TWI869276B - Processing system and method for detecting adhesive threads - Google Patents

Abstract

A processing system and method for detecting adhesive threads is provided. The processing system comprises a camera device and a detection device. The camera device captures a surface image of the adhesive; the detection device extracts a target block from the surface image; the detection device traverses each pixel in the first axis of the target block; the current pixel is the starting point of the region; the detection device obtains the first depth value of the starting point of the region; the detection device obtains the second depth values of the remaining pixels in the first axis; the detection device calculates the difference between the first depth value and each second depth value to obtain the depth difference set of the starting point of the region; the detection device judges whether the starting point of the region belongs to the adhesive category or non-adhesive category based on the depth difference set; the detection device performs contour marking processing on the non-adhesive category to generate at least one non-adhesive contour region; the detection device generates the detection result based on the non-adhesive contour region.

Description

Processing system and method for detecting adhesive defects

The invention relates to a digital image detection device and a detection method, and in particular to a processing system and a method for detecting adhesive coating defects.

Release liner is mainly used to carry objects with sticky surfaces, such as labels, stickers, tapes or double-sided tape, so that these sticky objects will not stick to each other during transportation and storage. Specifically, an adhesive (release agent) is applied to one side (or both sides) of the release liner, and then the surface material of the sticky object is attached to the release liner coated with adhesive, and then the cutting process and the sticky object edge material are removed. Release paper has appropriate adhesion to sticky objects and also has a good release effect. That is, when the sticky object is torn off, the adhesive layer will be completely attached to the surface material of the sticky object and there will be no residual glue on the release paper. Only when the adhesive is evenly applied can the user tear off the finished adhesive and stick it to the product object to achieve complete adhesion without falling off.

If the adhesive is not evenly applied, the adhesion between the adhesive object and the release paper will be insufficient. In other words, the adhesive object is likely to fall off the release paper or shift during the production process. When the finished adhesive object is torn off the release paper and attached to the product object, it will also fall off due to insufficient adhesion. When personnel discover these problems with adhesive objects during finished product testing or in the later stages of the production process, a large amount of unevenly applied release paper has often been produced.

In addition to the aforementioned method of applying adhesive on release paper, adhesive can also be applied on the back of the surface material of the sticky object so that the sticky object can be attached to the release paper. Due to uneven application, the sticky object will easily fall off or shift from the release paper, and when the finished sticky object is separated from the release paper and attached to a product object, it will cause insufficient adhesion and fall off.

In view of this, in one embodiment, the processing system for detecting adhesive defects includes a camera and a detection device. The camera is used to take an image of the adhesive surface; the detection device is connected to the camera, and the detection device captures the target block from the adhesive surface image. The detection device traverses each pixel in the first axis direction of the target block, and the currently selected pixel is the starting point of the region. The detection device obtains the first depth value of the starting point of the region, and the detection device obtains the second depth value of the remaining pixels in the first axis direction. The detection device uses the first depth value to calculate the depth of the target block. The difference between the first depth value and each second depth value is calculated to obtain multiple differences, which are recorded in a depth difference set. The detection equipment determines whether the starting point of the area belongs to a normal coating area or an abnormal coating area according to the depth difference set. The detection equipment performs contour marking processing on the abnormal coating area to generate at least one coating defect contour area. The detection equipment generates a detection result based on the coating defect contour area. The processing system for detecting coating defects can quickly identify whether a rubber film is coated on the release paper and the thickness of the rubber film. When the release paper on the production line is not coated with a rubber film or the thickness of the rubber film is insufficient, the detection equipment can immediately issue a detection result to notify the staff.

In one embodiment, a method for detecting adhesive coating defects includes the following steps: a detection device captures a target block from an adhesive surface image; the detection device calculates a depth difference set between multiple pixels in a first axis of the target block; the detection device classifies corresponding pixels into a normal coating area or an abnormal coating area according to each depth difference set; the detection device performs contour marking processing on the abnormal coating area to generate at least one adhesive coating defect contour area; the detection device generates a detection result based on the adhesive coating defect contour area.

The processing system and method for detecting glue defects provide a method for using digital images to detect the glue spraying status in real time, so as to prompt whether the film depth of the applied adhesive meets the set standard. Since the detection equipment can immediately confirm the depth of the glue film on the release paper after spraying, the detection equipment can immediately issue relevant warning notifications to prompt production line personnel to eliminate relevant errors.

Please refer to FIG. 1 and FIG. 2, which are respectively a partial schematic diagram of a spray coating release paper and a schematic diagram of a detection device component of an embodiment. FIG. 1 includes a spray coating device 100 and a processing system for detecting adhesive coating defects (referred to as the processing system 200). The processing system 200 can be independent of the spray coating device 100 or built into the spray coating device 100. For the convenience of explanation, the spray coating device 100 and the processing system 200 are described separately. The spray coating device 100 outputs the release paper 912 on which the adhesive has been sprayed. For the convenience of explanation, the release paper 912 is referred to as the coated release paper 912. Under normal conditions, when the adhesive is sprayed on the release paper 912, multiple groups of adhesive strips (or adhesive threads) will be formed, and the collection of the adhesive strips (or adhesive threads) will form an adhesive surface. When the spraying device 100 is blocked or other conditions occur, it will cause uneven distribution on the adhesive surface of the release paper 912, so that there will be a state of partial adhesive shortage or uneven thickness in the area, such as the formation of continuous or discontinuous stripes, filaments or spots of adhesive shortage.

