A method and device for accurately testing the strength of tempered glass

Abstract

The present invention relates to the technical field of strength testing, and in particular to a method and device for accurately testing the strength of tempered glass, comprising: collecting images of glass fragments after impact damage to obtain a connected domain of glass fragments; analyzing the grayscale features of through-cracks to mark suspected crack pixels in the connected domain of glass fragments; determining a target through-crack set in the connected domain of glass fragments according to the shape features of the through-cracks and the grayscale distribution features in the connected domain of glass fragments; obtaining the penetration of each element in the target through-crack set according to the contour features of the element; determining a real through-crack according to the penetration; determining the actual number of glass fragments in the glass fragment image based on the real through-cracks, and obtaining the strength test result of the tempered glass. The present invention obtains a more accurate number of glass fragments, and improves the accuracy of the strength test of tempered glass.

Description

Technical Field

[0001] The invention relates to the technical field of strength testing, and in particular to a method and a device for accurately testing the strength of tempered glass. Background Art

[0002] Tempered glass is a specially processed glass with high strength and heat resistance. The strength test of tempered glass is an important part of quality inspection, which helps to ensure the quality and safety of the product. The strength test of tempered glass usually includes impact test, compression test and bending test, among which the impact test is to judge the strength performance according to the number and shape of the broken glass fragments.

[0003] In the actual test process, edge detection is performed on the image of tempered glass fragments after the impact test to obtain the number of glass fragments. However, some glass fragments may have through cracks inside. These through cracks are relatively thin, making it impossible for the edge detection algorithm to fully detect them, resulting in a reduction in the accuracy of the strength test results of the tempered glass. Summary of the invention

[0004] The invention provides a method and a device for accurately testing the strength of tempered glass to solve the existing problems.

[0005] The present invention provides a method and device for accurately testing the strength of tempered glass using the following technical solutions:

[0006] An embodiment of the present invention provides a method for accurately testing the strength of tempered glass, the method comprising the following steps:

[0007] Collecting images of glass fragments after impact damage, and obtaining a connected domain of glass fragments in the images of glass fragments;

[0008] Analyze the grayscale features of the penetrating cracks in the glass fragments and mark the suspected crack pixels in the connected domain of the glass fragments;

[0009] Connect any two edge pixel points in the connected domain of the glass fragments to obtain a grayscale fluctuation straight line; determine the suspected through-crack area and the effective suspected through-crack area of ​​the straight line in the connected domain of the glass fragments by analyzing the distribution of the suspected crack pixel points on the grayscale fluctuation straight line;

[0010] According to the spatial proximity of the valid suspected through-crack regions, the valid suspected through-crack regions are merged, and combined with other valid suspected through-crack regions and straight line suspected through-crack regions, the target through-crack set in each glass fragment connected domain is obtained;

[0011] Obtaining the penetration of each element according to the profile characteristics of each element in the target through-crack set; and determining the real through-crack according to the penetration;

[0012] The actual number of glass fragments in the glass fragment image is determined based on the real through-cracks, and the strength test result of the tempered glass is obtained.

[0013] Preferably, the grayscale features of the penetrating cracks in the glass fragments are analyzed to mark the suspected crack pixels in the connected domain of the glass fragments, including the following specific methods:

[0014] For any glass fragment connected domain, the grayscale mean of all pixels in the glass fragment connected domain is recorded as the first threshold, all pixels in the glass fragment connected domain are traversed, and pixels greater than the first threshold are marked as suspected crack pixels.

[0015] Preferably, the method of connecting any two edge pixel points in the connected domain of the glass fragments to obtain the grayscale fluctuation straight line includes the following specific methods:

[0016] Taking any edge pixel point in the connected domain of glass fragments as the reference point, connect the reference point with other edge pixel points in the connected domain of glass fragments, and record the connecting line segment as the grayscale fluctuation line.

[0017] Preferably, the method of analyzing the distribution of the suspected crack pixel points on the grayscale fluctuation line to determine that the straight line in the glass fragment connected domain is suspected to pass through the crack area includes the following specific methods:

[0018] For any glass fragment connected domain, start from the edge line where any reference point is located and traverse all the grayscale fluctuation lines connecting the reference point and other edge pixel points in a clockwise direction;

[0019] When the ratio of the number of suspected crack pixels to the number of all pixels on the grayscale fluctuation line exceeds the preset parameter When , the area formed by all the pixel points on the grayscale fluctuation straight line is recorded as the straight line suspected to penetrate the crack area.

[0020] Preferably, the specific method for obtaining the effective suspected through-crack region is:

[0021] Obtain the target suspected area in the connected domain of the glass fragments;

[0022] The set formed by the coordinate values ​​of all the pixel points in the suspected through-crack set in the target suspected area in the horizontal direction of the global image is recorded as the horizontal position set of the suspected through-crack set;

[0023] Constructing a set length scatter plot according to the length of the horizontal position set of each suspected through-crack set in the target suspected area and the sequence number of the grayscale fluctuation line to which the suspected through-crack set belongs;

[0024] Analyze the distribution characteristics of data points in the set length scatter plot in the target suspected area to obtain the regional authenticity of the target suspected area;

[0025] According to the regional authenticity of the target suspected area, the target suspected area is determined to be a valid suspected through-crack area.

