coordinate system, wherein the complete camera projection model is shown in formula (6): (6); Wherein, the Is an internal reference matrix of the camera, and S is a scale factor.
5. The method for on-line detection of planar glass according to claim 1, further comprising the steps of 6, performing image analysis on the complete two-dimensional image obtained in the step 2, and determining pixel coordinates of four corner points of the actual physical contour of the glass plate to be detected in an image coordinate system through an edge detection and corner point identification algorithm; Selecting one pair of diagonal points, substituting the diagonal points into a formula (7) to calculate the length of the diagonal line; (7); Wherein, the , Is the coordinates of the upper right corner point, , Is the lower left corner coordinates.
6. An on-line detection device for planar glass, comprising: The image acquisition module is used for acquiring an image of the planar glass to be detected and comprises a first linear array camera, a second linear array camera and a third linear array camera which are sequentially arranged along the glass conveying direction, wherein the view field of the second linear array camera completely covers the width direction of the planar glass to be detected; The image splicing module is used for splicing the images acquired by the first linear array camera, the second linear array camera and the third linear array camera to obtain a complete image; the preprocessing module is used for preprocessing the complete image to obtain coordinate parameters acquired by each linear array camera; The first calculating module is used for calculating the actual width of the planar glass to be detected by using the formula (1): (1); Where W is the actual width of the strip, And For edge coordinates in the image acquired by the second line camera, Pixel equivalent coefficients for the second present camera; the second calculating module is used for calculating the actual length of the planar glass to be detected by using the formula (2): (2); wherein L is the actual length of the steel wire, And delta Respectively the pixel spans of the glass in the length direction in the images acquired by the first linear array camera and the third linear array camera, And Is the pixel equivalent coefficient of the corresponding linear array camera.
7. The online detection device for the planar glass according to claim 6, further comprising a third calculation unit, a first calculation unit and a second calculation unit, wherein the third calculation unit is used for carrying out image analysis on the complete two-dimensional image obtained in the step 2, and determining pixel coordinates of four corner points of the actual physical contour of the glass plate to be detected in an image coordinate system through an edge detection and corner point identification algorithm; Selecting one pair of diagonal points, substituting the diagonal points into a formula (7) to calculate the length of the diagonal line; (7); Wherein, the , Is the coordinates of the upper right corner point, , Is the lower left corner coordinates.
8. A computer readable storage medium, comprising a computer program stored thereon, characterized in that the computer program when executed by a processor realizes the steps of a flat glass online detection method according to any of claims 1 to 5.

Description

On-line detection method and device for plane glass Technical Field The invention relates to the field of image processing, in particular to a method and a device for online detection of planar glass. Background The flat glass is a key basic material in the industries of construction, automobiles, electronics and the like, and the geometric dimensional accuracy (length, width and diagonal line difference) and surface defects (ink points, white lines and unfilled corners) of the flat glass directly influence the product quality. At present, the online detection of the flat glass production line mainly faces the following technical problems: 1. The measurement accuracy is limited, namely, the traditional double-camera system adopts an indirect measurement mode to calculate the glass width, and the measurement accuracy is limited by the calibration accuracy of the distance between two cameras. Any mechanical mounting errors or thermal deformations introduce Abbe (Abbe) errors. 2. The acquisition synchronism is poor, the existing system is usually triggered by software at regular time or by a simple photoelectric switch, and when the speed of a conveyor belt fluctuates or glass slides, the acquired image can be stretched or compressed to deform, so that the measurement accuracy and the defect identification reliability are seriously affected. Disclosure of Invention In order to solve the problems, the invention provides a high-precision flat glass online detection method and device capable of eliminating Abbe errors and realizing hardware-level synchronous triggering. In order to achieve the above object, an aspect of the present invention provides a method for on-line detecting a flat glass, including: the method comprises the steps that 1, an image acquisition module is used for acquiring an image of planar glass to be detected, the image acquisition module comprises a first linear array camera, a second linear array camera and a third linear array camera which are sequentially arranged along the glass conveying direction, and the visual field of the second linear array camera completely covers the width direction of the planar glass to be detected; step 2, splicing the images acquired by the first linear array camera, the second linear array camera and the third linear array camera to obtain a complete image; step 3, preprocessing the complete image to obtain coordinate parameters acquired by each linear array camera; step 4, calculating the actual width of the planar glass to be detected by using a formula (1): (1); Where W is the actual width of the strip, AndFor edge coordinates in the image acquired by the second line camera,Pixel equivalent coefficients for the second present camera; step 5, calculating the actual length of the planar glass to be detected by using a formula (2): (2); wherein L is the actual length of the steel wire, And deltaRespectively the pixel spans of the glass in the length direction in the images acquired by the first linear array camera and the third linear array camera,AndIs the pixel equivalent coefficient of the corresponding linear array camera. Optionally, the step 1 includes: Step 11, when the photoelectric sensor detects the front edge of the planar glass to be detected, generating a trigger signal and sending the trigger signal to the controller; step 12, after receiving the trigger signal, the controller immediately latches the count value C 1 of the current rotary encoder; Step 13, calculating a target trigger position according to a preset calibration distance S by using a formula (3): (3); Wherein P is the resolution of the encoder, S is the distance from the photoelectric sensor to the camera, and delta is the triggering frequency division coefficient; step 14, when the count value reaches or exceeds the target trigger position When the acquisition start command is sent to the three cameras; And 15, synchronously acquiring image data of the plane glass to be detected by each camera under the pulse triggering of the encoder. Optionally, step 2 includes step 21, during image acquisition, each acquisition time triggered by an encoder pulse, the system records an accurate count value of the encoder at the time, and uses the value as a common position label of three lines of image data acquired by three cameras at the same time; Step 22, in the splicing stage, three rows of image data acquired under the same encoder count value are arranged according to the corresponding spatial sequence of the first, second and third linear cameras by taking the encoder count value as a sequence reference, and a complete transverse stripe image is formed; step 23, arranging all the transverse stripe images in sequence according to the sequence from the small value to the large value of the encoder, namely reconstructing a complete two-dimensional digital image corresponding to the glass physical area one by one; And step 24, obtaining a pixel-level mapping relation between the images of