A production quality detection method and system for insulating glass

Abstract

The present invention relates to the field of transparent glass detection technology, and specifically to a production quality detection method and system for insulating glass, including: obtaining an insulating glass dew point detection image; analyzing the grayscale concentration distribution characteristics of frosting in the insulating glass dew point detection image, and screening out glass frosting detection images from all insulating glass dew point detection images; analyzing the edge extension of the same image position area between adjacent glass frosting detection images, and dividing the frosting amplified texture; analyzing the single situation of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images, and obtaining the crack texture diffusion degree; according to the crack texture diffusion degree, combined with analyzing the interference of the frosting phenomenon between adjacent frosting detection images on the expression of fine crack information, obtaining the real crack degree, and adaptively identifying the fine cracks that coincide with the frosting information. The present invention improves the credibility of the air tightness detection results of the inner layer of the insulating glass.

Description

Technical Field

[0001] The present invention relates to the technical field of transparent glass detection, and in particular to a production quality detection method and system for insulating glass. Background Art

[0002] Insulating glass mainly refers to a sealed glass product formed by injecting a certain concentration of inert gas into the space between multiple pieces of glass separated by metal, and then bonding and sealing with a high-viscosity, high-airtightness adhesive. Compared with other single-layer glass, insulating glass mainly relies on the stable and non-reactive characteristics of the inert gas in the enclosed space, which makes insulating glass have better heat insulation and sound insulation capabilities and greatly extends the service life of the glass. Therefore, when conducting production quality inspections on insulating glass, conventional human eye inspections cannot effectively observe subtle condensation or frost, so digital imaging technology is usually used to detect the airtightness of the space formed by multiple pieces of glass.

[0003] The prior art usually uses a dew point meter to detect the dew point of insulating glass. The air tightness of the inner layer of the insulating glass is determined by lowering the temperature of the inner layer of the insulating glass within a standard temperature range to observe whether there is dew or frost. However, if there are tiny cracks on the surface of the insulating glass that cannot be observed visually, the glass surface in contact with the dew point meter during the dew point detection may cause the cracks to increase due to the low temperature of the contact surface, causing the frost and cracks generated in the inner layer of the glass to partially overlap in the image information, making it impossible to effectively distinguish the cracks, thereby reducing the credibility of the air tightness test results of the inner layer of the insulating glass and reducing the detection accuracy of the production quality of the insulating glass. Summary of the invention

[0004] The present invention provides a production quality detection method and system for insulating glass to solve the existing problem that during dew point detection, cracks on the glass surface in contact with a dew point meter may increase due to the low temperature of the contact surface, causing frost on the inner layer of the glass to partially overlap with the cracks in the image information, making it impossible to effectively distinguish the cracks.

[0005] The present invention provides a production quality inspection method and system for insulating glass using the following technical solutions:

[0006] The present invention proposes a production quality detection method for insulating glass, which comprises the following steps:

[0007] Obtain several insulating glass dew point detection images;

[0008] Analyze the grayscale distribution characteristics of frosting in the dew point detection map of insulating glass, and select several glass frosting detection maps from all the dew point detection maps of insulating glass;

[0009] The edge extension of the same image position area between adjacent glass frosting detection images is analyzed, and several frosting amplified textures are divided from the glass frosting detection images; the single situation of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images is analyzed, and the crack texture diffusion degree of each frosting amplified texture is obtained by combining the smoothness of the diffusion process;

[0010] According to the crack texture diffusion, combined with the analysis of the interference of frosting between adjacent frosting detection images on the expression of small crack information, the real crack degree of each frosting amplified texture is obtained;

[0011] According to the actual crack degree, small cracks that coincide with frosting information are adaptively identified and quality inspection is performed.

[0012] Preferably, the method for acquiring the glass frosting detection image is:

[0013] For any insulating glass dew point detection image, the aggregation distribution of grayscale information in the insulating glass dew point detection image is analyzed to obtain the frosting area of ​​the insulating glass dew point detection image; according to the overall content of grayscale information contained in the frosting area, several glass frosting detection images are screened out from all insulating glass dew point detection images.