The processing system 200 includes a camera device 210 , a storage device 220 and a detection device 230 . The detection device 230 may be, but is not limited to, a personal computer, a laptop, an embedded system or a server. The detection device 230 is connected to the camera device 210 and the storage device 220 . The camera device 210 may be connected to the detection device 230 through a cable, or may be connected to the detection device 230 through wireless communication. The camera device 210 is provided on the output side of the coating device. The camera device 210 captures the rubber surface image 410 output by the coating device, where the rubber surface image 410 has the sprayed release paper 912 . The camera device 210 can shoot the release paper 912 in real time or in sections. The above-mentioned sectioned shooting can be shooting the release paper 912 at fixed intervals or fixed lengths.

The storage device 220 stores the image detection program 221, the adhesive surface image 410 and the detection result 222. The detection device 230 executes the image detection program 221 and generates the corresponding detection result 222. The image detection program 221 executes the following steps, and please refer to FIG3: Step S310: The detection device captures the target block from the adhesive surface image; Step S320: The detection device calculates the depth difference set between multiple pixels in the first axis of the target block; Step S330: The detection device classifies the corresponding pixels into normal coating areas or abnormal coating areas according to each depth difference set; Step S340: The detection device performs contour marking processing on the abnormal coating area to generate at least one adhesive coating defect contour area; and Step S350: The detection device generates a detection result according to the adhesive coating defect contour area.

After the spraying device 100 applies adhesive on the paper surface, the camera device 210 shoots and outputs a laminated image 410, as shown in FIG4 . The detection device 230 extracts a target block 420 from the laminated image 410. The detection device 230 can extract a target block 420 of a corresponding size from the laminated image 410 using a frame of a preset size. The dashed frame in FIG4 means the target block 420 in the laminated image 410.

Generally speaking, the target block 420 has a first axis 431 and a second axis 432. The first axis 431 corresponds to the horizontal axis in FIG. 4 , and the second axis 432 corresponds to the vertical axis in FIG. 4 . The detection device 230 will traverse each pixel on the first axis 431. Each pixel has two attributes, namely, "pixel position" and "pixel value". In the following description, in order to distinguish the two attributes, "pixel" refers to "pixel position" and "pixel value" refers to the "value" of the pixel.

The currently selected pixel is referred to as the region starting point 711. Then, the detection device 230 obtains the pixel value of the region starting point 711 as the first depth value, and the pixel values of the remaining pixels as the second depth value. The detection device 230 performs difference calculations with the first depth value and all the second depth values to obtain the difference between the region starting point 711 and other pixels. The detection device 230 records these differences in a depth difference set. The detection device 230 determines whether the region starting point 711 belongs to the normal coating area or the abnormal coating area 510 based on the depth difference set. The normal coating area is a set of pixels having a glue surface formed by an adhesive in the release paper 912, which means that the adhesive is evenly coated in the area. The abnormal coating area 510 is a pixel set in which at least one thread-like portion (or strip) without adhesive or with insufficient film thickness exists in the release paper 912. In other words, the adhesive is not completely coated in the abnormal coating area 510.

In some embodiments, the detection device 230 determines whether each difference in the depth difference set is greater than a depth threshold. If there is any difference greater than the depth threshold in the depth difference set, the detection device 230 classifies the region start point 711 as the abnormal coating region 510, otherwise, the detection device 230 classifies the region start point 711 as the normal coating region. The depth threshold is determined by whether the spraying device 100 is configured with an external light source 911. Generally speaking, if the spraying device 100 is configured with an external light source 911, the detection device 230 can adjust the depth threshold to match the ambient brightness during on-site shooting.

After the detection device 230 traverses all pixels on the first axis 431, the detection device 230 can classify all pixels on the first axis as a normal coating area or an abnormal coating area 510. Then, the detection device 230 takes the first area starting point 711 as the main point and selects another pixel along the second axis 432, and performs steps S320 and S330 on the newly selected pixel and the first axis 431 to which it belongs, until the detection device 230 completes all pixels on the second axis 432 and the first axis 431. After the detection device 230 completes the processing of each pixel in the second axis direction 432, the detection device 230 completes the classification of the normal coating area and the abnormal coating area 510 in the target block 420, as shown in FIG. 5 .

FIG5 is a display result of the abnormal coating area 510 in the target block 420 corresponding to FIG4 and highlighted. While the detection device 230 classifies the abnormal coating area 510 and the normal coating area, the detection device 230 sets the object number for the abnormal coating area 510. The detection device 230 is capable of distinguishing the abnormal coating area 510 from the background in order to distinguish the normal coating area. The detection device 230 assigns specific values to the object number of the normal coating area and the object number of the abnormal coating area 510.

