TWI473989B - Automatic optical inspection system for defect detection of dental floss stick - Google Patents

Description

Dental floss 瑕疵 automatic optical detection system

BACKGROUND OF THE INVENTION 1. Field of the Invention This invention relates to an automatic optical inspection system, and more particularly to an automatic optical inspection system for recognizing a missing image, a line injury, and a dirty image signal of a dental floss.

With the development of the economy, modern people not only pay attention to the health of the diet, but also pay more attention to the health of the individual's mouth. Therefore, people often use floss sticks to clean between the teeth, and various flosses with clean teeth function. The stick came into being. Because the dental floss stick belongs to personal hygiene products, it is more demanding by consumers than the general plastic injection molded products in terms of the cleanliness and quality of the products.

Due to the manufacturing cost considerations of the floss bar products, most of the work on the inspection and elimination of defective products in the industry is carried out manually. Therefore, it is difficult to achieve consistency in the judgment standard of the products during the inspection process. The efficiency depends on the concentration of the inspectors, so the inaccurate management of the products limits the yield and quality of the products.

Plastic injection molding processing is one of the typical modern mass production methods. Its advantages are high consistency of processing products and fast production speed. It is a common processing method for producing consumable products. In the production line with large production and fast operation, how to have Efficient and reliable inspection of whether each product is flawed is a topic of concern to every large-scale producer since the industrialization. In today's computer technology, using computers and various sensors to sense and judge flaws, The use of computer-controlled motion mechanisms to exclude or collect counterfeit products is a common pattern in today's production line inspection procedures.

On the other hand, in the conventional industrial automatic inspection, the surface flaw of a test object is often checked by using an automated optical inspection (AOI). The automatic optical detecting device captures an image of one of the flat displays to be tested by means of a Charge Coupled Device (hereinafter referred to as CCD). In the image capturing process, the array CCD is moved along a image capturing path between a plurality of fixed points, and a unit detection image is captured for each unit of the capturing range at a certain point. The unit detection images are combined into a complete detection image for comparison detection. The detected image can be detected to have defect defects compared to the detected image. The industrial automatic optical inspection equipment has high precision, no artificial visual limitation, and the detection speed is fast. However, its cost is necessarily relatively high.

Therefore, it is necessary to provide an automatic optical detection system capable of recognizing the image signal of the dental floss, which can be far cheaper than the general industrial optical detection equipment, and at the same time, in the precision of the optical inspection, the plastic injection of the plastics for the people can be correctly tested. Product requirements.

The main object of the present invention is to provide an automatic optical detecting device suitable for detecting the flaw of a dental floss rod, and based on the existing injection molding production line, the minimum change of the existing equipment is premised, and the injection process is added in the production process. Fully automatic optical inspection of molded products. Therefore, this design uses a personal computer as the basis for control, and uses a webcam (WEBCAM) as the imaging medium. In terms of the construction of the device, it can be much cheaper than the general industrial optical inspection equipment, and at the same time in the precision of optical inspection. It can meet the requirements of correctly testing the plastic injection products of the people's livelihood.

In order to achieve the above object, the present invention provides an automatic optical detecting device suitable for detecting flaws of a dental floss bar, the automatic optical detecting device comprising: a conveying unit carrying a whole piece containing a plurality of dental floss rods to be detected for transporting Directional transport; a control unit responsible for the control of the start, stop and pause of the transport unit; an image capture unit for capturing a wide range of images of the entire floss bar and a plurality of individual floss bars of the detection device a range of image processing; an information processing unit for processing the input of the image signal of the floss stick captured by the image capturing unit, the identification of the image of the defect, and the movement control of the dental floss; and a structural support platform, Carrying all of the above components (ie, transport unit, control unit, imaging unit and information processing unit) and the function of transporting the dental floss bar, the automatic optical detection of the dental floss can be easily accomplished.

The invention is characterized in that a webcam (WEBCAM) is used as a photographing medium for measuring images, and an AC motor is provided to provide power for transporting a target object, and a structural support platform having functions of carrying all components and transporting objects, and The personal computer realizes signal integration and control. In the existing production process, an intermediate workstation is formed to detect various defects of the target, and data records or alarms are issued for reference by the operator.