[0014] Preferably, the method for obtaining the frosting augmented texture is:

[0015] For the first glass frosting detection image, the edge connectivity within the glass frosting detection image is analyzed to obtain several edge coordinate sequences within the glass frosting detection image; the extension of the image position occupied by the edge coordinate sequence in the surrounding area between subsequent adjacent glass frosting detection images is analyzed, and several frosting amplified textures are divided from the glass frosting detection image.

[0016] Preferably, the edge coordinate sequence is obtained by:

[0017] A plurality of edge pixel points in the glass frosting detection image are obtained, image coordinate information of the edge pixel points that are continuously distributed in the glass frosting detection image is analyzed, and a plurality of edge coordinate sequences in the glass frosting detection image are obtained.

[0018] Preferably, the method for obtaining the crack texture diffusion is:

[0019] For any frosting amplified texture in any glass frosting detection image, the prominence of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images is analyzed to obtain the process extension direction independence of the frosting amplified texture; based on the process extension direction independence and combined with the analysis of the smoothness of the diffusion process, the crack texture diffusion of the frosting amplified texture is obtained.

[0020] Preferably, the method for obtaining the process extension direction independence is:

[0021] Several suspected extension directions of the frosting augmented texture between adjacent glass frosting detection images are obtained; the length direction differences of the edge texture distribution between different suspected extension directions are compared to obtain the process extension direction independence of the frosting augmented texture.

[0022] Preferably, the method for obtaining the true crack degree is:

[0023] For any glass frosting detection image, according to the frosting amplified texture contained in the glass frosting detection image, the frosting reaction trend of the frosting phenomenon between the glass frosting detection image and the subsequent adjacent glass frosting detection images is analyzed to obtain the frosting dew point detection interference degree of the glass frosting detection image;

[0024] For any frosting amplified texture in the glass frosting detection image, the crack texture diffusion degree is combined with the frosting dew point detection interference degree to obtain the actual crack degree of the frosting amplified texture.

[0025] Preferably, the method for obtaining the frosting dew point detection interference degree is:

[0026] By comparing the growth trend of the frosted area between the glass frosting detection map and the subsequent adjacent glass frosting detection maps, the frosting dew point detection interference degree of the glass frosting detection map is obtained.

[0027] Preferably, after calculating the true crack extent, the method further comprises:

[0028] The actual crack extent is normalized.

[0029] The present invention also proposes a production quality detection system for insulating glass, comprising a memory and a processor, wherein the processor executes a computer program stored in the memory to implement the steps of the above-mentioned production quality detection method for insulating glass.

[0030] The beneficial effects of the technical solution of the present invention are: analyzing the edge extension of the same image position area between adjacent glass frost detection images, and dividing a number of frost amplified textures from the glass frost detection image; wherein the frost amplified texture reflects the change form of the edge texture of the fixed image position in the overall process of dew point detection temperature, and can better observe the edge extension characteristics of the same image position in the overall process of dew point detection temperature; analyzing the single situation of the main diffusion direction of the frost amplified texture between adjacent glass frost detection images, and combining the smoothness of the diffusion process, to obtain the crack texture diffusion degree of the frost amplified texture; wherein the crack texture diffusion degree reflects the tendency of the frost amplified texture to diffuse unidirectionally in the overall process of dew point detection temperature, which enhances the expression of small crack information. The method can obtain the real crack degree of the frosting amplified texture by analyzing the interference of the frosting phenomenon on the expression of small crack information between adjacent frosting detection images according to the crack texture diffusion degree. The real crack degree reflects the clarity of the small crack information when the frosting phenomenon interferes with the expression of small cracks, and reduces the interference of the frosting phenomenon on the frost texture. The present invention obtains the real crack degree by analyzing the extended monotonic features of the corresponding texture at the same image position during the dew point detection of the insulating glass, and adaptively identifies the small cracks that coincide with the frosting information according to the real crack degree to complete the quality detection. The small cracks that coincide with the frosting information can be effectively identified, thereby improving the credibility of the air tightness detection results of the inner layer of the insulating glass. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 A flowchart of the steps of a production quality inspection method for insulating glass according to the present invention;

[0033] Figure 2 It is a schematic diagram for comparing frosting areas of the present invention;

[0034] Figure 3 It is a schematic diagram of a crack that is not interconnected with the inner layer of the insulating glass according to the present invention. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the production quality detection method and system for insulating glass proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0037] The specific scheme of the production quality detection method and system for insulating glass provided by the present invention is described in detail below with reference to the accompanying drawings.