For example, the object number of the normal coating area can be set to "-1", and the object number of the abnormal coating area 510 can be initially set to "0", please refer to Figure 6. Figure 6 is a schematic diagram of the contour number table 610 of the local target block 420. The set of object numbers corresponding to the target block 420 is referred to as the contour number table 610. In other words, the size of the contour number table 610 is equal to the image size of the target block 420, and each pixel position of the target block 420 corresponds to each object number in the contour number table 610.

The detection device 230 selects one pixel from the contour number table 610, and the selected pixel is the starting pixel 611. Generally speaking, the detection device 230 can start selecting from the side boundary of the target block 420. Or, the detection device 230 starts selecting from the center axis of the target block 420. The detection device 230 performs contour marking processing on the starting pixel 611 to mark the object number of the starting pixel 611 and its surrounding pixels. Please refer to Figure 7 for the specific steps and operation of the contour marking processing. The initial value of the object number of the contour number table 610 includes the following steps: Step S710: The detection device sets the contour detection frame with the starting pixel as the center, and obtains the pixels adjacent to the starting pixel as the peripheral pixels; Step S720: The detection device adjusts the object number of the starting pixel according to the object numbers of each peripheral pixel; Step S730: If the object numbers of each peripheral pixel are "-1", the detection device sets the object number of the starting pixel; Step S740: If the object number of any peripheral pixel is not "-1" and the aforementioned object numbers are the same, the detection device sets the object number of the starting pixel with the object number of the peripheral pixel; Step S750: If the object number of any peripheral pixel is not "-1" and the aforementioned object numbers are all different, the detection device selects the object number with the minimum value from the aforementioned peripheral pixels, and updates the object number of the starting pixel with the selected object number; and Step S760: The detection device traverses all pixels in the target block and obtains the adjusted contour number table.

Initially, the object number of each pixel in the normal coating area of the target block 420 is set to "-1", and the object number of the abnormal coating area 510 is initially set to "1". The detection device 230 selects the starting pixel 611 in the target block 420 from left to right and from top to bottom one by one. The detection device 230 can select some peripheral pixels for comparison to reduce the overall computing load.

In Figure 8, the detection device 230 encounters a pixel with an object number of "1". The detection device 230 will use this pixel as the starting pixel 611 and select the corresponding contour detection frame 620. The detection device 230 sets the contour detection frame 620 with the starting pixel 611 as the center. Figure 8 is a continuation of the contour number table 610 of Figure 6. The first contour detection frame 621 and the second contour detection frame 622 are drawn in Figure 8. The thick black line frame in Figure 8 and the gray area in the column are represented. The contour detection frame 620 is set to an array of 3*3 pixels. In the contour detection frame 620, the starting pixel 611 is used as the center, and the pixels around the starting pixel 611 are called peripheral pixels (corresponding to step S710).

The detection device 230 compares the object number of the starting pixel 611 with the object number of its peripheral pixels, and updates the object number of the starting pixel 611 according to the comparison result (corresponding to step S720). The detection device 230 chooses to use 8 adjacent peripheral pixels for comparison, and can also select part of the peripheral pixels for comparison. For example, the detection device 230 selects the pixels at the "upper left", "directly above", "upper right", and "left" of the starting pixel 611 as peripheral pixels. According to the aforementioned selection method of the starting pixel 611, the contour detection frame 620 moves from the left side of the contour number table 610 to the right and from top to bottom. Therefore, two different positions are selected in the contour number table 610 as an example for explanation.

In the first contour detection frame 621, the object numbers of the peripheral pixels of the starting pixel 611 are "-1", "-1", "-1", and "-1", respectively. Since the object numbers of these peripheral pixels are all "-1", it means that these peripheral pixels all belong to the normal coating area, so the detection device 230 does not adjust the object number of the starting pixel 611 of the first contour detection frame 621. Then, the detection device 230 will determine whether the object numbers of these peripheral pixels are all "1". If the object numbers of the peripheral pixels are all "1", the detection device 230 will mark the starting pixel 611 with a new object number (corresponding to step S730).

The object numbers of the peripheral pixels of the second contour detection frame 622 are "1", "-1", "-1", and "-1", respectively. Therefore, the detection device 230 accumulates the object number of the starting pixel 611 from "1" to "2" (corresponding to step S740). At the same time, since the object number of the left pixel of the starting pixel 611 of the second contour detection frame 622 is "1". The aforementioned two adjacent pixels are connected, the object number of the left pixel will also be adjusted to "2" (corresponding to step S750). After the detection device 230 repeats the aforementioned steps, a contour number table 610 of the abnormal coating area 510 with multiple sets of different object numbers can be obtained as shown in Figure 9.

Next, the detection device 230 processes the new contour number table 610 in steps S710-S760 to merge the object numbers of the contour number table 610 of Fig. 9. The result after merging is shown in Fig. 10, which is a schematic diagram of the contour number table 610 after the second setting of the object number.