[0026] Preferably, the specific method for obtaining the target suspected area in the glass fragment connected domain is:

[0027] When the ratio of the number of suspected crack pixels to the number of all pixels on the grayscale fluctuation line does not exceed the preset parameter When , the set of suspected crack pixel points that continuously exist on the grayscale fluctuation line is recorded as the suspected set of the grayscale fluctuation line;

[0028] Acquire a horizontal position set of the suspected set, wherein the horizontal position set of the suspected set is acquired in the same manner as the horizontal position set of the suspected through-crack set;

[0029] When the horizontal position sets of the suspected sets of any two adjacent grayscale fluctuation straight lines in the connected domain of glass fragments intersect, the suspected sets of the two adjacent grayscale fluctuation straight lines are marked as suspected through-crack sets. Similarly, all the suspected sets of adjacent grayscale fluctuation straight lines in the connected domain of glass fragments are traversed to obtain several target suspected areas in the connected domain of glass fragments formed by the pixel points in all the suspected through-crack sets.

[0030] Preferably, the effective suspected through-crack regions are merged according to the spatial proximity of the effective suspected through-crack regions, including the specific method of:

[0031] For any two valid suspected through-crack regions in the connected domain of glass fragments, when the distance between any edge pixel point of any valid suspected through-crack region and any edge pixel point of the other valid suspected through-crack region in the vertical direction or horizontal direction is less than the preset parameter , the two valid suspected through-crack regions are merged using the regional growing algorithm, and the merged valid suspected crack region is recorded as the merged suspected through-crack region.

[0032] Preferably, the method of obtaining the penetration degree of each element according to the profile feature of each element in the target through-crack set includes:

[0033] Obtain the minimum circumscribed moments of all suspected through-crack regions in the target through-crack region set, and record the longest side of the minimum circumscribed moments as the first straight line; extend the first straight line until it intersects with two edges of the connected domain of the glass fragments, and record the straight-line distance between the two intersection points as the second straight line;

[0034] The specific calculation method of the penetration of any suspected through-crack region in any target through-crack region set is:

[0035]

[0036] In the formula, Indicates The target penetrates the crack area set The penetration of the suspected through-crack area; Indicates The target penetrates the crack area set The length of the first straight line suspected to penetrate the crack area; Indicates The target penetrates the crack area set The length of the first straight line suspected to penetrate the crack area; Describes the The target penetrates the crack area set The contour features of the suspected through-crack area.

[0037] Preferably, the determining the actual number of glass fragments in the glass fragment image based on the real through-cracks comprises the following specific methods:

[0038] On the glass fragment edge image, the grayscale values ​​of all pixels of real through-cracks are adjusted to be consistent with the grayscale values ​​of the pixels at the edge of the glass fragment;

[0039] The connected area formed by the crack area and the edge of the broken glass is marked as a separate glass fragment;

[0040] Get the number of glass fragments in the glass fragments image.

[0041] A device for accurately testing the strength of tempered glass, using any one of the methods for accurately testing the strength of tempered glass, the device comprises the following modules:

[0042] An image acquisition module is used to acquire images of glass fragments after impact damage and obtain a connected domain of glass fragments in the images of glass fragments;

[0043] The suspected crack pixel marking module is used to analyze the grayscale characteristics of the penetrating cracks in the glass fragments and mark the suspected crack pixels in the connected domain of the glass fragments;

[0044] The through-crack acquisition module is used to connect any two edge pixel points in the connected domain of glass fragments to obtain a grayscale fluctuation straight line; by analyzing the distribution of suspected crack pixels on the grayscale fluctuation straight line, the straight line suspected through-crack area and the effective suspected through-crack area in the connected domain of glass fragments are determined;

[0045] The target through-crack set module is used to merge the valid suspected through-crack regions according to their spatial proximity, and combine them with other valid suspected through-crack regions and straight line suspected through-crack regions to obtain the target through-crack set in each glass fragment connected domain;

[0046] A real through-crack calibration module is used to obtain the penetration of each element in the target through-crack set according to the profile characteristics of the element; and determine the real through-crack according to the penetration;

[0047] The strength detection module is used to determine the actual number of glass fragments in the glass fragment image based on the real through-cracks and obtain the strength test result of the tempered glass.

