Visual detection method and system for aperture of leakage nozzle

Abstract

The invention provides a visual detection method and system for the aperture of a leakage nozzle, and relates to the technical field of detection, and the method comprises the steps: S1, obtaining a to-be-detected glass fiber bushing leakage nozzle; s2, photographing the leakage nozzle of the glass fiber bushing plate to obtain image data of the leakage nozzle of the glass fiber bushing plate; s3, transmitting the image data to an upper computer, and calculating the coaxiality i, the inclination angle theta i and the hole inner diameter phi i of the leakage nozzle by the upper computer; and S4, comparing the coaxiality i with the coaxiality allowable error e, and comparing the inclination angle theta i with the perpendicularity allowable error theta e, and determining whether the hole inner diameter phi i of the leakage nozzle of the glass fiber leakage plate meets the precision requirement of the hole diameter in the ith leakage nozzle or not according to comparison results. According to the method, the inner aperture values of the leakage nozzles on the finished leakage plate are output in batches through an image processing algorithm, the time for manually reading and recording data is effectively shortened, the method for judging the coaxiality and the inclination of the leakage nozzles and measuring the aperture of the inclined leakage nozzles is provided, the detection efficiency is improved, and misjudgment and misjudgment caused by the coaxiality and the inclination are effectively avoided.

Description

Technical Field

[0001] The invention relates to the field of detection technology, and in particular to a visual detection method and system for a leakage nozzle aperture. Background Art

[0002] The grooving plate is one of the main devices in glass fiber production, and is shaped like a trough container. During the drawing process, molten glass flows into the grooving plate, which modulates it to a suitable temperature, then flows out through the grooving nozzle on the bottom plate, and is stretched into continuous glass fiber by the high-speed rotating drawing machine at the outlet; the mainstream glass fiber grooving plate models on the market currently have specifications between 200 and 6000 holes, and the requirements for the size of the grooving nozzle, especially the aperture of the grooving nozzle, are very strict.

[0003] At present, the detection of the aperture of the glass fiber leakage plate nozzle mainly adopts visual inspection and go / no-go gauge sampling inspection, but the above methods are time-consuming and laborious, and cannot accurately determine whether the concentricity of the nozzle is qualified, nor can they give the accurate size of the nozzle aperture. In addition, sampling inspection has certain risks and limitations, and has gradually failed to meet the detection needs of glass fiber leakage plate nozzles. In the process of welding the nozzle to the bottom plate of the leakage plate, a certain degree of deformation will occur due to the characteristics of the material itself. Therefore, the detection value of the nozzle aperture on the final finished leakage plate can represent its true value, which also makes some detection methods based on optical images inapplicable;

[0004] In addition, since the nozzle cannot be guaranteed to be welded to the base plate at a completely 90° vertical angle, there may be a certain inclination angle, causing some existing visual inspection equipment to have a large deviation in the detection of the actual aperture. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a visual inspection method and system for the aperture of a drain nozzle to solve the technical problem of quickly and accurately measuring the coaxiality and aperture of the drain nozzle on a finished drain plate in the prior art. A method for determining the coaxiality and inclination of the drain nozzle and measuring the aperture of the inclined drain nozzle is provided, which greatly improves the inspection efficiency and effectively avoids misjudgment and misjudgment caused by coaxiality and inclination.

[0006] The technical solution adopted by the present invention is:

[0007] A visual detection method for a nozzle aperture, comprising:

[0008] Step S1, obtaining the glass fiber plate nozzle to be tested;

[0009] Step S2, taking a picture of the fiberglass plate nozzle using an area array camera to obtain image data of the fiberglass plate nozzle;

[0010] Step S3: The image data is transmitted to the host computer through the data transmission module, and the host computer calculates the coaxiality of the nozzle through the image processing algorithm. i , tilt angle θ i and hole inner diameter

[0011] Step S4: adjust the coaxiality of the drain nozzle i Coaxiality tolerance◎ e、 Tilt angle θ i Perpendicularity tolerance θ e Compare and determine the inner diameter of the hole of the glass fiber bushing nozzle according to the comparison result Whether the inner aperture accuracy requirements of the i-th leak nozzle are met.

[0012] Preferably, in the visual detection method of the nozzle aperture, the step S2 of controlling the area array camera to take a picture of the fiberglass nozzle comprises:

[0013] Step S21. Place a small block with a high-precision known length next to the fiberglass plate nozzle to be tested, and determine the photographing position of the area array camera when taking pictures according to the positions of the small block and the fiberglass plate nozzle to be tested;

[0014] Step S22: Move the area array camera to the photographing position, and then use the area array camera to take photos.