[0038] See also Figure 1 , which shows a flowchart of a method for detecting the production quality of insulating glass provided by an embodiment of the present invention, the method comprising the following steps:

[0039] Step S001: Obtain several insulating glass dew point detection images.

[0040] It should be noted that the prior art usually uses a dew point meter to detect the dew point of insulating glass. The air tightness of the inner layer of the insulating glass is determined by lowering the temperature of the inner layer of the insulating glass within the standard temperature range to observe whether there is dew or frost. However, if there are tiny cracks on the surface of the insulating glass that cannot be observed intuitively, the glass surface in contact with the dew point meter during the dew point detection may cause the cracks to increase due to the low temperature of the contact surface, causing the frost and cracks produced on the inner layer of the glass to partially overlap in the image information, making it impossible to effectively distinguish the cracks.

[0041] In a specific implementation of the embodiment of the present invention, a method for obtaining a dew point detection image of a hollow glass is as follows: an industrial camera is fixed above the measuring surface of a dew point meter; a temperature control module of the dew point meter is started; Start, increase every 4 minutes , and take a hollow glass inspection image until the temperature is adjusted to Stop shooting until the end, obtain all the insulating glass detection images; grayscale process each insulating glass detection image to obtain a grayscale image, and use it as the insulating glass dew point detection image. Each insulating glass dew point detection image corresponds to the temperature of a temperature control module.

[0042] It should be noted that the insulating glass dew point detection diagram in this example is arranged in descending order by the temperature of the temperature control module by default.

[0043] It is particularly noted that the graying process is a well-known technology and will not be described in detail in this embodiment; in addition, the dew point meter in this embodiment is described by taking a semi-conductive refrigeration type insulating glass dew point meter as an example.

[0044] So far, several insulating glass dew point detection diagrams have been obtained through the above method.

[0045] Step S002: Analyze the grayscale concentration distribution characteristics of the insulating glass dew point detection map when frosting occurs, and select a number of glass frosting detection maps from all the insulating glass dew point detection maps.

[0046] It should be noted that the inner space of insulating glass is usually filled with high concentration of argon gas. The thermal conductivity of argon gas is lower than that of air to improve the thermal insulation capacity of insulating glass. When the insulating glass is tested for dew point, the surface of the insulating glass in contact with the measuring surface is limited, so the argon gas in the inner layer of the glass in contact with the measuring surface is mainly subjected to low temperature detection. The local low temperature is not easy to diffuse to the argon gas in other areas. Therefore, if the inner layer of the glass is frosted, the corresponding image grayscale information will generally be concentrated. Therefore, the grayscale concentrated distribution characteristics of frosting in the insulating glass dew point detection image can be analyzed, and several glass frosting detection images can be selected from all insulating glass dew point detection images. Please refer to Figure 2 , which shows a schematic diagram of the comparison of frosting areas.

[0047] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the glass frosting detection map is: for any insulating glass dew point detection map, analyze the aggregate distribution of grayscale information in the insulating glass dew point detection map to obtain the frosting area of ​​the insulating glass dew point detection map; and select a number of glass frosting detection maps from all insulating glass dew point detection maps according to the overall content of grayscale information contained in the frosting area. The specific process is as follows:

[0048] Take any insulating glass dew point detection image as an example, the neighborhood radius Preset to the radius of the measuring surface, minimum number of samples Preset as measurement surface The area is rounded to the nearest integer. The distance metric is preset as the grayscale difference between different pixels. , minimum number of samples And distance measurement, DBSCAN clustering is performed on the insulating glass dew point detection image to obtain several clusters. The mean of the grayscale values ​​of all pixels in each cluster is taken as the grayscale mean of each cluster. , minimum number of samples The process of clustering by distance metric is well known in the DBSCAN clustering algorithm and will not be described in detail in this embodiment.

[0049] Furthermore, a grayscale mean threshold is preset , the gray mean is greater than The image area occupied by the clusters is taken as the frosted area; the hollow glass dew point detection image containing the frosted area is taken as the glass frosting detection image. This example is described as an example, and this embodiment is not specifically limited. It may depend on the specific implementation situation.