The detection device 230 draws a corresponding area according to the contour number table 610 of the target block 420, and the drawn area is a coating defect contour area 631, please refer to FIG11. The coating defect contour area 631 is a collection of multiple adjacent abnormal coating areas 510, or a collection of multiple similar abnormal coating areas 510. In this embodiment, the vertical direction in FIG11 is used as a judgment basis for the collection of adjacent (or similar) abnormal coating areas 510. In other embodiments, the horizontal direction can also be used as a judgment basis. In some embodiments, the detection device 230 determines whether the area of each adhesive coating defect contour region 631 is greater than the area threshold. The detection device 230 selects the adhesive coating defect contour region 631 whose area is greater than the area threshold, and the detection device 230 adds the selected adhesive coating defect contour region 631 to the detection result 222.

The detection device 230 calculates a conversion ratio based on the actual width of the release paper 912 and the image width of the laminated surface image 410. In more detail, the detection device 230 calculates the conversion ratio based on the width of the release paper 912 and the number of pixels of the width of the laminated surface image 410. The detection device 230 calculates the actual area of the release paper 912 corresponding to the adhesive coating defect contour area 631 based on the conversion ratio. In addition, the detection device 230 can also calculate the position of each adhesive coating defect contour area 631 on the laminated surface image 410 based on the conversion ratio.

In some embodiments, the detection device 230 can perform grayscale processing on the laminated image 410 to obtain a grayscale image. The detection device 230 obtains the pixel value of each pixel along the first axis 431 of the grayscale image. The detection device 230 can obtain pixels from both sides of the laminated image 410, or can obtain the pixel value of each pixel from the middle axis of the laminated image 410 to both sides. The following table defines the image parameters of the grayscale image: X The coordinate of the 𝑥-axis (i.e. the first axis) Y The coordinate of the 𝑦-axis (i.e. the second axis) 𝐺𝑟𝑎𝑦(𝑥,𝑦) The pixel value at coordinate (𝑥,𝑦) 𝐺𝑟𝑎𝑦 _{𝑅𝑒𝑠𝑢𝑙gna} (𝑦,𝑦) 𝐺𝑟𝑎𝑦(𝑥,𝑦) is the pixel value after processing 𝑅𝑒𝑓𝑙𝑒𝑐սլ𝑣𝑒𝑇𝑜𝑝ɡ The 𝑦-axis value of the upper boundary of the reflection. The default value is "-1" 𝑅𝑒𝑓𝑙𝑒𝑐սյ𝑣𝑒𝐵𝑜𝑒𝑜𝑚 The 𝑦-axis value of the lower boundary of the reflection. The default value is "-1" Table 1. Definition of image parameters of grayscale images

The detection device 230 traverses the pixel values of each pixel of the grayscale image along the first axis 431. If the accessed pixel 𝐺𝑟𝑎𝑦(𝑥,𝑦) is less than or equal to "40", the detection device 230 sets 𝐺𝑟𝑎𝑦𝑅𝑒𝑠𝑢𝑙𝑡(𝑥,𝑦) to "255", which means that the pixel is set to white and indicates that the pixel has glue, thus forming glue reflection. If the accessed pixel 𝐺𝑟𝑎𝑦(𝑥,𝑦) is greater than "40", the detection device 230 sets 𝐺𝑟𝑎𝑦𝑅𝑒𝑠𝑢𝑙𝑡(𝑥,𝑦) to "0", which means that the pixel is set to black and indicates that the pixel has no glue reflection. And the detection device 230 checks whether there is a non-reflective feature for the pixel set to "black". The value "40" is for example, and can actually be determined according to the ambient brightness of the coating device or the intensity of the external light source 911.

After the detection device 230 finds that a "black" pixel is generated, the detection device 230 performs a difference calculation based on the "black" pixel 𝐺𝑟𝑎𝑦(𝑥,𝑦) and the pixel separated by a specified distance (Δ𝑦) on the second axis 432. The difference calculation is as follows: 𝐺𝑟𝑎𝑦(𝑥,𝑦-Δ𝑦)-𝐺𝑟𝑎𝑦(𝑥,𝑦). If the difference between the two pixels is less than or equal to "20", the detection device 230 will regard the "black" pixel 𝐺𝑟𝑎𝑦(𝑥,𝑦) as having glue reflection. The detection device 230 sets the 𝑦-axis value of the "black" pixel 𝐺𝑟𝑎𝑦(𝑥,𝑦) to the lower boundary of the reflection (𝑅𝑒𝑓𝑙𝑒𝑐𝑡𝑖𝑣𝑒𝐵𝑜𝑡𝑡𝑜𝑚𝑥). If the difference between the two pixels is greater than "20", it means that the brightness change of the two pixel values is large, so the detection device 230 determines that the "black" pixel 𝐺𝑟𝑎𝑦(𝑥,𝑦) is the reflection difference caused by no glue. The detection device 230 records all the pixel positions with no glue reflection on the first axis 431 as no reflection characteristics. Therefore, the detection device 230 can find the boundary of the laminated image 410 according to the non-reflective feature in the first axis 431 and the non-reflective feature in the second axis 432. The detection device 230 defines the range covered by the boundary set of the laminated image 410 as the target block 420, as shown in FIG. 12 .