[0015] Preferably, in the visual inspection method for the aperture of the drain nozzle, there are three photographing positions in step S22, which are the top of the drain nozzle, the bottom of the outer diameter of the drain nozzle and the bottom of the inner diameter of the drain nozzle, and in the process of the area array camera taking pictures of the product, the area array camera focuses on each of the photographing positions in turn, and takes pictures of the product at each of the photographing positions respectively, to obtain a first high-definition photo, a second high-definition photo and a third high-definition photo in turn.

[0016] Preferably, in the visual detection method of the nozzle aperture, the coaxiality in step S3 is i Obtained through the following methods:

[0017] Step S31. Extract the outer diameter contour and inner diameter contour of the top of the drain nozzle in the first high-definition photo through the contour extraction algorithm, determine its center through the minimum circumscribed circle algorithm, and calculate the coaxiality of the i-th drain nozzle. i .

[0018] Preferably, in the visual detection method of the nozzle aperture, the inclination angle θ in step S3 is i Obtained through the following methods:

[0019] Step S32. The first high-definition photo and the second high-definition photo are multi-focused by an image fusion algorithm to obtain a fourth high-definition photo of the top of the drain nozzle and the bottom of the outer diameter;

[0020] Step S33: extract the outer diameter contour of the top of the nozzle and the outer diameter contour of the bottom of the nozzle in the fourth high-definition photo by using the contour extraction algorithm, and calculate the center coordinates (x i1 ,y i1 ) and the center coordinates of the bottom outer diameter contour (x i2 ,y i2 ), calculate the distance l between the two centers based on the center coordinates of the top outer diameter contour and the bottom outer diameter contour i， According to l i Calculate the tilt angle θ i。

[0021] Preferably, in the visual detection method of the nozzle aperture, the distance l between the two centers of the circle in step S33 is i The calculation formula is as follows:

[0022]

[0023] Preferably, in the visual detection method of the nozzle aperture, the inclination angle θ in step S33 is i The calculation formula is as follows:

[0024] Where h is the length of the nozzle above the bottom plate, l i is the distance between the centers of the two circles.

[0025] Preferably, in the visual detection method of the nozzle aperture, the inner diameter of the hole in step S3 is Obtained by:

[0026] Step S34. According to the measurement accuracy requirement of the inner aperture of the drain nozzle, set the verticality threshold θ, θ<θ e , where θ e is the allowable error of verticality;

[0027] When θ i ≥θ, the image of the i-th nozzle is corrected by the first high-definition photo and the length value of a small block with high-precision known length is used as a reference. The front view image of the inner diameter of the top of the i-th nozzle is obtained by the contour extraction algorithm and the minimum circumscribed circle algorithm. The inner diameter of the nozzle is calculated.

[0028] When θ i<θ, according to the high-definition photo 3, the contour extraction algorithm and the minimum circumscribed circle algorithm are used to obtain the contour of the inner diameter of the bottom of the i-th drain nozzle, and the inner diameter of the drain nozzle is calculated.

[0029] Preferably, in the visual inspection method for the aperture of the leakage nozzle, the judgment rule for the leakage nozzle that does not meet the requirements in step S4 is as follows:

[0030] Step S41. Compare i Coaxiality tolerance◎ e , if◎ i >◎ e , record the number i of the leak nozzle that does not meet the requirements;

[0031] Step S42. Compare θ i Perpendicularity tolerance θ e , if θ i >θ e , record the number i of the leak nozzle that does not meet the requirements;

[0032] Step S43. When the inner aperture accuracy requirement of the i-th nozzle is not met, the nozzle number i that does not meet the requirement is recorded.

[0033] The present invention also provides a visual detection system for the aperture of a leak nozzle, which comprises:

[0034] Place the platform;

[0035] A fixed bracket, wherein the fixed bracket is arranged on the placement platform;

[0036] An area array camera, wherein the area array camera is arranged on a fixed bracket, the fixed bracket is used to drive the area array camera to move, and the area array camera is used to take pictures of the glass fiber leaking plate leaking nozzle to be detected;

[0037] A data transmission module, used for transmitting the image data of the glass fiber leak plate leak nozzle obtained by the area array camera to the host computer;

[0038] The host computer is used to calculate the coaxiality of the nozzle according to the image processing algorithm of the glass fiber nozzle plate. i , tilt angle θ i and hole inner diameter And adjust the coaxiality of the nozzle◎ i Coaxiality tolerance◎ e、 Tilt angle θ i Perpendicularity tolerance θ e Make a comparison;

[0039] The bottom plate is placed on the placement platform, and a plurality of glass fiber leaking plate leaking nozzles to be detected are fixed on the bottom plate.