The automatic optical detecting device of the present invention is preferably applied to detect the flaw of a dental floss stick, but is not limited thereto, and can be used as long as the flaw characteristic of the object to be tested can be photographed and visualized by the optical detecting device. On various plastic or electronic products, the identified form of 瑕疵 can be surface contamination of the object to be tested, contoured corners or part displacement.

In an embodiment of the present invention, the transport unit includes a motor mounted on the structural support platform and a pulley module connected to the motor. The motor is a variable frequency motor capable of adjusting the rotational speed within a certain distance range. .

In an embodiment of the present invention, the control unit includes a plurality of proximity switches mounted on the structural support platform to provide a non-contact signal as a basis for positioning the floss bars, and the connection unit and the information processing An input/output control card for the unit.

In an embodiment of the invention, the image capturing unit comprises a plurality of network cameras, one of the network cameras is responsible for one of the whole floss bars, and the rest is taken in a plurality of side by side manners. The entire floss bar is arranged in a plurality of individual floss bars.

In an embodiment of the invention, the information processing unit is a personal computer.

In an embodiment of the present invention, the information processing unit further includes a startup determination application software program, determining whether the entire piece of the dental floss bar is completely displayed in a range of a wide range of images, and an input application software program is responsible for inputting the Whole floss bar image signal.

In an embodiment of the present invention, the information processing unit further includes an image recognition software library for identifying the image signal of the dental floss, such as lack of material, line damage and dirt.

In an embodiment of the invention, the structural support platform further comprises a guiding plate, the guiding plate limits the degree of freedom of the floss bar, and ensures that the feeding of the whole floss bar can be detected by the detecting device. come out.

In an embodiment of the present invention, the material shortage is caused by the fact that the flow of the floss bar semi-finished product during the injection molding stage cannot completely fill the cavity, and the wire defect comprises breaking, loosening, and The unevenness of the thickness; and the contamination of the thimble into the mold causes the stain on the ejector portion of the semi-finished product of the dental floss.

In an embodiment of the present invention, the information processing unit identifies the missing material of the entire dental floss image and identifies the image of the dental floss bar to be compared by an image processing software identification library. The line hurts the dirt.

In an embodiment of the present invention, the information processing unit identifies the image of the missing piece of the whole piece of the dental floss includes the following steps: using the block comparison module of the identification library according to a preset standard tooth The bar image searches for the position of each of the identical floss bars; the standard floss bar is subtracted from the to-be-matched floss bar; and the analysis module using the identification library Obtain the missing material information of the floss bar to be compared.

In an embodiment of the present invention, the information processing unit is configured to identify the image of the line flaw to be compared with the floss bar, and further includes the following steps: the edge detection module of the identification library is for the waiting ratio Detecting an image of a peripheral edge portion of one of the floss rods to obtain an area occupied by the wire rod to be compared with the floss bar; the binarization module of the identification library is for the wire region The image is binarized, and the gray value is counted to obtain an area of the wire of the to-be-matched floss bar in the binarized image; the area of the wire is divided by the length of the wire to obtain the The line scar data can be obtained by comparing the width of the wire; and by comparing the width of the previously stored standard wire with the width of the wire to be compared.

In an embodiment of the present invention, the information processing unit further comprises the following steps: identifying the image of the dirty tooth to be compared with the dental floss bar, and capturing the image of the to-be-matched floss bar in a light-emitting environment. Obtaining a preliminary image of a black stain; binarizing the preliminary image of the black stain by the binarization module of the identification library; and using the analysis module of the identification library to obtain the image All the analysis parameters, after deducting the standard analysis parameters stored in advance, can obtain the analysis parameters formed by the stain.

In an embodiment of the present invention, the information processing unit further includes the following steps for avoiding the residual image caused by the image subtraction on the edge of the missing image: expanding the module by the pixel adding module of the identification library The boundary of the edge of the floss bar in the missing image; the reverse binarization module of the identification library is used to exchange the black and white missing image; the image dividing module using the identification library Dividing the inverse binarized missing image by the pre-stored standard complete image, obtaining a complete image of the grayscale value of the missing material portion maintaining the original value; and the missing grayscale value image and the inverse Subtracting the image of the missing material to obtain a result image of the missing portion is lower than the surrounding image; and using the analysis module of the identification library to obtain the analysis value of the missing portion, The judgment, classification and storage of the cockroach can be further performed.