[0048] The technical solution of the present invention has the following beneficial effects: first, by analyzing the grayscale features of through cracks in glass fragments, all suspected crack pixels are marked in the connected domain of glass fragments; then, any edge pixel in the connected domain of glass fragments is connected to other edge pixel points to obtain a grayscale fluctuation straight line; according to the distribution of suspected crack pixels on the grayscale fluctuation straight line, the straight suspected through crack region and the target suspected region in each connected domain of glass fragments are determined; by analyzing the change characteristics of the suspected through crack set in the target suspected region on the grayscale fluctuation straight line, the effective suspected through crack region is obtained; according to the spatial proximity of the effective suspected through crack regions, the effective suspected through crack regions are merged, and combined with other effective suspected through crack regions and the straight suspected through crack region, the target through crack set in each connected domain of glass fragments is obtained; according to the contour features of the suspected through crack region, the penetration of the suspected through crack region is obtained; the real through crack is determined by the penetration of each suspected through crack region in the target through crack region set; finally, the actual number of glass fragments in the glass fragment image is determined based on the real through crack, and the strength test result of the tempered glass is obtained. The above process effectively detects the through cracks in the glass fragments by analyzing the shape characteristics of the through cracks and the grayscale distribution characteristics in the connected domain of the glass fragments; thereby obtaining a more accurate number of glass fragments and improving the accuracy of the tempered glass strength test. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] 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.

[0050] Figure 1 A flowchart of the steps of a method for accurately testing the strength of tempered glass according to the present invention;

[0051] Figure 2 The present invention is a structural block diagram of a device for accurately testing the strength of tempered glass. DETAILED DESCRIPTION

[0052] 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 a method and device for accurately testing the strength of tempered glass proposed by the present invention, its specific implementation, structure, features and effects, in conjunction 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 may be combined in any suitable form.

[0053] 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.

[0054] The specific scheme of a method and device for accurately testing the strength of tempered glass provided by the present invention is described in detail below with reference to the accompanying drawings.

[0055] See also Figure 1 , which shows a flowchart of a method for accurately testing the strength of tempered glass provided by an embodiment of the present invention, the method comprising the following steps:

[0056] Step S001: collecting images of glass fragments after impact damage, and obtaining a glass fragment connected domain in the glass fragment image.

[0057] It should be noted that the strength performance of tempered glass is a key indicator for evaluating the safety and reliability of the tempered glass; in strength tests, the number of tempered glass fragments after impact damage is usually used as a strength performance evaluation indicator; however, after the tempered glass is damaged by impact, the fragments formed usually contain through cracks, and the grayscale value of the through crack pixels is low, which is difficult to be fully detected, resulting in misjudgment of the number of glass fragments, and then producing inaccurate results for the strength performance evaluation of the tempered glass.

[0058] Specifically, in order to implement the method for accurately testing the strength of tempered glass proposed in this embodiment, it is first necessary to collect images of tempered glass fragments after impact damage. The specific process is as follows:

[0059] Step (1.1), collecting images of tempered glass fragments after impact damage, all recorded as glass fragment images, and graying the glass fragment images to obtain glass fragment grayscale images.

[0060] The specific collection method is: place the tempered glass to be tested horizontally in the middle of the test platform, use a punch to destroy the tempered glass within 25mm of the center of the tempered glass; install an industrial camera directly above the test platform to collect images of glass fragments after impact damage.

[0061] Step (1.2), using the Sobel operator to perform edge detection on the grayscale image of the glass fragments, obtaining the edge image of the grayscale image of the glass fragments, and using the depth-first search algorithm to perform connected domain detection on the edge image to obtain several connected domains of the glass fragments.

[0062] It should be noted that the Sobel operator can effectively capture edge pixels with significant changes in grayscale values ​​in glass fragment images, but cannot detect all through cracks with relatively low grayscale gradient values ​​in glass fragments; therefore, each connected domain of glass fragments obtained is regarded as a separate glass fragment, and then all through cracks are obtained based on the characteristics of the through cracks in the glass fragments.

[0063] It should be further explained that the Sobel operator and the depth-first search algorithm are both existing image processing methods, and thus are not described in detail in this embodiment.

[0064] So far, all the glass fragment connected domains in the glass fragment image are obtained through the above method.

[0065] Step S002: Analyze the grayscale features of the penetrating cracks in the glass fragments, and mark the suspected crack pixels in the connected domain of the glass fragments.

[0066] It should be noted that for glass fragments with through-cracks, because the grayscale value of the through-crack pixels is greater than the grayscale value of the pixels in the normal area of ​​the glass fragments, but less than the grayscale value of the edge of the glass fragments, the suspected crack pixels in the connected domain of the glass fragments are obtained based on the grayscale features of the through-cracks.

[0067] Specifically, for any glass fragment connected domain, the grayscale mean of all pixels in the glass fragment connected domain is recorded as the first threshold, all pixels in the glass fragment connected domain are traversed, and pixels greater than the first threshold are marked as suspected crack pixels.

[0068] So far, the suspected crack pixels in the connected domain of all glass fragments are obtained through the above method.