[0040] Advantages of the present invention:

[0041] The visual inspection method for the aperture of the leak nozzle of the present invention has the advantages of simple operation, high efficiency and high inspection accuracy; the method outputs the inner aperture values ​​of the leak nozzles on the finished leak plate in batches through an image processing algorithm, effectively reducing the time for manual reading and recording of data; the numbers and reasons of the leak nozzles that do not meet the requirements are given, which is convenient for subsequent production personnel to handle; a method for determining the coaxiality and inclination of the leak nozzle and measuring the aperture of the inclined leak nozzle is given, which greatly improves the inspection efficiency and effectively avoids misjudgment and misjudgment caused by coaxiality and inclination. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a structural schematic diagram of the glass fiber bushing nozzle of the present invention.

[0043] Figure 2 It is a structural schematic diagram of the visual detection system of the nozzle aperture of the present invention.

[0044] Figure 3 The first high-definition photo shows the coaxiality detection effect of the fiberglass nozzle.

[0045] Figure 4 The third high-definition photo shows the coaxiality detection effect of the glass fiber nozzle.

[0046] Figure 5 Output the glass fiber leak detection structure diagram to the host computer. DETAILED DESCRIPTION

[0047] The present invention will be further described below in conjunction with specific embodiments.

[0048] Example 1

[0049] A visual detection method for a nozzle aperture, comprising:

[0050] Step S1, obtaining the glass fiber plate nozzle to be tested, the structure of the glass fiber plate nozzle is as follows: Figure 1 ;

[0051] Step S2, taking a picture of the fiberglass plate nozzle using an area array camera to obtain image data of the fiberglass plate nozzle;

[0052] Step S3: The image data is transmitted to the host computer through the data transmission module, and the host computer calculates the coaxiality of the nozzle through the image processing algorithm. i , tilt angle θ i and hole inner diameter

[0053] Step S4: adjust the coaxiality of the drain nozzle i Coaxiality tolerance◎ e、 Tilt angle θ iPerpendicularity tolerance θ e Compare and determine the inner diameter of the hole of the glass fiber bushing nozzle according to the comparison result Whether the inner diameter accuracy requirement of the i-th nozzle is met, when When the i-th nozzle inner aperture accuracy requirement is not met, the nozzle number i that does not meet the requirement is recorded, and the host computer outputs its location and indicates the reason for not meeting the requirement;

[0054] In step S2, controlling the area array camera to take a picture of the fiberglass plate nozzle includes:

[0055] Step S21. Place a small block with a high-precision known length next to the fiberglass plate nozzle to be tested, and determine the photographing position of the area array camera when taking pictures according to the positions of the small block and the fiberglass plate nozzle to be tested;

[0056] Step S22. Move the area array camera to the photographing position, and then use the area array camera to take photos. There are three photographing positions, which are the top of the nozzle, the bottom of the outer diameter of the nozzle, and the bottom of the inner diameter of the nozzle. During the process of the area array camera taking photos of the product, the area array camera focuses on each of the photographing positions in turn, and takes photos of the product at each of the photographing positions, respectively, to obtain a first high-definition photo, a second high-definition photo, and a third high-definition photo in turn.

[0057] Coaxiality in step S3◎ i Obtained through the following methods:

[0058] Step S31. Extract the outer diameter contour and inner diameter contour of the top of the drain nozzle in the first high-definition photo through the contour extraction algorithm, determine its center through the minimum circumscribed circle algorithm, and calculate the coaxiality of the i-th drain nozzle. i .

[0059] Among them, the contour extraction algorithm can be selected from gradient-based algorithm, Laplace-based algorithm, and deep learning-based algorithm according to actual conditions.

[0060] In step S3, the tilt angle θ i Obtained through the following methods:

[0061] Step S32. The first high-definition photo and the second high-definition photo are multi-focused by an image fusion algorithm to obtain a fourth high-definition photo of the top of the drain nozzle and the bottom of the outer diameter;

[0062] The image fusion algorithm can be selected based on the spatial domain, the change domain, the region segmentation and the deep learning according to the actual situation.