The features, implementation and efficacy of the present invention will be described in detail with reference to the drawings.

Floss bar, the material usually contains two parts of plastic and nylon thread. The processing machine required in the manufacturing process has a plastic injection molding machine and a cutting machine, and the manufacturing process firstly puts the wire of the floss into the wire. The plastic is injected into the cavity of the mold. This procedure is called the pull wire. The implementation medium of the wire drawing process is usually carried out by a single degree of freedom mechanical arm. During this insertion process, the wire may be placed with skew or tension. Unevenness, which causes defects in the wire, such as slack, displacement or different degrees of breakage of the wire, as shown in Figures 6A to 6B.

After the wire is placed, the mold is sealed and filled, and the plastic material is filled into the cavity by heating and pressing. When the filling is completed and cooled, the finished product is removed from the cavity, and the mold is ejected. The design of the needle, and the ejector pin is to be telescoped in the pinhole, so that the lubricating oil is injected into the portion where the friction occurs around it, and the lubricating oil may penetrate into the cavity during the plastic filling process. Dirty, as shown in Figures 7A through 7B.

During the process of filling the cavity, there may be a lack of heating or pressurization of the mold part, so that the plastic material is not completely filled in the cavity, and the lack of volume of the rod body is formed at the tip of the floss bar.瑕疵, or the unsuccessful demolding process may also cause bending of the tip portion, as shown in Figures 5A-5B.

Please refer to Figure 1 and refer to Figure 2 together. 1 is a schematic plan view of an automatic optical detecting device according to an embodiment of the present invention. Fig. 2 is a perspective assembled view of an automatic optical detecting device according to an embodiment of the present invention.

The automatic optical detecting device 10 shown in Figs. 1 and 2 is suitable for identifying the defect of a dental floss bar. The automatic optical detecting device 10 comprises: a conveying unit 100, which carries a whole piece of dental floss 60 and contains a plurality of The detecting floss bar is transported in a conveying direction; a control unit 200 is responsible for controlling the start, stop and pause of the conveying unit 100; and an image capturing unit 300 for capturing a wide range of images of the entire floss bar 60 a plurality of small-scale images of the individual floss bars of the detecting device; an information processing unit 400 for processing the input of the floss image signal captured by the image capturing unit 300, the identification of the image of the defect, and the movement control of the floss bar; And a structural support platform 500, which assembles and carries all of the above components (ie, the transport unit 100, the control unit 200, the image capturing unit 300, and the information processing unit 400) and the transport floss bar.

Referring to FIG. 1 and referring to the automatic optical detecting device 10 shown in FIG. 2, the conveying unit 100 of FIG. 1 includes a motor 110 mounted on the structural support platform and a belt connected to the motor. The wheel module 120, as shown in Fig. 2, includes a motor 110 that is capable of adjusting the rotational speed over a range of distances.

Please refer to Figure 3 and refer to Figure 1 together. 3 is a perspective view of a proximity switch 210 according to an embodiment of the present invention, wherein the control unit 200 includes a plurality of proximity switches 210 mounted on the structural support platform 500 to provide a non-contact signal as a basis for positioning the floss bar. And an input/output control card (not shown) connected to the transport unit 100 and the information processing unit 400. However, the present invention can also change the proximity switch 210 on the control unit 200 that provides the non-contact signal as the basis for the positioning of the floss bar, and is not limited to the embodiment disclosed in the present invention.

The automatic optical detecting device 10 as shown in FIG. 1 , wherein the image capturing unit 300 comprises a plurality of network cameras, one of which is responsible for one of the entire floss bars 60, and the rest is plural. Side by side, draw a plurality of individual floss bars of the entire floss bar 60. It should be noted that the network camera used in this embodiment is not limited to the actual use timing. The present invention can select different camera types according to actual use requirements.

The automatic optical detecting device 10 according to an embodiment of the present invention, wherein the information processing unit 400 is a personal computer. The information processing unit 400 includes a startup determination application software program for determining whether the entire floss bar 60 is completely displayed in a range of a wide range of images, and an input application software program for inputting the image signal of the entire floss bar 60. The information processing unit 400 further includes an image processing software identification library for identifying image signals such as missing bars, line injuries and dirt.