[0069] Step S003: connect any two edge pixel points in the connected domain of glass fragments to obtain a grayscale fluctuation straight line; determine the straight line suspected through-crack area and the target suspected area in each connected domain of glass fragments according to the distribution of suspected crack pixels on the grayscale fluctuation straight line; obtain the effective suspected through-crack area by analyzing the change characteristics of the suspected through-crack set in the target suspected area on the grayscale fluctuation straight line.

[0070] S1: Connect any two edge pixel points in the connected domain of glass fragments to obtain a grayscale fluctuation straight line; determine that the straight line in each connected domain of glass fragments is suspected to pass through the crack area and the target suspected area according to the distribution of the suspected crack pixel points on the grayscale fluctuation straight line.

[0071] It should be noted that the through cracks in the glass fragments are usually straight through cracks and curved through cracks, so the target suspected area in the glass connected domain is determined according to the shape characteristics of the through cracks.

[0072] Specifically, any edge pixel point in the connected domain of glass fragments is used as a reference point, and the reference point is connected with any pixel point on other edges of the connected domain of glass fragments, and the connecting line segment is recorded as a grayscale fluctuation line; all grayscale fluctuation lines in the connected domain are obtained.

[0073] It should be further explained that the distribution range and characteristics of the suspected crack pixels on the grayscale fluctuation line describe the distribution of the suspected crack pixels on the grayscale fluctuation line, and the suspected through-crack area in each glass fragment connected domain is determined based on the distribution; for straight through cracks, when most of the pixels on the grayscale fluctuation line are suspected crack pixels, then the grayscale fluctuation line may be a straight through crack area; for curved through cracks, due to their irregular shapes, it is necessary to analyze the positions and range changes of the suspected crack pixels on adjacent grayscale fluctuation lines to determine the suspected curved through crack area.

[0074] Specifically, first, for any glass fragment connected domain, all grayscale fluctuation lines connecting the reference point and other edge pixel points are traversed clockwise from the edge line where any reference point is located; when the ratio of the number of suspected crack pixels to the number of all pixels on the grayscale fluctuation line exceeds When (preset parameters based on experience , which can be adjusted according to actual conditions and is not specifically limited in this embodiment), the area formed by all pixel points on the grayscale fluctuation straight line is recorded as the straight line suspected to penetrate the crack area.

[0075] Then, when the ratio of the number of suspected crack pixels to the number of all pixels on the grayscale fluctuation line does not exceed When , the set of suspected crack pixel points that exist continuously on the grayscale fluctuation line is recorded as the suspected set; the set formed by the coordinate values ​​of all pixel points in the suspected set in the horizontal direction of the global image is recorded as the horizontal position set of the suspected set.

[0076] It should be noted that a single suspected crack pixel point that appears independently on the grayscale fluctuation line is also regarded as a suspected set.

[0077] Finally, if there is a curved crack in any glass fragment connected domain, for the glass fragment connected domain, all grayscale fluctuation lines connecting the reference point and other edge pixel points are traversed clockwise from the edge line where any reference point is located; when the horizontal position sets of two suspected sets on adjacent grayscale fluctuation lines have an intersection relationship (for example: the horizontal position set of the first suspected set in the first grayscale fluctuation line is ; The horizontal position set of the first suspected set in the adjacent second grayscale fluctuation line is , the horizontal position set of the second suspected set is ; then the horizontal position set of the first suspected set in the first grayscale fluctuation straight line and the second suspected set in the second grayscale fluctuation straight line have an intersection, that is, an intersection relationship), and these two suspected sets are marked as suspected through-crack sets.

[0078] For the first suspected set on the first grayscale fluctuation straight line, when there is an intersection relationship with the horizontal position set of the suspected set on the adjacent grayscale fluctuation straight line (i.e., the second grayscale fluctuation straight line), the two suspected sets are marked as suspected through-crack sets; based on the intersection relationship, and by analogy, the suspected through-crack sets on all grayscale fluctuation straight lines are obtained; all the obtained suspected through-crack sets are recorded as the first target suspected area.

[0079] Similarly, for the second suspected set on the first grayscale fluctuation line, according to the set relationship, and so on, obtain the suspected through-crack sets on all remaining grayscale fluctuation lines; all the obtained suspected sets are recorded as the second target suspected area.

[0080] The remaining suspected sets on the first grayscale fluctuation straight line are analyzed by the above method; then, the unanalyzed suspected sets on the remaining grayscale fluctuation straight lines are analyzed.

[0081] So far, all the target suspected regions in any glass fragment connected domain are obtained through the above method.

[0082] S2: Obtain the effective suspected through-crack area by analyzing the change characteristics of the set of suspected through-cracks in the target suspected area on the grayscale fluctuation line.

[0083] It should be noted that for the curve penetrating crack, when traversing the grayscale straight line according to the order of obtaining the grayscale fluctuation straight line, during the traversal process, it will first gradually approach the actual position of the curve penetrating crack, so the number of suspected crack pixels will gradually increase; then, continue to traverse the grayscale fluctuation straight line, the number of suspected crack pixels will reach a peak; as the traversal continues and leaves the curve penetrating crack area, the number of suspected crack pixels will gradually decrease. For the curve penetrating crack, as the grayscale fluctuation straight line is traversed, the number of suspected crack pixels gradually increases and then gradually decreases. Therefore, the effective suspected penetrating crack area is judged based on this feature.

