JPH04236343A - Detecting method of defect of glass edge - Google Patents

Abstract

PURPOSE:To provide a detecting method for detecting defects of a glass edge such as 'chippings', 'points' or the like generated after the glass for automobiles, etc., is cut and polished in a manufacturing line, with applying an image processing technique. CONSTITUTION:The image data in the vicinity of the polishing face 3 of an edge of a glass 2 is input and stored in an image processor 4 by a CCD camera 1. The image data is converted to a binary image (512X512 pixels) 5 expressed only by white and black. Then, pixels in each line are retrieved from above the line (in a direction of an arrow X) and below the line (in a direction of an arrow Y). Coordinates values of a point (pixel) A and a point B where black is changed to white for the first time are obtained. The difference Di= Ci-Ci-1 of a difference Ci-1 of a next line from a difference Ci=Bi-Ai wherein Ai is a coordinate value of the point A and Bi is a coordinate value of the point B is obtained. If Ci=0 or Di>L (judging value), it is determined that the inspecting glass 2 is defective.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention uses image processing technology to detect defects such as chips and horns that occur on glass edges after cutting and polishing in automobile glass production lines. This invention relates to a defect detection method.

0002

2. Description of the Related Art Conventionally, flat glass after cutting and polishing has been visually inspected on an automobile glass production line.

0003

[Problems to be Solved by the Invention] Since the edges of the flat glass after cutting and polishing are visually inspected, this has hindered the complete automation of the glass manufacturing line.

0004

[Means for Solving the Problems] In order to solve the above problems, the present invention inputs the edge portion of the glass to an image processing device using an imaging device, processes it using a predetermined algorithm, and determines whether or not there is a defect in the glass edge. It is something to judge. In addition, the algorithm converts the image data input to the image processing device into a binary image expressed only in white and black, and then searches the pixels in each column from the top of the row to find the coordinates of the pixel that changes from black to white for the first time. Find the value Ai and the coordinate value Bi of the pixel that changes from black to white by searching from below, and also calculate the coordinate value Ai and the coordinate value B.
It is preferable to determine the difference Ci between i and the difference Di between the difference Ci and the difference Ci-1 in the adjacent column, and when Ci=0 or Di is larger than the determination value L, it is determined that the glass has a defect.

[0005]

[Operation] Determine whether there is a defect in the glass edge.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be described below with reference to the accompanying drawings. FIG. 1(A) is an explanatory diagram of the glass edge defect detection method according to the present invention, and FIG. 1(B) is a 512×512
It is a figure which shows the binarized image of the glass edge in a pixel.

A detection device for carrying out the glass edge defect detection method includes a CCD camera 1 as an imaging device and a CC
It consists of an image processing device 4 that stores and processes image data of the edge polished surface 3 of the glass 2 inputted from the D camera 1, and determines whether there are defects such as "chips" or "horns". There is. The CCD camera 1 is attached to an industrial robot (not shown) and programmed in the industrial robot so that it can take images of the vicinity of the edge polished surface 3 of various glasses 2 of different sizes and shapes over the entire circumference of the glass 2. ing.

The algorithm of the method for detecting defects on glass edges will be explained with reference to the flowchart shown in FIG. C.C.
The edge polished surface 3 of the glass 2 is inspected with the D camera 1.
(S1), and image data of a grayscale image output from the CCD camera 1 is input to a memory provided in the image processing device 4 (S2). Next, the image data of the grayscale image stored in the memory is binarized at a certain level,
Convert to binary image 5 expressed only in white and black (S3)
. Then, noise in the binarized image is removed by a noise processing routine. (S4).

Furthermore, as shown in FIG. 1(B), 512×
Each column (5
12 columns) from the upper row and from the lower row (S5), and find the coordinate values of the point (pixel) where black becomes white for the first time (S6). The point (pixel) column found by searching from the top of the row (arrow X direction) is the A point column (512 point columns).
The point (
(pixel) sequence is B point sequence (512 dot sequence) (S
7). If a white point (pixel) is not found even after searching in this way, point A is set to have the same coordinate values as the adjacent point (pixel), and point B is set to have the same coordinate values as point A.

Next, find the difference Ci between the coordinate value Ai of point A and the coordinate value Bi of point B in column i (i=1 to 512) (S8)
. This difference Ci=Bi-Ai is usually the polishing width of the glass 2. Further, the difference Di=Ci-Ci-1 between Ci in each column and Ci-1 in the adjacent column is determined (S9). However, the first difference Di=0. Then, if the difference Ci=0, it is determined that the edge is not polished or the edge portion is chipped (S10), and a defect detection signal is output (S12). Moreover, even if the difference Ci≠0, the difference Di
If is larger than the judgment value L, it is determined that the polishing width is larger than the standard or that the edge portion is chipped and a recess has occurred on the glass surface (S11), and a defect detection signal is output (S12). .

[0011]

As described above, according to the present invention, it is possible to process image data using a relatively simple device and a simple algorithm to determine whether or not there is a defect in the glass edge. Additionally, automation of the inspection process can be achieved.

[Brief explanation of the drawing]

[Fig. 1] (A) is an explanatory diagram of the glass edge defect detection method according to the present invention, and (B) is a diagram showing a binarized image of the glass edge in 512 x 512 pixels.

[Figure 2] Diagram showing a flowchart of the algorithm in the glass edge defect detection method

[Explanation of symbols]

1... CCD camera (imaging device), 2... Glass, 3... Glass edge polished surface, 4... Image processing device, 5... Binarized image,
Ai, Bi...coordinate value, Ci...difference between coordinate value Ai and coordinate value Bi, Di...difference between Ci and Ci-1, L...judgment value.

Claims (2)

[Claims] 1. A method for detecting a defect in a glass edge, comprising inputting an edge portion of the glass to an image processing device using an imaging device, and processing the input using a predetermined algorithm to determine whether or not there is a defect in the glass edge. 2. The algorithm converts the image data input to the image processing device into a binary image expressed only in white and black, and then searches the pixels in each column from the top of the row so that the pixels become white only from black. Find the coordinate value Ai of the pixel and the coordinate value Bi of the pixel that becomes white for the first time from black by searching from below, and calculate the difference Ci between the coordinate value Ai and the coordinate value Bi, and the difference Ci-1 between the difference Ci and the adjacent column. 2. The glass edge defect detection method according to claim 1, wherein the difference Di is determined and when Ci=0 or Di is larger than a determination value L, it is determined that the glass has a defect.