KR100277364B1 - Missing defect automatic detection method and device - Google Patents

Abstract

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a method for automatically detecting a missing defect and an apparatus thereof, and more particularly, to a method for automatically detecting a missing defect and a device for automatically detecting a defective state of a sewing machine by using an image recognition method when a wrapping paper is missing. Missing defective automatic detection method of the present invention comprises the steps of photographing the wrapping paper that is missing; Capturing a captured image to form a still image; Determining a threshold for distinguishing a paper and a thread from each pixel of the still image; Binarizing an image according to a threshold and a value of each pixel; Detecting a value corresponding to the yarn from the binarized image and determining the length of the yarn according to the number of times the value is continuously detected to determine whether the wrapping paper is missing.

Description

Missing defect automatic detection method and device

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a method for automatically detecting a missing defect and a device thereof, and more particularly, to a method for automatically detecting a missing defect and a device for automatically detecting a defective state by using an image recognition method when a wrapping paper is missing.

Generally, an automatic sewing machine is used to make a wrapping paper using the fabric of the wrapping paper. The automatic sewing machine seals and wraps the wrapping paper fabric transferred through the conveyor belt one by one. However, if the sewing needle is defective during sewing, or if the thread is not fixed, or if the tension of the wrapping paper is not constant, the sewing machine cannot be precisely formed and the sewing on the wrapping paper is not formed at regular intervals. Occurs as shown in 1.

Therefore, if the defective needle part is present on the packaging paper, the contents leak out of the packaging paper during transportation or storage, especially when the contents in powder form are filled. Therefore, in order to solve this problem, it was conventionally checked whether the sewing machine was correctly made on the wrapping paper depending on the person. However, this method does not accurately check for missing defects, and relying solely on people when a large amount of wrapping paper is automatically missing can result in very inaccurate results.

In order to solve the above problems, an object of the present invention is to provide a method for automatically detecting a missing defect that can automatically detect a missing defect of a wrapping paper by using an image recognition technique.

Another object of the present invention is to provide a missing automatic detection device which can automatically detect the missing of the wrapping paper by using the above-mentioned automatic missing detection method.

Missing fault automatic detection method according to the present invention to achieve the above object is

Photographing the wrapping paper to be missed;

Capturing the captured image to form a still image;

Determining a threshold value for distinguishing a paper and a thread from each pixel constituting the still image;

Binarizing the image according to the threshold and the value of each pixel;

Detecting a value corresponding to the yarn from the binarized image and determining the length of the yarn in accordance with the number of times the value is detected continuously characterized in that it comprises the step of determining whether the wrapping paper missing.

In addition, the missing automatic detection device according to the present invention

A CCD camera for photographing the missing state of the wrapping paper;

A frame grabber for capturing a still image from an image captured by the camera;

Receiving a still image output from the frame grabber to determine a threshold value for distinguishing a paper and a thread from the pixels constituting the still image to binarize the image according to the threshold value and each pixel value And an image signal processing apparatus for detecting a value corresponding to the yarn from the image, and determining the length of the yarn according to the continuously detected value of the yarn to determine whether the wrapping paper is missing or not.

1 shows an example of a case where a sewing machine is made defective. (Figure 1)

2 is a view schematically showing an example of the use of the missing sewing automatic detection device according to the present invention. (Figure 2)

3A and 3B are histograms showing the captured image and the pixel-by-pixel distribution corresponding to the image. (The picture below ② cuts out only the necessary parts, and ③ corresponds to the histogram creation.)

4 is a view showing an image in which binarization processing has been performed. (④ below the binarization process)

* Description of the symbols for the main parts of the drawings *

10 ... wrapping paper 20 ... transfer section

30 ... CCD camera 40 ... lighting unit

50 ... frame grabber 60 ... video signal processing device

Hereinafter, with reference to the accompanying drawings, a description will be given of a method for automatically detecting a missing defect and an apparatus for automatically detecting a missing defect using the same.

2 is a view schematically showing an example of the use of the missing sewing automatic detection device according to the present invention. 2 to 10 is a CCD camera for photographing the missing part of the wrapping paper to be transported on the wrapping paper, 20 is a conveying part used to transport the wrapping paper 10, 30 is a conveying paper 20 is detected in the present invention, 40 is a lighting unit for adjusting the lighting of the portion photographed by the CCD camera 30, 50 is a frame grabber for capturing the image taken by the CCD camera 30 and 60 is the analysis of the image transmitted from the frame grabber It is an image signal processing device for determining whether or not missing.