In some embodiments, the detection device 230 may perform binarization processing on the plastic surface image 410 and obtain a binarized image. The detection device 230 obtains the non-reflective features in the first axial direction 431 and the second axial direction 432 from the binarized image. The detection device 230 obtains the first boundary 441 and the second boundary 442 according to the non-reflective features of the two axial images, thereby obtaining the target block 420 in the binarized image.

In some embodiments, after the detection device 230 obtains the target block 420, it also performs image smoothing processing on the target block 420, wherein the image smoothing processing can be Gaussian smoothing (Gaussian Blur), mean blur, median blur or exponential blur, etc. The target block 420 can be obtained from the laminated image 410, the grayscale image or the binary image. The output result after the image smoothing processing is referred to as the smoothed image below. The pixels in the smoothed image are called smoothed pixels, and in the above manner, the smoothed pixel is the "pixel position" of the pixel, and the smoothed pixel value is the "pixel value" of the pixel. The following is the definition of each image parameter of the laminated image 410: X The coordinate of the 𝑥-axis (i.e. the first axis) Y The coordinate of the 𝑦-axis (i.e. the second axis) 𝐺𝑟𝑎𝑦(𝑥,𝑦) The pixel value of the laminated image at coordinate (𝑥,𝑦) 𝑅𝑒𝑓𝑙𝑒𝑐սլ𝑣𝑒𝑇𝑜𝑝ɡ The 𝑦-axis value of the upper boundary of the reflection. The default value is "-1" 𝑅𝑒𝑓𝑙𝑒𝑐սյ𝑣𝑒𝐵𝑜𝑒𝑜𝑚 The 𝑦-axis value of the lower boundary of the reflection. The default value is "-1" 𝑅𝑒𝑓𝑙𝑒𝑐ստ𝑣𝑒𝐿𝑒𝑓𝑡𝑅𝑒𝑠𝑢𝑙𝑡 The x-axis value of the left edge of the reflection. The default value is "-1" 𝑅𝑒𝑓𝑙𝑒𝑐սս𝑣𝑒𝑅　𝑔ℎ𝑡𝑅𝑒𝑠𝑢𝑙ս The x-axis value of the right edge of the reflection. The default value is "-1" 𝑆𝑚𝑜𝑜gnaℎ(𝑦,𝑦) Smooth the pixel values of the image 𝑉𝑎𝑔𝑢𝑒𝑁 The difference between the pixels of the laminated image and the smoothed pixels 𝑉𝑎𝑔𝑢𝑒𝐴𝑣𝑒𝑟𝑎𝑔𝑒 The average difference between the pixels of the laminated image and the smoothed pixels 𝑉𝑎𝑔𝑢𝑒(%) Fuzziness Table 2. Definition of fuzziness related parameters of target blocks

The detection device 230 traverses the pixels of the target block 420 along the first axis 431, and selects the pixel value of the corresponding position in the smooth image according to the pixels obtained from the target block 420. Among them, all pixels of the target block 420 on the first axis 431 are called pixel distribution set 641. The detection device 230 is used to obtain a smooth pixel distribution set 642 in the smooth image and corresponding to the pixel distribution set 641. The detection device 230 performs a difference calculation on each pixel of the pixel distribution set 641 and the smooth pixel distribution set 642, please refer to FIG. 13. The horizontal axis of FIG. 13 is the pixels on the first axis 431 of the target block 420 and part of the plane image, and the vertical axis is the pixel value of the pixel. The detection device 230 obtains the collective difference of the target block 420 according to the results of all difference calculations, as shown in FIG14. The blue curve in FIG14 is the difference between the pixel values of the target block 420 and the plane image in FIG12. The thick black horizontal line in FIG13 is the blur 643, the calculation of which will be described later.

The collective difference represents the difference between each pixel of the laminated image 410 (which may also be a grayscale image or a binary image) and the smoothed image in the first axis 431, and takes the absolute value of the difference, that is, the collective difference 𝑉𝑎𝑔𝑢𝑒𝑁=｜𝑆𝑚𝑜𝑜𝑡ℎ(𝑥,𝑦)-𝐺𝑟𝑎𝑦(𝑥,𝑦)｜. When the difference of the pixel is not 0, the detection device 230 accumulates "1" into the collective difference. Then, the detection device 230 sums up all the differences and divides the sum by the aggregate difference, that is, the difference average 𝑉𝑎𝑔𝑢𝑒𝐴𝑣𝑒𝑟𝑎𝑔𝑒=∑(VagueN)/N.

The detection device 230 adjusts the numerical range of the difference average value so that the numerical range of the adjusted difference average value falls between 0 and 100. In other words, the adjusted difference average value is the blurriness 643 (%). The detection device 230 can pre-subtract the reference value from the difference average value of each pixel. When the reference value is subtracted from the difference average value, the output result may be less than "0". Therefore, the detection device 230 makes a judgment on each output result. If the output result is less than "0", the detection device 230 sets the difference average value of the pixel to "0". If the output result is greater than "2", the detection device 230 sets the difference average value of the pixel to "2". For pixels whose output results fall between "0" and "2", the detection device 230 does not adjust the difference average value of the pixel.