[0063] Step S33: extract the outer diameter contour of the top of the nozzle and the outer diameter contour of the bottom of the nozzle in the fourth high-definition photo by using the contour extraction algorithm, and calculate the center coordinates (x i1 ,y i1 ) and the center coordinates of the bottom outer diameter contour (x i2 ,y i2 ), l i is the distance between the two centers, h is the length of the drain nozzle above the bottom plate, and the inclination angle

[0064] Inner diameter of the hole in step S3 Obtained by:

[0065] Step S34. According to the measurement accuracy requirement of the inner aperture of the drain nozzle, set the verticality threshold θ, θ<θ e , where θ e is the allowable error of verticality;

[0066] When θ i ≥θ, the image of the i-th nozzle is corrected by the first high-definition photo and the length value of a small block with high-precision known length is used as a reference. The front view image of the inner diameter of the top of the i-th nozzle is obtained by the contour extraction algorithm and the minimum circumscribed circle algorithm. The inner diameter of the nozzle is calculated.

[0067] Among them, the image geometry correction algorithm can be selected based on the spatial domain, the frequency domain, or the optimization according to the actual situation.

[0068] When θ i <θ, according to the high-definition photo 3, the contour extraction algorithm and the minimum circumscribed circle algorithm are used to obtain the contour of the inner diameter of the bottom of the i-th drain nozzle, and the inner diameter of the drain nozzle is calculated.

[0069] The judgment rules for leak nozzles that do not meet the requirements in step S4 are as follows:

[0070] Step S41. Compare i Coaxiality tolerance◎ e , if◎ i >◎ e , record the number i of the leak nozzle that does not meet the requirements;

[0071] Step S42. Compare θ i Perpendicularity tolerance θ e , if θ i >θ e, record the number i of the leak nozzle that does not meet the requirements;

[0072] Step S43. When the inner aperture accuracy requirement of the i-th nozzle is not met, the nozzle number i that does not meet the requirement is recorded;

[0073] Figure 3 The inner and outer diameter contours of the drain nozzle top can be extracted with high precision, and the coaxiality can be calculated; Figure 4 The inner diameter profile of the bottom of the drain nozzle can be extracted with high precision; Figure 5 It can realize batch detection of glass fiber leaks and give the location of leaks that do not meet the requirements. The unqualified leaks are marked in red.

[0074] Example 2

[0075] like Figure 2 , a visual inspection system for a leak nozzle aperture, comprising:

[0076] Place platform 3;

[0077] A fixed bracket 2, wherein the fixed bracket 2 is arranged on a placement platform 3;

[0078] An area array camera 1, wherein the area array camera 1 is disposed on a fixed bracket 2, and the fixed bracket 2 is used to drive the area array camera 1 to move, and the area array camera 1 is used to take pictures of the glass fiber plate nozzle to be detected;

[0079] A data transmission module, used for transmitting the image data of the glass fiber plate nozzle obtained by the area array camera 1 to the host computer 5;

[0080] Host computer 5 is used to calculate the coaxiality of the nozzle of the glass fiber plate according to the image processing algorithm. i , tilt angle θ i and hole inner diameter And adjust the coaxiality of the nozzle◎ i Coaxiality tolerance◎ e、 Tilt angle θ i Perpendicularity tolerance θ e Make a comparison.

[0081] The bottom plate 4 is placed on the placement platform 3, and a plurality of glass fiber leak plate leak nozzles to be tested are welded on the bottom plate 4.

[0082] Finally, it should be noted that the above specific implementation methods are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A visual detection method for the aperture of a nozzle, characterized in that: include: Step S1, obtaining the glass fiber plate nozzle to be tested; Step S2, taking a picture of the fiberglass plate nozzle using an area array camera to obtain image data of the fiberglass plate nozzle; Step S3: The image data is transmitted to the host computer through the data transmission module, and the host computer calculates the coaxiality of the nozzle through the image processing algorithm. i , tilt angle θ i and hole inner diameter Step S4: adjust the coaxiality of the drain nozzle i Coaxiality tolerance◎ e、 Tilt angle θ i Perpendicularity tolerance θ e Compare and determine the inner diameter of the hole of the glass fiber bushing nozzle according to the comparison result Whether the inner aperture accuracy requirements of the i-th leak nozzle are met.

2. The visual detection method of the nozzle aperture according to claim 1 is characterized in that: In step S2, controlling the area array camera to take a picture of the fiberglass plate nozzle includes: Step S21. Place a small block with a high-precision known length next to the fiberglass plate nozzle to be tested, and determine the photographing position of the area array camera when taking pictures according to the positions of the small block and the fiberglass plate nozzle to be tested; Step S22: Move the area array camera to the photographing position, and then use the area array camera to take photos.