[0050] So far, several glass frosting detection images are obtained through the above method.

[0051] Step S003: Analyze the edge extension of the same image position area between adjacent glass frost detection images, and divide several frosting amplified textures from the glass frost detection images; analyze the single situation of the main diffusion direction of the frosting amplified texture between adjacent glass frost detection images, and combine the smoothness of the diffusion process to obtain the crack texture diffusion degree of each frosting amplified texture.

[0052] It should be noted that when conducting dew point detection, if there are small cracks that are difficult to observe on the glass surface in contact with the dew point meter temperature measuring platform; then in the process of obtaining the grayscale image of the insulating glass as the temperature changes continuously, these crack edges will appear in a fixed area on the glass surface, and with different temperature changes, there will be different differences in the heat inside and outside the middle glass, resulting in different pressure differences, causing cracks to expand to varying degrees; therefore, the edge extension of the same image position area between adjacent glass frost detection images can be analyzed, and several frost amplification textures can be divided from the glass frost detection image.

[0053] It should be further explained that, under normal circumstances, if there are small cracks on the glass surface, these cracks are usually scratches in the glass manufacturing process, and they extend in a certain direction. The part extending in this direction is more fragile than other glass areas. Compared with it, the texture of frost is relatively chaotic and disordered. Therefore, the single situation of the main diffusion direction of the frost amplified texture between adjacent glass frost detection images can be analyzed, and combined with the smoothness of the diffusion process, the crack texture diffusion degree of each frost amplified texture can be obtained. Among them, if the crack texture diffusion degree is larger, it means that the extension direction of the frost amplified texture in the process of increasing the dew point detection temperature is more independent, and the interference from the edges extending in other directions is less, reflecting that the frost amplified texture is more likely to be the edge of the crack in the inner layer of the insulating glass.

[0054] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the frosting augmented texture is as follows: for the first glass frosting detection image, analyzing the edge connectivity in the glass frosting detection image, obtaining a number of edge coordinate sequences in the glass frosting detection image; analyzing the extension of the image position occupied by the edge coordinate sequence in the surrounding area between subsequent adjacent glass frosting detection images, and dividing a number of frosting augmented textures from the glass frosting detection image. The specific process is:

[0055] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the edge coordinate sequence is: obtaining a number of edge pixel points in the glass frosting detection image, analyzing the image coordinate information of the edge pixel points that are continuously distributed in the glass frosting detection image, and obtaining a number of edge coordinate sequences in the glass frosting detection image. The specific process is as follows:

[0056] Obtain all connected domains in each glass frosting detection image, and use each pixel on the edge of each connected domain as a connected edge pixel; take any connected domain in the first glass frosting detection image as an example, use the sequence of all connected edge pixels of the connected domain arranged in a clockwise direction as an edge sequence, and use the line segment formed by all connected edge pixels in the edge sequence as the initial frosting texture line; use the sequence formed by the coordinate position of each pixel in the edge sequence as an edge coordinate sequence; obtain several edge coordinate sequences. Each connected domain corresponds to an edge coordinate sequence. In addition, the process of obtaining a connected domain is a well-known technology and will not be repeated in this embodiment.

[0057] It is particularly noted that the first pixel point in the edge sequence takes the pixel point on the edge of the first connected domain in the horizontal left direction of the connected domain center as the starting point.

[0058] Furthermore, the remaining glass frost detection images except the first glass frost detection image are used as reference glass dew point detection images; taking any edge coordinate sequence in the first glass frost detection image as an example and the first reference glass dew point detection image as the starting point, if there are connected edge pixels in the first reference glass dew point detection image within the image position range of the eight-neighborhood of any coordinate position in the edge coordinate sequence, then these connected edge pixels are also incorporated into the initial frost texture line; if there are connected edge pixels in the second reference glass dew point detection image within the image position range of the eight-neighborhood of any coordinate position in the edge coordinate sequence, then these connected edge pixels are also incorporated into the initial frost texture line; if there are connected edge pixels in the third reference glass dew point detection image within the image position range of the eight-neighborhood of any coordinate position in the edge coordinate sequence, then these connected edge pixels are also incorporated into the initial frost texture line; and so on, until all reference glass dew point detection images are traversed and all initial frost texture lines are obtained.