Finally, the detection device 230 multiplies the difference average of all pixels by the value "50". Since all the difference averages fall within the range of 0 to 2, the result after multiplying by the value "50" will fall within the range of 0 to 100. Finally, the detection device 230 subtracts the multiplier result from the value "100" to obtain the blur 643, and the blur 𝑉𝑎𝑔𝑢𝑒＝100-(𝑉𝑎𝑔𝑢𝑒𝐴𝑣𝑒𝑟𝑎𝑔𝑒*50) is obtained. Generally speaking, the lower the blur 643 is, the cleaner the lens of the camera device 210 is, which means that there is no glue stuck on the lens of the camera device 210. On the contrary, the higher the blur 643 is, the more glue is stuck on the lens. After the detection device 230 obtains the blur 643, the detection device 230 determines whether the blur 643 is greater than a blur threshold value. When the blur 643 is greater than a blur threshold value, the detection device 230 outputs a cleaning signal to prompt the staff to clean the camera device 210.

In some embodiments, the detection device 230 is used to obtain each depth value in the target block 420 according to the pixel value of the area starting point 711, the blur 643 and the brightness compensation value. The target block 420 can be obtained from the plastic surface image 410, the grayscale image or the binary image. The detection device 230 adjusts the pixel value of the area starting point 711 according to the blur 643 and the brightness compensation value. The brightness compensation value is the corresponding compensation value of the ambient brightness of the spraying device 100. The detection device 230 calculates the light intensity value Light of each pixel according to the brightness compensation value and the blur 643. The light intensity value Light can be expressed as follows:

Wherein, Δ1 and Δ2 are constants. The detection device 230 calculates the depth value for each pixel. For the convenience of explanation, the detection device 230 refers to the depth value of the currently selected pixel as the first depth value. Similarly, the detection device 230 calculates and obtains the depth value of the remaining pixels according to the pixel depth and blur 643. The detection device 230 stores the depth values of all pixels in the depth difference set.

Please refer to FIG. 15 . The table T on the left side of FIG. 15 records the pixel position, pixel value, pixel depth and depth value list T of the target block 420 on the first axis 431. After traversing all the pixels of the target block 420 on the first axis 431, the detection device 230 obtains the corresponding depth value according to the pixel value, light intensity value and pixel depth of each pixel. Then, the detection device 230 calculates the difference between the pixel value of the region starting point 711 and the pixel values of the remaining pixels and records each difference in the depth difference set, as shown in the table L on the right side of FIG. 15 .

In FIG. 15 , the pixel at the pixel position "1834" is the region start point 711. The detection device 230 uses the depth value of the region start point 711 as the depth threshold value, and compares the difference between each second depth value. As can be seen from FIG. 15 , the pixel value of the pixel position "1835" is "127". Therefore, the difference between the region start point 711 and the pixel position "1835" is "-3". The depth threshold value of the region start point 711 is "11.904", so the difference "-3" of the pixel position "1835" is less than the depth threshold value "11.904" of the region start point 711. Therefore, the detection device 230 determines that the comparison result between the pixel position "1834" and the pixel position "1835" is "not exceeded". In other words, the pixel position "1834" is judged as a normal coating area with "no glue".

In some embodiments, the detection device 230 can be centered on the region starting point 711 and set a range of the first quantity. The range of the first quantity is at least 1 pixel. Taking Figure 15 as an example, the region starting point 711 is the pixel position "1834", and the first quantity can be set to "10" pixels, which means that the pixel position "1834" is compared with the pixel positions "1835" to "1844". When the difference of any pixel in the range of the first quantity is greater than the depth threshold, it means that the region starting point 711 belongs to the abnormal coating area 510, and vice versa. Therefore, the difference between the region starting point 711 (corresponding to the pixel position "1834") and the pixel position "1839" in Figure 15 is greater than the depth threshold, and the pixel position "1839" is within the range of the pixel position "1834" and the first quantity. Therefore, the pixel position "1834" is classified as the abnormal coating area 510 of "having glue", which means that the pixel position "1834" is "having glue".

Compared with the comparison method of two adjacent pixels, the comparison method of the region starting point 711 and the first quantity can avoid the misjudgment of no glue caused by the unlikely difference of continuous pixel changes. As long as the detection device 230 determines that the difference of any group in the first quantity is "exceeding", it can be judged as "there is glue". And the detection device 230 can select the next pixel as the new region starting point 711 and perform a new difference judgment, so as to increase the speed of inspection.