Please refer to Figure 4. Figure 4 is a schematic view of a guide plate according to an embodiment of the present invention. The automatic optical detecting device 10 of the embodiment, wherein the structural support platform 500 further comprises a guiding plate 510, the guiding plate 510 limits the degree of freedom of the floss bar, and ensures that the feeding of the whole floss bar 60 can be The detecting device detects that the invention can select different types of guiding plates according to actual use requirements. .

An automatic optical detection method for identifying a defect of a dental floss in accordance with an embodiment of the present invention includes the following steps: taking a standard floss stick from an image capturing unit 300 to obtain a standard image, As a basis for determining the defect of the floss bar; whether the entire floss bar 60 is completely taken into the range containing a large range of images of the whole floss bar 60 by the image capturing unit 300; The unit 100 transports the entire floss bar 60 and transports it in a transport direction; the control unit 200 controls the transport unit 100 to pause at more than one image capturing position; and the image taking unit 300 captures all the floss bars 60 Comparing the floss bars to more than one image capturing position to obtain a small range of images of the floss bars to be compared; processing a large range of floss bars 60 by an information processing unit 400 to identify the entire piece of teeth The lack of material of the wire rod 60, and then input a small range of images to identify the image of the floss bar to be compared, determine the line injury and dirtiness of the floss bar to be compared; and for the specific type of flaw, by detection Value comparison Storage of the standard values to get the complete information of the defect.

Please refer to Figures 5A through 7B. FIG. 5A is an image of a defect of the missing material of the dental floss semi-finished product in the injection molding stage due to the inability of the filling material to completely fill the cavity. FIG. 5B is a result of direct subtraction of the missing image and the complete image of the dental floss semi-finished product according to an embodiment of the present invention, and the image of the residual pixel is generated at the edge of the contour of the floss bar. Fig. 6A is an image of a dental floss bar containing a slack of a wire used in an embodiment of the present invention. Fig. 6B is a view showing an image of a broken tooth of a dental floss in accordance with an embodiment of the present invention. Fig. 7A is an image of stains caused by stains and the like caused by the oil stain of the thimble of the mold due to the penetration of the thimble of the mold to the ejector portion of the semi-finished product of the dental floss. Fig. 7B is an image of the stain of the dental floss of the dental floss in accordance with an embodiment of the present invention, and the dirty portion presents a black image.

The automatic optical detecting method according to an embodiment of the present invention, as shown in FIG. 5A and FIG. 5B, wherein the material shortage is caused by the flow material in the injection molding stage cannot completely fill the cavity, and the defect is formed. As shown in Figs. 6A and 6B, wherein the line injury includes breakage, slack, and unevenness of the thickness of the wire used; and as shown in Figures 7A and 7B, where the dirt is the thimble infiltration of the mold The oil stain causes the stain of the ejector portion of the semi-finished product of the dental floss.

Please refer to Figures 5A through 7B. An automatic optical detection method according to an embodiment of the present invention, wherein the information processing unit 400 identifies the missing material of the entire dental floss 60 image and identifies the dental floss bar by an image processing software identification library. The line of the image is hurt and dirty.

Please refer to Figures 5A and 5B and refer to Figure 5C. FIG. 5C is a flow chart for identifying the missing material of the dental floss according to an embodiment of the present invention. An automatic optical detection method according to an embodiment of the present invention, wherein the information processing unit 400 recognizes the image of the missing piece of the entire dental floss 60 includes the following steps: using the identification module of the identification library according to the preset setting A standard floss image searches for the same position of each floss bar; the standard floss bar is subtracted from the floss bar to be compared; and the analysis module using the identification library Get the missing material information of the floss bar to be compared. However, the present invention can also be implemented by a missing function identification software of similar functions, and is not limited to the embodiment disclosed by the present invention.