[0084] Specifically, a set length scatter plot is constructed based on the length of the horizontal position set of each suspected through-crack set in the target suspected area and the serial number of the grayscale fluctuation straight line to which the suspected through-crack set belongs; and the least squares method is used to fit all the data points in the set length scatter plot into a curve, namely, the set length curve.

[0085] The horizontal coordinates of all data points in the set length scatter plot of any target suspected area are arranged in ascending order, and the data point corresponding to the horizontal coordinate point in the middle position of the arranged data point sequence is recorded as the middle data point; the set length curve is divided from the middle data point to obtain two sub-curve segments, and the sub-curve segment on the left side of the middle data point is recorded as the first sub-curve segment, and the sub-curve segment on the right side of the middle data point is recorded as the second curve segment.

[0086] It should be noted that if the number of data points in the set length scatter plot of the target suspected area is an even number, the data point on the left of the two middle data points is selected as the middle data point.

[0087] Then the regional authenticity calculation expression of any target suspected area in any glass fragment connected domain is:

[0088]

[0089] In the formula, Indicates the regional authenticity of the nth target suspected area in the mth glass fragment connected domain; Represents the mean ordinate value of all data points on the first sub-curve segment; Represents the mean ordinate value of all data points on the second sub-curve segment; represents the average slope of the first sub-curve segment; represents the average slope of the first sub-curve segment; Indicates obtaining the absolute value; Represents an exponential function with a natural constant as base.

[0090] It should be noted that the smaller the difference between the mean values ​​of the ordinates of all data points on the first sub-curve segment and the second sub-curve segment in the set length curve, the more similar the lengths of the horizontal position sets corresponding to the data points on the two sub-curves are; the larger the sum of the average slopes of the first sub-curve segment and the second sub-curve segment in the set length curve, the more similar the slopes of the two sub-curve segments are. The larger the value of , the greater the probability that the length changes of the horizontal position sets corresponding to the data points on the two sub-curve segments show opposite trends, that is, the more obvious the change feature of first increasing and then decreasing is, indicating that the regional authenticity of the corresponding target suspected area is greater, that is, the greater the probability that the target suspected area is a crack through which the real curve penetrates.

[0091] Finally, the regional authenticity of all target suspected areas is obtained. When the regional authenticity is greater than the preset threshold When , it indicates that the probability that the target suspected area is a curved through-crack is relatively high, and therefore the target suspected area is determined to be a valid suspected through-crack area.

[0092] It should be further explained that, in this embodiment, a threshold is given , all the suspected target areas in the glass fragment connected domain are judged; in addition, the threshold is preset according to experience , can be adjusted according to actual conditions, and this embodiment does not make any specific limitation.

[0093] So far, all the valid suspected through-crack regions in the connected domain of the glass fragments are obtained through the above method.

[0094] Step S004: according to the spatial proximity of the valid suspected through-crack regions, the valid suspected through-crack regions are merged, and combined with other valid suspected through-crack regions and straight line suspected through-crack regions to obtain a target through-crack set in each glass fragment connected domain.

[0095] It should be noted that, since the shape of the curved through crack is irregular, the effective suspected through cracks are connected according to the closeness of the spatial distances between the effective suspected through crack regions in the connected domain of the glass fragments.

[0096] Specifically, for any two valid suspected through-crack regions in the connected domain of glass fragments, when the distance between any edge pixel point of any valid suspected through-crack region and any edge pixel point of the other valid suspected through-crack region in the vertical direction or horizontal direction is less than the preset parameter (Preset parameters based on experience pixel points, which can be adjusted according to actual conditions and are not specifically limited in this embodiment), the two valid suspected through-crack regions are merged using a region growing algorithm, and the merged valid suspected crack region is recorded as a merged suspected through-crack region.

[0097] The merged suspected through-crack regions, effective suspected through-crack regions, and straight line suspected through-crack regions in any glass fragment connected domain are recorded as a target suspected through-crack region set; and each target suspected through-crack region set corresponds to a glass fragment connected domain, and the elements in the target through-crack set are collectively referred to as suspected through-crack regions.

[0098] It should be noted that when the distance between any edge pixels of any two valid suspected through-crack regions in the vertical direction or the horizontal direction represents the spatial proximity of the two valid suspected through-crack regions, the distance is less than the preset parameter , indicating that the set distances of the two valid suspected through-crack regions are relatively close and they can be merged.

[0099] It should be further explained that the region growing algorithm is a well-known technology and will not be described in detail in this embodiment.

[0100] At this point, the target through-crack set of all glass fragment connected domains is obtained through the above method.