3. The visual detection method of the nozzle aperture according to claim 2, characterized in that: There are three photographing positions in step S22, which are the top of the nozzle, the bottom of the outer diameter of the nozzle, and the bottom of the inner diameter of the nozzle, respectively. During the process of the area array camera taking pictures of the product, the area array camera focuses on each of the photographing positions in turn, and takes pictures of the product at each of the photographing positions, respectively, to obtain a first high-definition photo, a second high-definition photo, and a third high-definition photo, respectively.

4. The visual detection method of the nozzle aperture according to claim 3 is characterized in that: Coaxiality in step S3◎ i Obtained through the following methods: Step S31. Extract the outer diameter contour and inner diameter contour of the top of the drain nozzle in the first high-definition photo through the contour extraction algorithm, determine its center through the minimum circumscribed circle algorithm, and calculate the coaxiality of the i-th drain nozzle. i .

5. The visual detection method of the nozzle aperture according to claim 3, characterized in that: In step S3, the tilt angle θ i Obtained through the following methods: Step S32. The first high-definition photo and the second high-definition photo are multi-focused by an image fusion algorithm to obtain a fourth high-definition photo of the top of the drain nozzle and the bottom of the outer diameter; Step S33: extract the outer diameter contour of the top of the nozzle and the outer diameter contour of the bottom of the nozzle in the fourth high-definition photo by using the contour extraction algorithm, and calculate the center coordinates (x i1 ,y i1 ) and the center coordinates of the bottom outer diameter contour (x i2 ,y i2 ), calculate the distance l between the two centers based on the center coordinates of the top outer diameter contour and the bottom outer diameter contour i， According to l i Calculate the tilt angle θ i。 6. The visual detection method of the nozzle aperture according to claim 5, characterized in that: The distance l between the two centers of the circle in step S33 i The calculation formula is as follows:

7. The visual detection method of the nozzle aperture according to claim 6, characterized in that: In step S33, the tilt angle θ i The calculation formula is as follows: Where h is the length of the nozzle above the bottom plate, l i is the distance between the centers of the two circles.

8. The visual detection method of the nozzle aperture according to claim 3, characterized in that: Inner diameter of the hole in step S3 Obtained by: Step S34. According to the measurement accuracy requirement of the inner aperture of the drain nozzle, set the verticality threshold θ, θ<θ e , where θ e is the allowable error of verticality; When θ i ≥θ, the image of the i-th nozzle is corrected by the first high-definition photo and the length value of a small block with high-precision known length is used as a reference. The front view image of the inner diameter of the top of the i-th nozzle is obtained by the contour extraction algorithm and the minimum circumscribed circle algorithm. The inner diameter of the nozzle is calculated. When θ i < θ, according to the high-definition photo 3, the contour of the inner diameter of the bottom of the i-th nozzle is obtained through the contour extraction algorithm and the minimum circumscribed circle algorithm, and the inner diameter of the nozzle is calculated 9. The visual detection method of the nozzle aperture according to claim 3, characterized in that: The judgment rules for leak nozzles that do not meet the requirements in step S4 are as follows: Step S41. Compare i Coaxiality tolerance◎ e , if◎ i >◎ e , record the number i of the leak nozzle that does not meet the requirements; Step S42. Compare θ i Perpendicularity tolerance θ e , if θ i >θ e , record the number i of the leak nozzle that does not meet the requirements; Step S43. When the inner aperture accuracy requirement of the i-th nozzle is not met, the nozzle number i that does not meet the requirement is recorded.

10. A visual inspection system for the aperture of a leak nozzle, characterized in that: include: Placement platform (3); A fixed support (2), wherein the fixed support (2) is arranged on a placement platform (3); An area array camera (1), the area array camera (1) being arranged on a fixed bracket (2), the fixed bracket (2) being used to drive the area array camera (1) to move, and the area array camera (1) being used to take a picture of the glass fiber leaking plate leaking nozzle to be detected; A data transmission module, used for transmitting image data of the glass fiber leak plate leak nozzle obtained by the area array camera (1) to a host computer (5); The host computer (5) is used to calculate the coaxiality of the nozzle according to the image processing algorithm of the glass fiber nozzle plate nozzle. i , tilt angle θ i and hole inner diameter And adjust the coaxiality of the nozzle◎ i Coaxiality tolerance◎ e、 Tilt angle θ i Perpendicularity tolerance θ e Make a comparison; A bottom plate (4) is placed on the placement platform (3), and a plurality of glass fiber leak plate leak nozzles to be tested are fixed on the bottom plate (4).

Images