Please refer to Figures 6A to 6D. Fig. 6C is a view showing the line flaw of the dental floss rod according to an embodiment of the present invention, and the average width value of the converted wire is used to obtain an image of a high value of the line wound portion in the image. Fig. 6D is a flow chart showing the identification of the wire scar of the dental floss according to an embodiment of the present invention. To check the occurrence or failure of the line injury in the image, the logic used is to divide the area occupied by the wire to be recognized in the image by the length of the wire in the image, thus obtaining the average width of the wire in the image. For a wire with a wire injury, the average width will be much higher or lower than the average width of the normal wire, as shown in Figure 6C. An automatic optical detection method according to an embodiment of the present invention, wherein the information processing unit 400 further includes the following steps in the image identifying the line flaw of the floss bar: the edge detection module of the identification library is The image of the peripheral edge of one of the wires of the floss bar is detected, and the area occupied by the wire rod of the floss bar is obtained; and the image of the wire area is determined by the binarization module of the identification library. Performing binarization and counting the gray value thereof to obtain an area of the wire to be compared with the floss bar in the binarized image; dividing the area of the wire by the length of the wire to obtain the width of the wire to be compared; As well as the ratio of the width of the standard wire stored in advance to the width of the wire to be compared, the complete information of the wire defect can be obtained. However, the present invention can also be implemented by a similarly functional line scar recognition software, and is not limited to the embodiment disclosed by the present invention.

Please refer to Figures 7A and 7B and refer to Figure 7C. Fig. 7C is a flow chart showing the identification of the dirt of the dental floss stick according to an embodiment of the present invention. The automatic optical detecting method according to an embodiment of the present invention, wherein the information processing unit 400 identifies the image of the dirty sputum to be compared, further includes the following steps: treating the flossing bar in a light-emitting environment Obtaining a preliminary image of a black stain; binarizing the preliminary image of the black stain by the binarization module of the identification library; and using the analysis module of the identification library to obtain all the analysis parameters in the image The analytical parameters formed by the stain can be obtained by subtracting the standard analysis parameters stored in advance. However, the present invention can also be implemented by a similar function of the dirt recognition software, and is not limited to the embodiment disclosed by the present invention.

Please refer to Figures 8A, 8B and 8C. Fig. 8A is an image of the original missing line of the dental floss (Fig. 5A) according to an embodiment of the present invention, which is enlarged by the pixel growth to expand the edge line of the dental floss in the missing material. FIG. 8B is an image of an image result of a dental floss stick undergoing pixel growth and inverse binarization according to an embodiment of the present invention. FIG. 8C is an image of the result of the pixel sticking and the inverse binarization processing of the dental floss rod according to an embodiment of the present invention, and subtracting the image from the complete image.

In the embodiment, the information processing unit 400 further includes the following steps for avoiding the residual image caused by the image subtraction on the edge of the missing image image: expanding the missing image in the missing image by the pixel adding module of the identification library The boundary of the floss bar edge; the black and white two-color missing image is exchanged by the inverse binarization module of the identification library; the image binning module of the identification library is used to inversely binarize the missing image Dividing the pre-stored standard complete image to obtain a complete image of the gray-scale value of the missing material portion and maintaining the original value; subtracting the missing gray-scale value image from the inverse binarized missing image to obtain a The gray level of the missing part is lower than the surrounding image; and the analysis module of the identification library is used to obtain the analysis value of the missing part, which can be further judged, classified and stored.

In the automatic optical detection method of the embodiment, when two different images are subjected to image subtraction, the subtracted image generates residual pixels at the edge of the subject due to the slight difference in the angle at which the object is placed. And become the image noise to hinder the identification work. In order to avoid edge residual pixels after image subtraction, the logical method used is to enlarge the edge boundary of the target, and the edge is used as a neglected area, so that when the image subtraction is performed, the edge of the target will remain. The portion of the pixel will be covered by the enlarged ignored area, and thus the result of the non-residual pixel processing after the image is subtracted. Therefore, the processing step first uses the pixel growth function of the recognition library to expand the edge boundary of the floss bar in the missing material map.

Please refer to Figure 9. Fig. 9 is a view showing an image of a thin wire scar of an enhanced floss stick according to an embodiment of the present invention. The step of strengthening the identification of the flaws of the filaments is performed by the binarization of the identification library module for the processing of the filament detection area, and then the analysis module of the identification library module obtains the white dots of the filament detection area. Pixel data.