[0101] Step S005: Obtain the penetration of the suspected through crack region according to the contour features of the suspected through crack region; determine the real through crack by the penetration of each suspected through crack region in the target through crack region set.

[0102] It should be noted that the length of each suspected through-crack region in the target through-crack region set is different, so it is necessary to calculate the penetration of the suspected through-crack region through the contour features of the suspected through-crack region; and determine the real through-crack based on the penetration.

[0103] Specifically, the minimum circumscribed moments of all suspected through-crack regions in the target through-crack region set are obtained, and the longest side in the minimum circumscribed moment is recorded as a first straight line; the first straight line is extended until it intersects with the edge line of the connected region of the glass fragments to obtain two intersection points, and the straight-line distance between the two intersection points is recorded as a second straight line;

[0104] Then the calculation expression of the penetration degree of any suspected through-crack region in any target through-crack region set is:

[0105]

[0106] In the formula, Indicates The target penetrates the crack area set The penetration of the suspected through-crack area; Indicates The target penetrates the crack area set The length of the first straight line suspected to penetrate the crack area; Indicates The target penetrates the crack area set The length of a second straight line suspected to pass through the crack area.

[0107] It should be further explained that The target through-crack region set is The target penetration crack area set in the connected domain of glass fragments; contour features The larger the value, the The target penetrates the crack area set The greater the penetration of the suspected through crack area, the The greater the probability that a suspected through-crack region is a real through-crack region.

[0108] Get the penetration of all suspected through-crack areas in the target through-crack area set. When the penetration is greater than the preset threshold When (preset threshold based on experience , which can be adjusted according to actual conditions and is not specifically limited in this embodiment), and the suspected through-crack area is recorded as a real through-crack.

[0109] So far, the real through-cracks in the connected domains of all glass fragments are obtained through the above method.

[0110] Step S006: determining the actual number of glass fragments in the glass fragment image based on the real through-cracks, and obtaining the strength test result of the tempered glass.

[0111] It should be noted that the actual number of glass fragments in the glass fragment image is determined based on the real through-cracks; and the strength performance of the tempered glass is tested according to the number of glass fragments.

[0112] Specifically, first, on the glass fragment edge image, the grayscale values ​​of all pixels of real through-cracks are adjusted to be consistent with the grayscale values ​​of the pixels of the glass fragment edge; and the connected domain formed by the through-crack area and the glass fragment edge is recorded as a glass fragment.

[0113] Then, the number of connected domains in the edge image of the grayscale image of the glass fragments is determined;

[0114] Finally, according to the contents disclosed in Section 5.6 of the standard number GB15763.2-2005 "Safety Glass for Buildings Part 2: Tempered Glass", the minimum number of fragments in any 50 mm×50 mm area in the tempered glass fragment image is obtained.

[0115] The strength of tempered glass is determined by the minimum number of fragments in any 50mm×50mm area in the tempered glass fragment image. The specific method is as follows:

[0116] Get the strength parameters of tempered glass. The specific calculation expression is:

[0117]

[0118] In the formula, Indicates the strength parameter of the tempered glass to be tested; Indicates the minimum number of fragments within the 50mm×50mm area of ​​the tempered glass to be tested; Indicates the number of fragments in any 50mm×50mm area of ​​the tempered glass to be tested.

[0119] when When , the quality of the tempered glass to be tested is unqualified; when When , the tempered glass to be tested is of low strength; when When , the tempered glass to be tested is of medium strength; when When the tempered glass to be tested is of high strength.

[0120] Preset thresholds based on experience , can be adjusted according to actual conditions, and this embodiment does not make any specific limitation.

[0121] It should be noted that the parameters are preset based on experience. and , can be adjusted according to actual conditions, and this embodiment does not make any specific limitation.

[0122] It should be further explained that, in the calculation of the strength parameters of the tempered glass to be tested, the minimum number of fragments in the 50mm×50mm area is related to the nominal thickness of the tempered glass; when the nominal thickness of the tempered glass is 3mm, the minimum number of fragments is 30 pieces; when the nominal thickness is 4-12mm, the minimum number of fragments is 30 pieces; when the nominal thickness is greater than or equal to 15mm, the minimum number of fragments is 30 pieces.

[0123] Through the above steps, the strength test of tempered glass is completed.

[0124] See also Figure 2 , which shows a structural block diagram of a device for accurately testing the strength of tempered glass provided by an embodiment of the present invention, the device includes the following modules:

[0125] An image acquisition module is used to acquire images of glass fragments after impact damage and obtain a connected domain of glass fragments in the images of glass fragments;

[0126] The suspected crack pixel marking module is used to analyze the grayscale characteristics of the penetrating cracks in the glass fragments and mark the suspected crack pixels in the connected domain of the glass fragments;

[0127] The through-crack acquisition module is used to connect any two edge pixel points in the connected domain of glass fragments to obtain a grayscale fluctuation straight line; by analyzing the distribution of suspected crack pixels on the grayscale fluctuation straight line, the straight line suspected through-crack area and the effective suspected through-crack area in the connected domain of glass fragments are determined;

[0128] The target through-crack set module is used to merge the valid suspected through-crack regions according to their spatial proximity, and combine them with other valid suspected through-crack regions and straight line suspected through-crack regions to obtain the target through-crack set in each glass fragment connected domain;

[0129] A real through-crack calibration module is used to obtain the penetration of each element in the target through-crack set according to the profile characteristics of the element; and determine the real through-crack according to the penetration;

[0130] The strength detection module is used to determine the actual number of glass fragments in the glass fragment image based on the real through-cracks and obtain the strength test result of the tempered glass.