FIG. 16 is a partial area of the target block 420 of FIG. 11 , which corresponds to 6 pixel positions on the first axis 431, and the 6 pixel positions are respectively "1831 (pixel value 118)", "1832 (pixel value 128)", "1833 (pixel value 133)", "1834 (pixel value 130)", "1835 (pixel value 127)", and "1836 (pixel value 125)". The pixel values, depth threshold values, and differences corresponding to each pixel are also listed in FIG. 16 . The pixel position "1831" is taken as the region starting point 711, and the corresponding depth threshold value is "11.928". The difference between the pixel positions "1833" and "1831" is "15", so the difference value exceeds the depth threshold value. Therefore, the detection device 230 determines that the pixel position "1831" is the abnormal coating area 510. The detection device 230 can select the next pixel position "1832" to perform the next round of difference judgment.

The detection device 230 may also choose to perform a difference comparison at intervals of at least one pixel. In the aforementioned paragraph, the detection device 230 performs a difference comparison at intervals of one pixel. In other embodiments, the detection device 230 may perform a difference comparison at intervals of two or more pixels. For example, the detection device 230 uses the pixel position "1834" as the region start point 711 and chooses to perform a comparison at intervals of two pixels. The detection device 230 will select the pixel position "1836" and the pixel position "1834" for a difference comparison to determine whether the region start point 711 belongs to the abnormal coating region 510.

The detection device 230 can perform the aforementioned difference comparison from any direction of the first axis 431 of the target block 420. For example, the detection device 230 can perform a difference comparison on each pixel from the left side to the right side of the first axis 431. Alternatively, the difference comparison can be performed simultaneously from the left and right sides to the center of the first axis 431. After the detection device 230 traverses all pixels on the first axis 431, the detection device 230 will move to the next pixel in the second axis 432 and begin to traverse all pixels in the first axis 431 and compare the differences.

The processing system 200 and processing method for detecting adhesive defects provide a method of using digital images to detect the spraying status of adhesive defects in real time, so as to prompt whether the depth of the applied adhesive film meets the set standard. Since the detection device 230 can immediately confirm the depth of the adhesive film on the release paper 912 after spraying, the detection device 230 can immediately issue relevant warning notifications to prompt production line personnel to eliminate relevant errors.

100: Spraying equipment 200: Processing system 210: Camera equipment 220: Storage equipment 221: Image detection program 222: Detection results 230: Detection equipment 410: Glue surface image 420: Target block 431: First axis 432: Second axis 441: First boundary 442: Second boundary 510: Abnormal coating area 610: Contour number table 611: Starting pixel 620: Contour detection frame 621: First contour detection frame 622: Second contour detection frame 631: Glue defect contour area 641: Pixel distribution set 642: Smooth pixel distribution set 643: Blur 711: Region start point 911: External light source 912: Release paper T: List L: Table S310, S320, S330, S340, S350, S710, S720, S730, S740, S750, S760: Steps

FIG. 1 is a partial schematic diagram of a spray-coated release paper of an embodiment. FIG. 2 is a schematic diagram of a detection device component of an embodiment. FIG. 3 is a schematic diagram of an image detection program flow of an embodiment. FIG. 4 is a schematic diagram of a rubber surface image and a target block of an embodiment. FIG. 5 is a schematic diagram of a target block and an abnormal coating area of an embodiment. FIG. 6 is a schematic diagram of a contour numbering table of a local target block of an embodiment. FIG. 7 is a schematic diagram of a contour marking process of an embodiment. FIG. 8 is a schematic diagram of the movement of a contour detection frame of an embodiment. FIG. 9 is a schematic diagram of an object number of a contour numbering table of an embodiment. FIG. 10 is a schematic diagram of a contour numbering table for setting object numbering for the second time of an embodiment. FIG. 11 is a schematic diagram of the outline area of the glue defect in the target block of an embodiment. FIG. 12 is a schematic diagram of the first boundary and the second boundary of the glue surface image of an embodiment. FIG. 13 is a schematic diagram of the pixel distribution set and the smooth pixel distribution set of an embodiment. FIG. 14 is a schematic diagram of the difference in pixel values between the target block and the plane image of an embodiment. FIG. 15 is a schematic diagram of the depth difference set of a part of the target block of an embodiment. FIG. 16 is a schematic diagram of the difference comparison of different regional starting points of an embodiment.

200:Processing system

210: Camera equipment

410: Laminated image

230: Testing equipment

220: Storage equipment

221: Image detection program

222:Test results

Claims (17)