In detail, to check whether the occurrence of the filament damage in the shadow occurs, the logic used is to divide the total number of white points of the wire by the length of the wire, and to determine the numerical value of the wire 原, multiply the value by the wire. The length will get a rectangular area, which is the area where all the white points of the wire are concentrated toward the center line, so a normal wire will not have any white spots outside this area, but has a thin wire. The wire is not the same, so the use of this phenomenon to design a new identification logic, that is, for the wire detection outside the wire collection area, to obtain white point statistics and classification processing.

Please refer to Figure 10. Fig. 10 is a view showing an image for enhancing the stain of the dental floss stick according to an embodiment of the present invention. The steps of enhancing the identification of the dirty sputum are as follows: the image matching module of the identification library obtains the position of the escaping area of the floss stick; the edge reinforced module of the identification library is used for the top of the floss bar The edge enhancement is performed in the out-of-area; the black image is interchanged with the white background image by using the inverse binarization module of the identification library; and the analysis and calculation values of the contamination of the ejector portion are obtained by using the analysis module of the identification library. As the basis for determining the contamination of the ejector.

The logic used to enhance the performance verification of the soiling detection is to first take the image of the floss bar in an effective lighting environment, and to map the position of the ejecting rod to the top of the floss bar and the position of the pre-stored floss bar ejecting area. Block comparison ensures that the detection processing frame can be correctly set to the relative position of the image to be tested, and then the edge enhancement is performed in the detection frame, so that the intersection of the dirty feature and the floss bar is black and white, and the processed area is reversed. Value, simplify the image computing environment, and finally count the white pixels in the detection frame except the total area pixels and multiply by the percentage. Analyze the white point pixels (that is, the black top is dirty) as a percentage of the total ejector area. In the case of white, the white pixel is zero, and zero is divided into any number equal to zero, and zero is used as the basis for determining the normality. Therefore, when the dirtyness is detected, if the value is calculated as a constant, it is judged as the top. The out is dirty.

Please refer to Figure 11. Fig. 11 is a view showing an image of a missing material at the end of the dental floss stick according to an embodiment of the present invention. The step of enhancing the missing defect of the end of the dental floss is as follows: the image matching module of the identification library obtains the position of the end region of the floss bar; the binarization mode of the identification library is used. The group binarizes the image of the end region of the floss bar; the edge detection module of the identification library is used to obtain the pixel length of the end region of the floss bar; and the width of the standard tail region stored in advance is to be By comparing the end regions of the alignment, the complete data of the tail end defects can be obtained.

In order to identify such a method, the identification method adopted by the present invention is roughly divided into two steps. One is to compare the workpiece to be tested with the complete piece, and obtain the relative position of the end of the floss bar and set the effective detection frame. Next, the left edge of the floss bar detection frame is detected by the left and right black edges, and the white X coordinate pixel is obtained to the origin X coordinate of the detection frame, and the length of the floss bar tail pixel is calculated. Generally, the end of the dental floss bar lacks the white pixel length, which is usually smaller than the white pixel value at the end of the normal floss bar. With a large number of normal floss bar tests, the normal floss bar white pixel is used as the index basis, if it is smaller than the index number. It is a shortage of materials.

The automatic optical detecting device according to another embodiment of the present invention learns from the research that a PC can be used to integrate the video capture of five web cameras (WEBCAM) by one program and image the captured pictures. Processing, and automatic control of external devices through the I/O card. In terms of performance verification of the defect identification, the production line after the trial run and adjusted to the optimal parameters can be measured with an area larger than the gray scale difference 0.09mm wound wire ² of the 35 or more dirty defects, lack of 3mm and a length greater than After the measurement results of the ten dental floss test pieces, the system detected the sputum detection rate of 90.4%, 80.5% of the soil, and 84.8% of the material.

The present invention has been disclosed in its preferred embodiments, and is not intended to limit the invention, and the present invention may be modified and modified without departing from the spirit and scope of the invention. The scope of protection is subject to the definition of the scope of the patent application.

10. . . Automatic optical inspection device

100. . . Conveyor unit

110. . . motor

120. . . Pulley module

200. . . control unit

210. . . Proximity switch

220. . . Input and output control card

300. . . Image capture unit

400. . . Information processing unit

500. . . Structural support platform

510. . . Guiding plate

60. . . a whole piece of dental floss

1 is a schematic plan view of an automatic optical detecting device according to an embodiment of the present invention.