[0131] In this embodiment, firstly, by analyzing the grayscale features of the through cracks in the glass fragments, all the suspected crack pixels are marked in the connected domain of the glass fragments; then, any edge pixel in the connected domain of the glass fragments is connected with other edge pixel points to obtain a grayscale fluctuation straight line; according to the distribution of the suspected crack pixels on the grayscale fluctuation straight line, the straight suspected through crack region and the target suspected region in each connected domain of the glass fragment are determined; by analyzing the change characteristics of the suspected through crack set in the target suspected region on the grayscale fluctuation straight line, the effective suspected through crack region is obtained; according to the spatial proximity of the effective suspected through crack regions, the effective suspected through crack regions are merged, and combined with other effective suspected through crack regions and the straight suspected through crack region, the target through crack set in each connected domain of the glass fragment is obtained; according to the contour features of the suspected through crack region, the penetration of the suspected through crack region is obtained; the real through crack is determined by the penetration of each suspected through crack region in the target through crack region set; finally, the actual number of glass fragments in the glass fragment image is determined based on the real through crack, and the strength test result of the tempered glass is obtained. The above process effectively detects the through cracks in the glass fragments by analyzing the shape characteristics of the through cracks and the grayscale distribution characteristics in the connected domain of the glass fragments; thereby obtaining a more accurate number of glass fragments and improving the accuracy of the tempered glass strength test.

[0132] It should be noted that the The model is only used to represent negative correlation and constrain the output of the model to be in In the specific implementation, it can be replaced by other models with the same purpose. This embodiment is only based on The model is described as an example without any specific limitation. is the input to the model.

[0133] 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 method for accurately testing the strength of tempered glass. It is characterized in that The method comprises the following steps: Collecting images of glass fragments after impact damage, and obtaining a connected domain of glass fragments in the images of glass fragments; Analyze the grayscale features of the penetrating cracks in the glass fragments and mark the suspected crack pixels in the connected domain of the glass fragments; Connect any two edge pixel points in the connected domain of the glass fragments to obtain a grayscale fluctuation straight line; determine the suspected through-crack area and the effective suspected through-crack area of ​​the straight line in the connected domain of the glass fragments by analyzing the distribution of the suspected crack pixel points on the grayscale fluctuation straight line; According to the spatial proximity of the valid suspected through-crack regions, the valid suspected through-crack regions are merged, and combined with other valid suspected through-crack regions and straight line suspected through-crack regions, the target through-crack set in each glass fragment connected domain is obtained; Obtaining the penetration of each element in the target through-crack set according to the profile characteristics of the element; and determining the real through-crack according to the penetration; Determine the actual number of glass fragments in the glass fragment image based on the real through-cracks, and obtain the strength test result of the tempered glass; The method of analyzing the distribution of the suspected crack pixel points on the grayscale fluctuation line to determine that the straight line in the glass fragment connected domain is suspected to penetrate the crack area includes the following specific methods: For any glass fragment connected domain, start from the edge line where any reference point is located and traverse all the grayscale fluctuation lines connecting the reference point and other edge pixel points in a clockwise direction; When the ratio of the number of suspected crack pixels to the number of all pixels on the grayscale fluctuation line exceeds the preset parameter When , the area formed by all the pixel points on the grayscale fluctuation straight line is recorded as the area where the straight line is suspected to penetrate the crack; The specific method for obtaining the effective suspected through-crack area is: Obtain the target suspected area in the connected domain of the glass fragments; The set formed by the coordinate values ​​of all the pixel points in the suspected through-crack set in the target suspected area in the horizontal direction of the global image is recorded as the horizontal position set of the suspected through-crack set; Constructing a set length scatter plot according to the length of the horizontal position set of each suspected through-crack set in the target suspected area and the sequence number of the grayscale fluctuation line to which the suspected through-crack set belongs; Analyze the distribution characteristics of data points in the set length scatter plot in the target suspected area to obtain the regional authenticity of the target suspected area; According to the regional authenticity of the target suspected area, determining the target suspected area as a valid suspected through-crack area; The specific method for obtaining the target suspected area in the glass fragment connected domain is: When the ratio of the number of suspected crack pixels to the number of all pixels on the grayscale fluctuation line does not exceed the preset parameter When , the set of suspected crack pixel points that continuously exist on the grayscale fluctuation line is recorded as the suspected set of the grayscale fluctuation line; Acquire a horizontal position set of the suspected set, wherein the horizontal position set of the suspected set is acquired in the same manner as the horizontal position set of the suspected through-crack set; When the horizontal position sets of the suspected sets of any two adjacent grayscale fluctuation straight lines in the connected domain of glass fragments intersect, the suspected sets of the two adjacent grayscale fluctuation straight lines are marked as suspected through-crack sets. Similarly, all the suspected sets of adjacent grayscale fluctuation straight lines in the connected domain of glass fragments are traversed to obtain several target suspected areas in the connected domain of glass fragments formed by the pixel points in all the suspected through-crack sets.