Referring to the operation of the missing automatic detection device having a configuration as described above in detail.

The CCD camera 30 photographs a lower end including a missing portion of the wrapping paper 10, for example, a wrapping paper, which is transported by the transporting unit 20 formed of a roller and a conveyor belt at regular time intervals. The CCD camera 30 used a black and white camera, the resolution of which is 768 x 493 pixels, which corresponds to a size of 240 mm x 154 mm (32 pixels / cm).

When the missing part of the wrapping paper is photographed by the CCD camera 30, the photographed image is transmitted to the frame grabber 50. The frame grabber 50 captures an image two frames per second, and divides one piece of wrapping paper into two processes. In the present invention, as a frame grabber, VIP-103 Line of Bargold Electronics of Israel was used. The I / O port consists of 8 bits.

The image signal captured by the frame grabber 50 is input to the image signal processing device 60. The video signal processing apparatus cuts out only the part including the thread among the missing parts of the wrapping paper from the input video signal. That is, the size of the image captured by the camera was initially 768 × 493 pixels, but it was cut out to a size of 240 mm × 16 mm, which is 768 × 50 pixels. This is to reduce the processing of unnecessary parts by cutting only the part to be processed.

As described above, a histogram is generated from the digital image data after the image is cut to a predetermined size. Since data obtained from the frame grabber 50 is composed of 8 bits, each pixel is represented by 2 ⁸ , that is, 256 values. Therefore, each pixel of the image data has a value between 0 and 255, where 0 represents black and 255 represents white.

However, the image data differs in the pixel values of the yarn and the paper part depending on the state at the time of reading, that is, the color of the paper or yarn or the degree of brightness of the illumination. Therefore, the threshold for distinguishing the color of paper and thread should be obtained. To obtain this value, the threshold value can be determined by creating a histogram of all the data and determining the conversion state.

As shown in FIGS. 3A and 3B, the reason for obtaining a threshold value by a histogram is to measure the length of a real node using image data. As shown in Fig. 3A, in the image data, the real part appears bright and the paper part appears dark. Therefore, it can be expected that the value of the pixels corresponding to the yarn is larger than the value of the pixels of the paper portion. However, even the values of the paper part have not only one value but usually have various values between 160 and 190, and the pixels corresponding to the thread have a value of 195 or more. Therefore, according to the distribution of the values between the paper and the thread, the threshold value, which is the point of change between the paper and the thread, must be determined in advance.

Therefore, the threshold value is determined according to the histogram, and according to the result, all pixel values are compared with the threshold value and mapped to 255 when the value is larger than the threshold value, and to 0 when the value is smaller. Change. That is, in the case of FIG. 3B, when the threshold is set to 195, the pixel values are converted to 0 or 255 based on this value. According to this conversion process, when the value of a specific pixel is 255, it may be determined as a threaded part, or conversely, as 0, a paper part.

The method of determining the threshold value as described above may be determined by a trial and error method in addition to the method determined by the histogram. As described above, the method used for the histogram indicates the number of pixels whose horizontal axis is between 0 and 255, and the vertical axis indicates the number of pixels having a given pixel value. As shown in FIG. 3B, the Gaussian distribution forms the paper part, and the right part represents the real part.

In the case of the trial and error method, each time the user increases / decreases the value corresponding to this threshold value and determines the final threshold value by looking at the result of processing the video signal displayed on the screen. In this case, when the threshold value input by the user is low, a lot of white dots will appear in the background part, and when the input threshold value is too high, black dots will appear in the real part. Therefore, it is possible to accurately determine whether there is a missing machine after setting the correct threshold value through trial and error several times.

In addition to the correct setting of the threshold, lighting also plays an important role. In order to perform binarization correctly, that is, in order for all of the paper parts to appear black and only the real part to appear white in the binarized image, the difference between the pixel value of the paper part and the pixel value of the real part in the original image must be large. Therefore, it is also desirable to adjust the illumination so that the difference in these pixel values can be large. In addition, the light shining from the lighting unit should be uniformly shined on the whole wrapping paper.

When the pixels are binarized using only the values of 0 and 255 through the same process as described above, a binarized image of only white and black is obtained as shown in FIG. 4.

When a binary image is obtained, it proceeds from the upper left to the lower right of the image, and also scans from left to right to detect the pixel value and measure the length of the yarn.