Fig. 2 is a perspective assembled view of an automatic optical detecting device according to an embodiment of the present invention.

FIG. 3 is a perspective view of a proximity switch according to an embodiment of the present invention.

Figure 4 is a schematic view of a guide plate according to an embodiment of the present invention.

FIG. 5A is an image of a defect in the semi-finished product of the dental floss in the injection molding stage which cannot be completely filled by the cavity.

FIG. 5B is a result of direct subtraction of the missing image and the complete image of the dental floss semi-finished product according to an embodiment of the present invention, and the image of the residual pixel is generated at the edge of the contour of the floss bar.

FIG. 5C is a flow chart for identifying the missing material of the dental floss according to an embodiment of the present invention.

Fig. 6A is an image of a dental floss bar containing a slack of a wire used in an embodiment of the present invention.

Fig. 6B is a view showing an image of a broken tooth of a dental floss in accordance with an embodiment of the present invention.

Fig. 6C is a view showing the line flaw of the dental floss rod according to an embodiment of the present invention, and the average width value of the converted wire is used to obtain an image of a high value of the line wound portion in the image.

Fig. 6D is a flow chart showing the identification of the wire scar of the dental floss according to an embodiment of the present invention.

Fig. 7A is an image of a stain of a stain such as a stain on a dentate portion of a dental floss semi-finished product due to oil infiltration of a thimble of a mold according to an embodiment of the present invention.

Fig. 7B is a view showing the image of the stain of the dental floss in the embodiment of the present invention, and the stained portion is black.

Fig. 7C is a flow chart showing the identification of the dirt of the dental floss stick according to an embodiment of the present invention.

Fig. 8A is an image of the original missing line of the dental floss according to an embodiment of the present invention. Fig. 5A shows an image of the edge boundary of the dental floss in the missing image through the pixel growth.

FIG. 8B is an image of a result of an image of a dental floss stick undergoing pixel growth and inverse binarization according to an embodiment of the present invention.

FIG. 8C is an image of the result of the pixel rod growing and inverse binarization processing and subtracting the image from the complete image according to an embodiment of the present invention.

Fig. 9 is a view showing an image of a thin wire scar of an enhanced floss stick according to an embodiment of the present invention.

Fig. 10 is a view showing an image for enhancing the stain of the dental floss stick according to an embodiment of the present invention.

Fig. 11 is a view showing an image of a missing material at the end of the dental floss stick according to an embodiment of the present invention.

10. . . Automatic optical inspection device

60. . . a whole piece of dental floss

100. . . Conveyor unit

200. . . control unit

300. . . Image capture unit

400. . . Information processing unit

500. . . Structural support platform

Claims (13)