2. According to the method for accurately testing the strength of tempered glass as claimed in claim 1, It is characterized in that The specific method of analyzing the grayscale features of the penetrating cracks in the glass fragments and marking the suspected crack pixels in the connected domain of the glass fragments includes: For any glass fragment connected domain, the grayscale mean of all pixels in the glass fragment connected domain is recorded as the first threshold, all pixels in the glass fragment connected domain are traversed, and pixels greater than the first threshold are marked as suspected crack pixels.

3. According to the method for accurately testing the strength of tempered glass as claimed in claim 1, It is characterized in that The specific method of connecting any two edge pixel points in the connected domain of the glass fragments to obtain the grayscale fluctuation straight line includes: Taking any edge pixel point in the connected domain of glass fragments as the reference point, connect the reference point with other edge pixel points in the connected domain of glass fragments, and record the connecting line segment as the grayscale fluctuation line.

4. According to the method for accurately testing the strength of tempered glass as claimed in claim 1, It is characterized in that The specific method of merging the effective suspected through crack regions according to the spatial proximity of the effective suspected through crack regions is as follows: For any two valid suspected through-crack regions in the connected domain of glass fragments, when the distance between any edge pixel point of any valid suspected through-crack region and any edge pixel point of the other valid suspected through-crack region in the vertical direction or horizontal direction is less than the preset parameter , the two valid suspected through-crack regions are merged using the regional growing algorithm, and the merged valid suspected crack region is recorded as the merged suspected through-crack region.

5. According to the method for accurately testing the strength of tempered glass as claimed in claim 1, It is characterized in that The specific method of obtaining the penetration degree of each element according to the profile feature of each element in the target through-crack set is as follows: Obtain the minimum circumscribed moments of all suspected through-crack regions in the target through-crack region set, and record the longest side of the minimum circumscribed moments as the first straight line; extend the first straight line until it intersects with two edges of the connected domain of the glass fragments, and record the straight-line distance between the two intersection points as the second straight line; The specific calculation method of the penetration of any suspected through-crack region in any target through-crack region set is: In the formula, Indicates The target penetrates the crack area set The penetration of the suspected through-crack area; Indicates The target penetrates the crack area set The length of the first straight line suspected to penetrate the crack area; Indicates The target penetrates the crack area set The length of the first straight line suspected to penetrate the crack area; Describes the The target penetrates the crack area set The contour features of the suspected through-crack area.

6. According to the method for accurately testing the strength of tempered glass as claimed in claim 1, It is characterized in that The specific method of determining the actual number of glass fragments in the glass fragment image based on the real through-cracks includes: On the glass fragment edge image, the grayscale values ​​of all pixels of real through-cracks are adjusted to be consistent with the grayscale values ​​of the pixels at the edge of the glass fragment; The connected area formed by the crack area and the edge of the broken glass is marked as a separate glass fragment; Get the number of glass fragments in the glass fragments image.

7. A device for accurately testing the strength of tempered glass, using a method for accurately testing the strength of tempered glass as claimed in any one of claims 1 to 6, It is characterized in that The device includes the following modules: An image acquisition module is used to acquire images of glass fragments after impact damage and obtain a connected domain of glass fragments in the images of glass fragments; The suspected crack pixel marking module is used to analyze the grayscale characteristics of the penetrating cracks in the glass fragments and mark the suspected crack pixels in the connected domain of the glass fragments; The through-crack acquisition module is used to connect any two edge pixel points in the connected domain of glass fragments to obtain a grayscale fluctuation straight line; by analyzing the distribution of suspected crack pixels on the grayscale fluctuation straight line, the straight line suspected through-crack area and the effective suspected through-crack area in the connected domain of glass fragments are determined; The target through-crack set module is used to merge the valid suspected through-crack regions according to their spatial proximity, and combine them with other valid suspected through-crack regions and straight line suspected through-crack regions to obtain the target through-crack set in each glass fragment connected domain; A real through-crack calibration module is used to obtain the penetration of each element in the target through-crack set according to the profile characteristics of the element; and determine the real through-crack according to the penetration; The strength detection module is used to determine the actual number of glass fragments in the glass fragment image based on the real through-cracks and obtain the strength test result of the tempered glass.

Images