[0059] Furthermore, the initial frosting texture line that is continuously incorporated into the connected edge pixel point process is used as the frosting augmented texture.

[0060] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the crack texture diffusion is: for any frosting amplified texture in any glass frosting detection image, analyze the prominence of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images to obtain the process extension direction independence of the frosting amplified texture; according to the process extension direction independence, combined with the analysis of the smoothness of the diffusion process, obtain the crack texture diffusion of the frosting amplified texture. The specific process is:

[0061] It should be noted that, under normal circumstances, compared with the chaotic and disordered texture change characteristics of frost texture, if there are fine cracks on the inner surface of insulating glass, although these fine cracks have certain texture extensions in multiple local directions, the overall texture extension in a specific direction will be highlighted. Therefore, the highlighting of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images can be analyzed to obtain the process extension direction independence of the frosting amplified texture. Among them, the greater the process extension direction independence, the smaller the effect of the texture extension in multiple local directions of the frosting amplified texture on the texture extension in the overall direction, reflecting that the frosting amplified texture has a stronger tendency to amplify along a specific direction during the dew point detection process.

[0062] Preferably, in some implementations of the present invention, the method for obtaining the process extension direction independence is: obtaining several suspected extension directions of the frosting augmented texture between adjacent glass frosting detection images; comparing the length direction difference of the edge texture distribution between different suspected extension directions to obtain the process extension direction independence of the frosting augmented texture. The specific process is as follows:

[0063] Taking any reference glass dew point detection image as an example, the initial frosting texture line of the frosting amplified texture in the reference glass dew point detection image is linearly fitted to obtain a fitted frosting line; the minimum angle between the fitted frosting line and the horizontal direction is taken as the suspected extension direction of the frosting amplified texture in the reference glass dew point detection image. Each frosting amplified texture corresponds to a suspected extension direction in each reference glass dew point detection image, and each suspected extension direction corresponds to a fitted frosting line.

[0064] Further, as an example, the process extension direction independence can be calculated by the following formula:

[0065]

[0066] In the formula, Indicates the process extension direction independence of the frosting augmented texture; Indicates the number of all reference glass dew point test graphs; Indicates that the frosting augmented texture is The suspected extension direction in the dew point detection image of the reference glass; Indicates that the frosting augmented texture is The suspected extension direction in the dew point detection image of the reference glass; Indicates that the frosting augmented texture is The length of the fitted frost straight line in the dew point detection chart of the reference glass; Indicates that the frosting augmented texture is The length of the fitted frost straight line in the dew point detection chart of the reference glass; Indicates taking the absolute value; Represents the normalization function.

[0067] It should be noted that the greater the independence of the process extension direction, the smaller the influence of the texture extension in multiple local directions of the frost amplified texture on the texture extension in the overall direction, which reflects that the tendency of the frost amplified texture to amplify along a specific direction during the dew point detection process is stronger.

[0068] Further, the initial frosting texture line of the frosting amplified texture in the control glass dew point detection diagram is fitted with a Bezier curve to obtain a fitted frosting curve; the mean curvature of all pixel points on the fitted frosting curve is used as a diffusion smoothing parameter; the product of the process extension direction independence of the frosting amplified texture and the diffusion smoothing parameter is used as the crack texture diffusion degree of the frosting amplified texture. The acquisition of the Bezier curve is a well-known technology and will not be repeated in this embodiment.

[0069] It should be noted that, if the crack texture diffusion degree is greater, it means that the extension direction of the frost amplification texture is more independent during the increase of the dew point detection temperature, and the interference from the edges extending in other directions is smaller, which reflects that the frost amplification texture is more likely to be the edge of a crack in the inner layer of the insulating glass.

[0070] At this point, the crack texture diffusion of each frosting amplified texture is obtained through the above method.

[0071] Step S004: According to the crack texture diffusion degree, combined with analyzing the interference of the frosting phenomenon between adjacent frosting detection images on the expression of small crack information, the real crack degree of each frosting amplified texture is obtained.