An automatic optical detecting device for identifying a defect of a dental floss bar, the automatic optical detecting device comprising: a conveying unit carrying the automatic optical detecting device for transporting in a conveying direction; and a control unit responsible for starting and stopping the conveying unit And a pause control; an image capture unit for capturing a wide range of images of a whole floss bar and a small range of images for individual floss bars, wherein the image capture unit comprises a plurality of network cameras, the network One of the cameras is responsible for the capture of a wide range of images of the entire floss bar, and the remaining plurality of floss bars of the entire floss bar are taken in a plurality of side by side manners; an information processing unit for processing the The input of the image signal of the dental floss taken by the image capturing unit, the identification of the image and the movement control of the dental floss; and a structural support platform for assembling and carrying the conveying unit, the control unit, the image capturing unit and the information Processing unit, and transporting the dental floss stick. The automatic optical detecting device of claim 1, wherein the conveying unit comprises a motor mounted on the structural support platform and a pulley module connected to the motor, the motor is a variable frequency motor for Adjust the speed within the distance range. The automatic optical detecting device of claim 1, wherein the control unit comprises a plurality of proximity switches mounted on the structural support platform for providing a non-contact signal as a basis for positioning the floss bar, and the connection An input/output control card of the transport unit and the information processing unit. The automatic optical detecting device of claim 1, wherein the information processing unit further comprises a startup determining application software program to determine whether the entire dental floss bar is completely present in the range of the wide range of images, and an input application. The software program is responsible for inputting the image signal of the whole piece of dental floss. The automatic optical detecting device of claim 1, wherein the information processing unit further comprises an image processing software identification library for identifying the image signal of the dental floss, wire wound and dirty flaw. The automatic optical detecting device of claim 1, wherein the structural support platform further comprises a guiding plate, the guiding plate limits the degree of freedom of the dental floss bar, and ensures that the feeding of the whole dental floss bar can be The detection device detects it. An automatic optical detection method for identifying a defect of a dental floss comprises the steps of: taking a standard floss stick by an image taking unit to obtain a standard image as a basis for determining the defect of the floss bar; The image taking unit captures whether a whole piece of floss bar completely enters a range of images including one of the whole floss bars; the whole floss bar is transported by a transport unit, and is transported in a transport direction; The unit controls the conveying unit to pause at more than one image capturing position; And acquiring, by the image capturing unit, all the to-be-matched floss bars of the whole piece of the floss bar at the one or more image capturing positions to obtain a small range image of all the to-be-matched floss bars; The processing unit processes the large-scale image of the whole piece of dental floss to identify the missing defect of the whole piece of the dental floss, and then inputs the small-range image to identify the image of the dental floss to be compared, and determines the waiting The line damage and dirt of the floss bar are compared; and for the particular type, the test value is compared to its standard value to obtain the complete data of the file. The automatic optical detecting method according to claim 8, wherein the material shortage is that the flow material of the dental floss semi-finished product in the injection molding stage cannot completely fill the cavity, and the defect is formed; the wire injury includes the break of the wire used. , slack, and unevenness of thickness; and the contamination is caused by the oozing of the thimble of the mold, which causes the stain on the ejector portion of the semi-finished product of the dental floss. The automatic optical detection method of claim 8, wherein the information processing unit identifies the missing material of the entire dental floss image by identifying an image recognition library of an image processing software and identifies the to-be-matched This line of floss image damages the dirt. The automatic optical detection method of claim 8, wherein the information processing unit identifies the image of the missing piece of the whole piece of the dental floss comprises the following steps: using the identification module of the identification library according to the preset One of the standard floss stick images finds the same each to be compared to the floss stick Position: the standard floss bar is subtracted from the to-be-matched floss bar; and the missing module of the floss bar to be compared is obtained by using the analysis module of the identification library. The automatic optical detection method according to claim 8, wherein the information processing unit identifies the image of the line flaw to be compared with the floss bar, and further comprises the following steps: an edge detection module of the identification library Detecting an image of the peripheral edge portion of the wire rod of the floss bar to be obtained, and obtaining an area occupied by the wire rod of the floss bar to be compared; the binarization module of the identification function library is The image of the wire area is binarized, and the gray value is counted to obtain an area of the wire of the to-be-matched floss bar in the binarized image; the area of the wire is divided by the wire The length is obtained by obtaining the width of the wire to be compared; and the line scar data is obtained by comparing the width of the previously stored standard wire with the width of the wire to be compared. The automatic optical detection method of claim 8, wherein the information processing unit further comprises the following steps: identifying the image of the dirt to be compared with the dental floss, and the step of comparing the teeth in a light-emitting environment. The bar is taken to obtain a preliminary image of a black stain; the binarization module of the identification library is aimed at the black stain The step image is binarized; and the analysis module of the identification library is used to obtain all the analysis parameters in the image, and the analysis parameters formed by the stain are obtained by subtracting the standard analysis parameters stored in advance. The automatic optical detection method of claim 11, wherein the information processing unit further comprises the following steps for avoiding the residual image caused by the image subtraction at the edge of the missing image: increasing by the pixel of the identification library The module enlarges the boundary of the edge of the missing floss in the missing image; and uses the inverse binarization module of the identification library to exchange black and white missing images; using the identification library The image dividing module divides the reverse binarized missing image by the pre-stored standard complete image, and obtains a complete image of the gray level value of the missing portion to maintain the original value; Subtracting the image of the missing material after the inverse binarization, obtaining a grayscale of a missing portion lower than the surrounding image; and using the analysis module of the identification library to obtain the missing portion By analyzing the values, you can further judge, classify and store the defects.