First, the length of the yarn proceeds from top to bottom to determine whether there are white spots. That is, it is checked whether the pixel value is 255 while increasing the row number for a given column in a matrix of pixels of a binary image. At this time, if the pixel value is found to be 255 while increasing the row number, the pixel value is checked while sequentially increasing from the first row without moving to the next (right) column without considering further rows. In the above, only the length of the thread is determined, so the thickness of the thread is not considered. Thus, if a pixel with a value of 255 is found in a given column, the next row is skipped without considering the value in the next row.

At this time, the column number of the pixel having the value of 255 that appears first is set as the starting point of the yarn, and then moved to the first row of the next column. It continuously checks to see if there is a pixel with a value of 255 in each column. If the value of 255 is no longer found in a particular column, it indicates the end of the thread, so the value of the current column minus the number of the column of the previously found starting point indicates the thread length (L).

The length of the yarn thus obtained is compared with the reference length. If the comparison value is larger than the predetermined reference value, it is determined as a missing machine and the user is notified of the result. The user can remove or reprocess the wrapping paper determined to be defective. If the comparison value is smaller than the reference value, the sewing work is done properly, so proceed to the next column and repeat the above process.

As described above, by capturing the image of the wrapping paper by the automatic missing detection method and the automatic detecting device according to the present invention, the image is digitized and binarized to measure the length of the yarn from the value of each pixel, and thus, the sewing machine automatically fails the wrapping paper. By detecting this, it is possible to determine whether the defect is more accurate, there is an effect that can process the task accurately and quickly.

While the invention has been particularly shown and described with reference to the above embodiments, it has been used for the purpose of illustration and those of ordinary skill in the art, having the spirit and scope of the invention as defined in the appended claims. Various modifications can be made without departing.

Claims (9)

Photographing the wrapping paper to be missed; Capturing the captured image to form a still image; Determining a threshold value for distinguishing a paper and a thread from each pixel constituting the still image; Binarizing the image according to the threshold and the value of each pixel; And detecting a value corresponding to a yarn from the binarized image, and determining a thread length according to the number of times the value is continuously detected. The method of claim 1, further comprising adjusting a size of the image to include only a missing portion of the image after the forming of the still image. The method of claim 1 or 2, wherein the determining whether the missing a) checking whether the pixel value of the binary image consisting of rows and columns is a value indicating a white point of the pixel while moving the pixel value in a row unit; b) when the pixel indicating the white point is found in the checking step, setting a column of the first discovery position as a start position; c) incrementing the column sequentially and setting the column at the corresponding position as the final position if no pixel indicating a white point is found; d) calculating the pixel length corresponding to the length of the yarn according to the final position and the starting position obtained in the steps c) and b), and comparing the calculated value with a predetermined reference value to determine whether there is a miss. Missing fault automatic detection method, characterized in that. 4. The method of claim 3, wherein when the calculated value in step d) is less than or equal to the reference value, the sewing machine is normally performed. When the calculated value is larger than the reference value, it is determined that the sewing machine is performed poorly. Automatic detection method of missing sewing. The method of claim 4, wherein if the sewing is normally performed in step d), the process proceeds to step a) and repeatedly performs the steps. Missing failure automatic detection method comprising a. CCD camera for taking pictures of the wrapping paper A frame grabber for capturing a still image from an image captured by the camera; Receiving a still image output from the frame grabber to determine a threshold value for distinguishing a paper and a thread from the pixels constituting the still image to binarize the image according to the threshold value and each pixel value And a video signal processing device for detecting a value corresponding to a thread from an image and determining a thread length according to a value continuously detected based on the value. . The apparatus of claim 6, wherein the apparatus for detecting a missing defect further comprises an illumination unit for controlling a difference between pixel values of a paper part and a real part of an image photographed by the CCD camera. The apparatus of claim 6 or 7, wherein the image signal processing apparatus is configured to determine whether a missing image is missing from the binarized image. a) moving the pixel values of the binary image consisting of rows and columns in units of rows and checking whether the white points of the pixels are displayed; b) if a pixel indicating a white point is found in the checking step, the column of the first discovery position is set as a start position; c) increasing the columns sequentially and if no pixels indicating white spots are found, set the columns at the corresponding positions to the final positions, d) calculating the pixel length corresponding to the length of the yarn in accordance with the final position and the starting position obtained in c) and b), and comparing the calculated value with a predetermined reference value to determine whether or not there is a miss. Missing automatic detection device. The image signal processing apparatus of claim 8, wherein the image signal processing apparatus performs the steps a) and repeatedly performs the steps if the sewing machine is normally performed in d), and notifies the user that the sewing state is defective. Missing defect automatic detection method, characterized in that.