[0072] It should be noted that, in the process of dew point detection temperature changes, the inner layer of the insulating glass will also form frost of different thicknesses due to different heat differences, which will interfere with the crack information to varying degrees; therefore, the real crack degree of each frosted amplified texture can be obtained based on the crack texture diffusion degree and the interference of frosting between adjacent frosting detection images on the expression of fine crack information. The greater the real crack degree, the more the frosted amplified texture can eliminate the interference information of the frost texture, reflecting that the frosted amplified texture is more likely to be a fine crack that actually exists on the surface of the insulating glass.

[0073] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the real crack degree is: for any glass frosting detection image, according to the frosting amplified texture contained in the glass frosting detection image, the frosting reaction trend of the frosting phenomenon between the glass frosting detection image and the subsequent adjacent glass frosting detection images is analyzed to obtain the frosting dew point detection interference degree of the glass frosting detection image; for any frosting amplified texture in the glass frosting detection image, the crack texture diffusion degree is combined with the frosting dew point detection interference degree to obtain the real crack degree of the frosting amplified texture. The specific process is:

[0074] It should be noted that as the temperature in the temperature control module continues to rise during the dew point detection process, the smaller the heat difference pressure between the inner space and the outer space of the insulating glass, the smaller the force that promotes the frosting phenomenon, and the corresponding frosting thickness and area will be reduced; therefore, according to the frosting amplification texture contained in the glass frosting detection map, the frosting reaction trend of the frosting phenomenon between the glass frosting detection map and the subsequent adjacent glass frosting detection map can be analyzed to obtain the frosting dew point detection interference of the glass frosting detection map. Among them, if the frosting dew point detection interference is greater, it means that when the insulating glass is subjected to dew point detection, the frosting phenomenon changes more obviously, reflecting that there is a greater possibility that cracks exist on the surface of the insulating glass, causing the inner layer of argon gas to leak and reduce the thermal insulation capacity.

[0075] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the frosting dew point detection interference degree is: comparing the growth trend of the frosting area between the glass frosting detection diagram and the subsequent adjacent glass frosting detection diagram to obtain the frosting dew point detection interference degree of the glass frosting detection diagram. The specific process is as follows:

[0076] As an example, the interference degree of frost dew point detection can be calculated by the following formula:

[0077]

[0078] In the formula, Indicates Frost dew point detection interference degree of glass frosting detection diagram; Indicates The number of all glass frost detection images after the glass frost detection image; Indicates The area of ​​the frosted area in the glass frosting detection map; Indicates After the glass frost detection picture The area of ​​the frosted area in the glass frosting detection map; Indicates The mean gray value of all pixels in the frosted area of ​​the glass frosting detection image; Indicated in After the glass frost detection picture The average grayscale value of all pixels in the frosted area in the glass frosting detection image; Indicates taking the absolute value.

[0079] It should be noted that the greater the interference of frosting dew point detection, the more obvious the frosting phenomenon will be when the insulating glass is subjected to dew point detection, which means that there is a high possibility that cracks will exist on the surface of the insulating glass, causing the inner layer of argon gas to leak, resulting in reduced thermal insulation capacity.

[0080] Furthermore, a normalized value of the ratio of the crack texture diffusion degree of the frosting amplified texture to the frosting dew point detection interference degree of the glass frosting detection image is used as the actual crack degree of the frosting amplified texture.

[0081] It is particularly noted that in this embodiment The normalization process is performed by taking the function as an example, wherein the normalization function can be determined according to the specific implementation situation.

[0082] It should be noted that the greater the degree of the actual cracks, the more the frosting amplified texture can eliminate the interference information of the frost texture, which reflects that the frosting amplified texture is more likely to be the actual tiny cracks on the surface of the insulating glass.

[0083] At this point, the actual crack extent of each frosting-amplified texture is obtained through the above method.

[0084] Step S005: According to the actual crack degree, small cracks that coincide with the frosting information are adaptively identified and quality inspection is performed.

[0085] In a specific implementation of the embodiment of the present invention, the quality detection process is as follows:

[0086] Preset a real crack degree threshold , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. It may depend on the specific implementation situation.

[0087] 1. If there is no frost detection image, it means that the air tightness of the insulating glass meets the standard. If there are cracks on the surface, the cracks are not connected to the inner layer of the insulating glass; please refer to Figure 3 , which shows a schematic diagram of a crack that is not interconnected with the inner layer of the insulating glass.