A processing system for detecting adhesive coating defects, comprising: a camera device for photographing an adhesive surface image; and a detection device connected to the camera device, the detection device captures a target block from the adhesive surface image, the detection device traverses each pixel in a first axis direction of the target block, the currently selected pixel is a region starting point, the detection device obtains a first depth value of the region starting point, the detection device obtains a second depth value of the remaining pixels in the first axis direction, and the detection device compares the first depth value with each pixel in the first axis direction. The second depth value is subjected to difference calculation to obtain multiple differences, and the differences are recorded in a depth difference set. The detection device determines whether the starting point of the area belongs to an abnormal coating area or a normal coating area according to the depth difference set. The detection device performs a contour marking process on the normal coating area to generate at least one coating defect contour area. The detection device generates a detection result according to the coating defect contour areas. A processing system for detecting adhesive defects as described in claim 1, wherein the detection equipment performs a grayscale processing on the adhesive surface image to obtain a grayscale image. A processing system for detecting adhesive defects as described in claim 2, wherein the detection device obtains a first boundary based on a non-reflective feature in the first axis of the grayscale image, the detection device obtains a second boundary based on another non-reflective feature in the second axis of the grayscale image, and the detection device captures the target block from the grayscale image based on the first boundary and the second boundary. A processing system for detecting adhesive defects as described in claim 1, wherein the detection equipment performs a binarization process on the adhesive surface image to obtain a binary image. A processing system for detecting adhesive defects as described in claim 4, wherein the detection device obtains a first boundary based on a non-reflective feature in the first axis of the binary image, the detection device obtains a second boundary based on the non-reflective feature in the second axis of the binary image, and the detection device captures the target block from the binary image based on the first boundary and the second boundary. A processing system for detecting adhesive defects as described in claim 1, 3 or 5, wherein the detection device performs an image smoothing process on the target block to obtain a smooth pixel distribution set, the detection device also obtains a pixel distribution set of the target block, the detection device obtains a set difference based on the difference between the smooth pixel distribution set and the pixel distribution set, and the detection device obtains a blur based on the set difference. A processing system for detecting adhesive defects as described in claim 6, wherein the detection device determines that the fuzziness is greater than a fuzziness threshold value, and the detection device outputs a clean signal. A processing system for detecting adhesive defects as described in claim 6, wherein the detection device is used to obtain the first depth value and the second depth value of the remaining pixels based on a pixel value at the starting point of the area, the blurriness and a brightness compensation value. A processing system for detecting adhesive defects as described in claim 1, wherein the detection device takes the starting point of the area and selects a first number of pixels in the first axis direction, and the detection device calculates the first depth value and the second depth values of the starting point of the area and the first number of pixels. A processing system for detecting adhesive defects as described in claim 9, wherein the detection device determines whether the differences in the depth difference set are all greater than a depth threshold value. If any of the differences is greater than the depth threshold value, the detection device determines that the starting point of the area is the abnormal coating area. A processing system for detecting glue defects as described in claim 1, wherein the detection device establishes a contour number table based on an object number of the abnormal coating area and the normal coating area, the position of each pixel of the target block corresponds to a different object number in the contour number table, the detection device selects one of the pixels from the contour number table, the selected pixel is a starting pixel, the detection device performs the contour marking process on the starting pixel, re-marks the starting pixel and the adjacent pixels with an object number, and the contour marking process assigns the same object number to the same glue defect contour area. A processing system for detecting coating defects as described in claim 1, wherein the detection device selects a region of the coating defect contour region whose area is greater than an area threshold value, and the detection device adds the selected coating defect contour region to the detection result. A processing method for detecting adhesive coating defects, comprising: A detection device captures a target block from an adhesive surface image; The detection device calculates a depth difference set between a plurality of pixels in a first axial direction of the target block; The detection device classifies the corresponding pixel into an abnormal coating area or a normal coating area according to each depth difference set; The detection device performs a contour marking process on the normal coating area to generate at least one adhesive coating defect contour area; and The detection device generates a detection result according to the adhesive coating defect contour area. The processing method for detecting adhesive defects as described in claim 13, wherein the step of extracting the target block from the adhesive surface image by the detection device includes: The detection device performs a grayscale processing on the adhesive surface image to obtain a grayscale image; The detection device obtains a first boundary based on a non-reflective feature in the first axis of the grayscale image; The detection device obtains a second boundary based on the non-reflective feature in the second axis of the grayscale image; and The detection device extracts the target block from the grayscale image based on the first boundary and the second boundary. The processing method for detecting adhesive defects as described in claim 13, wherein the step of capturing the target block from the adhesive surface image by the detection device includes: The detection device performs a binarization process on the adhesive surface image to obtain a binarized image; The detection device obtains a first boundary based on a non-reflective feature in the first axis of the binarized image; The detection device obtains a second boundary based on the non-reflective feature in the second axis of the binarized image; and The detection device captures the target block from the binarized image based on the first boundary and the second boundary. The processing method for detecting adhesive coating defects as described in claim 13, wherein the step of the detection device capturing the target block from the adhesive surface image includes: The detection device performs an image smoothing process on the adhesive surface image to obtain a smooth pixel distribution set; The detection device obtains a pixel distribution set of the adhesive surface image; The detection device obtains a blur according to the smooth pixel distribution set and the pixel distribution set; and The detection device determines that the blur is greater than a blur threshold value, and the detection device outputs a clean signal. The processing method for detecting adhesive defects as described in claim 13, wherein the step of the detection device calculating the depth difference set between the plurality of pixels in the first axis of the target block includes: The detection device traverses each pixel in the first axis of the target block and takes the currently selected pixel as a region starting point; The detection device obtains a first depth value of the region starting point; The detection device obtains a second depth value of the remaining pixels in the first axis; and The detection device calculates the difference between the first depth value and each second depth value to obtain the depth difference set of the region starting point.