[0088] 2. If When there is no frost detection image, if there is no real crack greater than The frosting amplified texture indicates that the air tightness of the insulating glass meets the standards.

[0089] 3. If When there is a frost detection image, if there is no real crack greater than The frosting amplified texture indicates that the air tightness of the insulating glass does not meet the standards, and the main reason is not caused by cracks on the surface of the insulating glass.

[0090] 4. If When there is a frost detection image, if there is a real crack greater than The frosting amplified texture indicates that the air tightness of the insulating glass does not meet the standards, and the main reason is the cracks on the surface of the insulating glass.

[0091] Through the above steps, a production quality inspection method for insulating glass is completed.

[0092] Another embodiment of the present invention provides a production quality inspection system for insulating glass, the system comprising a memory and a processor, and when the processor executes the computer program stored in the memory, the above method steps S001 to S005 are executed.

[0093] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A production quality inspection method for insulating glass, characterized in that: The method comprises the following steps: Obtain several insulating glass dew point detection images; Analyze the grayscale distribution characteristics of frosting in the dew point detection map of insulating glass, and select several glass frosting detection maps from all the dew point detection maps of insulating glass; The edge extension of the same image position area between adjacent glass frosting detection images is analyzed, and several frosting amplified textures are divided from the glass frosting detection images; the single situation of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images is analyzed, and the crack texture diffusion degree of each frosting amplified texture is obtained by combining the smoothness of the diffusion process; According to the crack texture diffusion degree, combined with the analysis of the interference of frosting phenomenon between adjacent glass frosting detection images on the expression of small crack information, the real crack degree of each frosting amplified texture is obtained; According to the actual crack degree, the system adaptively identifies small cracks that coincide with frosting information and performs quality inspection; The method for obtaining the frosting augmented texture is: For the first glass frosting detection image, edge connectivity in the glass frosting detection image is analyzed to obtain a number of edge coordinate sequences in the glass frosting detection image; the extension of the image position occupied by the edge coordinate sequence in the surrounding area between subsequent adjacent glass frosting detection images is analyzed to divide a number of frosting amplified textures from the glass frosting detection image; The method for obtaining the edge coordinate sequence is: Acquiring a number of edge pixel points in the glass frosting detection image, analyzing the image coordinate information of the edge pixel points that are continuously distributed in the glass frosting detection image, and acquiring a number of edge coordinate sequences in the glass frosting detection image includes: Obtain all connected domains in each glass frosting detection image, and use each pixel point on the edge of each connected domain as a connected edge pixel point; for any connected domain in the first glass frosting detection image, use a sequence of all connected edge pixel points of the connected domain arranged in a clockwise direction as an edge sequence, and use a line segment formed by all connected edge pixel points in the edge sequence as an initial frosting texture line; use a sequence formed by the coordinate position of each pixel point in the edge sequence as an edge coordinate sequence; obtain a number of edge coordinate sequences; wherein each connected domain corresponds to an edge coordinate sequence; Methods for obtaining frosting augmented textures include: The remaining glass frost detection images except the first glass frost detection image are used as reference glass dew point detection images; for any edge coordinate sequence in the first glass frost detection image, take the first reference glass dew point detection image as the starting point, if there are connected edge pixels in the first reference glass dew point detection image within the image position range of the eight neighborhoods of any coordinate position in the edge coordinate sequence, then these connected edge pixels are also incorporated into the initial frost texture line; if there are connected edge pixels in the second reference glass dew point detection image within the image position range of the eight neighborhoods of any coordinate position in the edge coordinate sequence, then these connected edge pixels are also incorporated into the initial frost texture line; if there are connected edge pixels in the third reference glass dew point detection image within the image position range of the eight neighborhoods of any coordinate position in the edge coordinate sequence, then these connected edge pixels are also incorporated into the initial frost texture line; and so on, until all reference glass dew point detection images are traversed to obtain all initial frost texture lines; The initial frosting texture line that is continuously merged into the connected edge pixel point process is used as the frosting augmented texture; The method for obtaining the crack texture diffusion is: For any frosting amplified texture in any glass frosting detection image, the prominence of the main diffusion direction of the frosting amplified texture between adjacent glass frosting detection images is analyzed to obtain the process extension direction independence of the frosting amplified texture; according to the process extension direction independence, combined with the analysis of the smoothness of the diffusion process, the crack texture diffusion of the frosting amplified texture is obtained, including: performing Bezier curve fitting on the initial frosting texture line of the frosting amplified texture in the control glass dew point detection image to obtain a fitted frosting curve; taking the mean curvature of all pixel points on the fitted frosting curve as a diffusion smoothing parameter; taking the product of the process extension direction independence of the frosting amplified texture and the diffusion smoothing parameter as the crack texture diffusion of the frosting amplified texture; The method for obtaining the independence of the process extension direction is: Obtain several suspected extension directions of the frosting augmented texture between adjacent glass frosting detection images; compare the length direction differences of the edge texture distribution between different suspected extension directions, and obtain the process extension direction independence of the frosting augmented texture, including: For any reference glass dew point detection image, a straight line fitting is performed on the initial frosting texture line of the frosting amplified texture in the reference glass dew point detection image to obtain a fitted frosting straight line; the minimum angle between the fitted frosting straight line and the horizontal direction is used as the suspected extension direction of the frosting amplified texture in the reference glass dew point detection image; wherein each frosting amplified texture corresponds to a suspected extension direction in each reference glass dew point detection image, and each suspected extension direction corresponds to a fitted frosting straight line; The process extension direction independence is calculated by the following formula: In the formula, Indicates the process extension direction independence of the frosting augmented texture; Indicates the number of all reference glass dew point test graphs; Indicates that the frosting augmented texture is The suspected extension direction in the dew point detection image of the reference glass; Indicates that the frosting augmented texture is The suspected extension direction in the dew point detection image of the reference glass; Indicates that the frosting augmented texture is The length of the fitted frost straight line in the dew point detection chart of the reference glass; Indicates that the frosting augmented texture is The length of the fitted frost straight line in the dew point detection chart of the reference glass; Indicates taking the absolute value; represents the normalization function; The method for obtaining the true crack degree is: For any glass frosting detection image, according to the frosting amplified texture contained in the glass frosting detection image, the frosting reaction trend of the frosting phenomenon between the glass frosting detection image and the subsequent adjacent glass frosting detection images is analyzed to obtain the frosting dew point detection interference degree of the glass frosting detection image; For any frosting amplified texture in the glass frosting detection map, combining the crack texture diffusion degree with the frosting dew point detection interference degree to obtain the real crack degree of the frosting amplified texture includes: taking the normalized value of the ratio of the crack texture diffusion degree of the frosting amplified texture to the frosting dew point detection interference degree of the glass frosting detection map as the real crack degree of the frosting amplified texture; The method for obtaining the frosting dew point detection interference degree is: Compare the growth trend of the frosted area between the glass frosting detection map and the subsequent adjacent glass frosting detection map to obtain the frosting dew point detection interference degree of the glass frosting detection map; The interference degree of frosting dew point detection is calculated by the following formula: In the formula, Indicates Frost dew point detection interference degree of glass frosting detection diagram; Indicates The number of all glass frost detection images after the glass frost detection image; Indicates The area of ​​the frosted area in the glass frosting detection map; Indicates After the glass frost detection picture The area of ​​the frosted area in the glass frosting detection map; Indicates The mean gray value of all pixels in the frosted area of ​​the glass frosting detection image; Indicated in After the glass frost detection picture The average grayscale value of all pixels in the frosted area in the glass frosting detection image; Indicates taking the absolute value.

2. A production quality inspection method for insulating glass according to claim 1, characterized in that: The method for obtaining the glass frosting detection image is as follows: For any insulating glass dew point detection image, the aggregation distribution of grayscale information in the insulating glass dew point detection image is analyzed to obtain the frosting area of ​​the insulating glass dew point detection image; according to the overall content of grayscale information contained in the frosting area, several glass frosting detection images are screened out from all insulating glass dew point detection images.

3. A production quality inspection method for insulating glass according to claim 1, characterized in that: After calculating the true crack extent, it also includes: The actual crack extent is normalized.

4. A production quality inspection system for insulating glass, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of a production quality detection method for insulating glass as described in any one of claims 1 to 3 are implemented.

Images