Multi-angle three-dimensional imaging label 3D shape detection system and method thereof

Abstract

The invention provides a multi-angle three-dimensional imaging label 3D morphology detection system and method, and belongs to the technical field of industrial detection, and the system comprises a data acquisition module which is used for acquiring related data, and the related data comprises a curved surface angle of a labeling area of a target object and the area of a target label; and the data construction module is used for constructing the 3D morphology of the labeling area according to the area and the curved surface angle and detecting the labeling quality of the labeling area in the constructed 3D morphology, and the detection comprises the fitting degree of the target label in the labeling area and the target object. According to the method, the local defects are accurately recognized by constructing the 3D morphology and introducing a multi-dimensional fitting degree analysis method, the method adapts to a complex curved surface scene, the crossing of label detection from a two-dimensional plane to a three-dimensional space is achieved, the detection precision is improved, and meanwhile the method is compatible with industrial diversity requirements.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial detection, in particular to a multi-angle stereoscopic imaging label 3D topography detection system and method thereof. BACKGROUND

[0002] With the rapid development of manufacturing industry, the requirements for product quality and anti-counterfeiting identification are becoming higher and higher, therefore, the label attachment in the fields of medicine packaging, electronic components, food packaging, etc. usually needs to be quality detected.

[0003] For this research, the application file with the application number CN202310789766.1 provides a 3D imaging system and its calibration method, device, equipment and storage medium, which calibrates the telecentric lens according to the calibration image obtained by collecting the first calibration target by the telecentric lens, obtains the first initial calibration parameter of the telecentric lens, determines the second coordinate of the reference point corresponding to the first coordinate relative to the target projection light machine coordinate system according to the first initial calibration parameter, calibrates the target projection light machine according to the second coordinate, and obtains the second initial calibration parameter of the target projection light machine. This technical solution not only can improve the imaging accuracy of the calibrated 3D imaging system, but also can avoid the problem of missing 3D information of components caused by occlusion.

[0004] Another application file with the application number CN202210674122.3 provides a flexible motion imaging multi-line structure stereovision three-dimensional topography measurement method and system, which includes: a group of laser projectors capable of projecting multiple laser surfaces and a monocular camera are fixedly connected, and the structure light is projected and the image is collected on the measured surface from different viewpoints, the sparse overlapping point matching between the local three-dimensional point clouds obtained by adjacent viewpoints is carried out, and the complete dense three-dimensional reconstruction result of the observed surface is obtained by splicing and fusion. This technical solution observes in the motion process, does not require the measured surface to make accurate controllable relative motion, and can freely select a suitable observation angle, which can better solve the problem of structure occlusion.

[0005] However, when detecting the label on the curved surface object, the above technical solution still has defects, and there are occlusion blind areas in actual detection, which leads to insufficient detection accuracy of complex curved surface labels, and cannot detect three-dimensional defects such as label warping and wrinkling. SUMMARY

[0006] In view of the above problems existing in the prior art of industrial detection technology field, the present application is proposed.

[0007] Therefore, one of the purposes of the present application is to provide a multi-angle stereoscopic imaging label 3D topography detection system and method thereof, which realizes the transition from "two-dimensional plane" to "three-dimensional stereoscopic" by constructing 3D topography, introducing multi-dimensional fitting degree analysis method, accurately identifying local defects, and adapting to complex curved surface scenes, improves the detection accuracy, and is compatible with industrial diversity requirements.

[0008] To solve the above technical problems, the present application provides the following technical solutions:

[0009] In one aspect, the present application provides a multi-angle stereoscopic imaging label 3D topography detection system, comprising:

[0010] A data acquisition module for acquiring relevant data, the relevant data including the surface angle of the labeling area of the target object and the area of the target label;

[0011] A data construction module for constructing the 3D topography of the labeling area according to the area and the surface angle, and detecting the labeling quality of the labeling area in the constructed 3D topography, the detection including the fitting degree of the target label in the labeling area and the target object;

[0012] A feature processing module, the feature processing module responding to the detection of the labeling quality, for giving a fitting point at the center of the labeling area, marking the fitting point as a main fitting point, giving at least four vice fitting points around the main fitting point based on the main fitting point, and calculating the front and back distances between the main fitting point and each of the vice fitting points in the constructed 3D topography;

[0013] A data fusion processing module, the data fusion processing module responding to the feature processing module, for analyzing and judging the fitting degree according to the calculation results; the data fusion processing module includes an acquisition unit, an analysis unit, a calculation unit and a judgment unit;

[0014] The acquisition unit is used to acquire the protrusion of the labeling area after being covered by the target label, mark the protrusion as a protrusion point, and acquire the shape corresponding to the protrusion point;

[0015] The analysis unit is used to analyze the edge of the shape, acquire the relevant features of the protrusion point based on the analysis results, and the relevant features include the trend of the protrusion point;

[0016] The calculation unit is used to perform relevant calculations according to the acquired trend;

[0017] The judgment unit responds to the calculation unit, and is used to determine the quality of the fitting degree according to the results of the relevant calculations.

[0018] As a preferred scheme of the present application, in the calculation unit, the correlation calculation step is as follows:

[0019] Based on the shape, the center of the protruding point is obtained, and a protruding part farthest from the center is collected;

[0020] The edge of a target label closest to the protruding part is obtained;

[0021] The highest point and the lowest point of the edge are obtained, and the highest point and the lowest point are two points horizontally and vertically on the edge;

[0022] Based on the highest point and the lowest point, a horizontal edge is generated, and the distance between the edge of the target label and the horizontal edge is calculated;

[0023] According to the calculation result, if there is a distance, and the edge of the target label is closer to one side of the protruding part, the system determines that the fit degree has a warp; otherwise, it is not determined.

[0024] As a preferred scheme of the present application, in the data construction module, the 3D topography of the labeling area is constructed according to the area and the surface angle, and the steps include:

[0025] Based on the surface angle of the labeling area, at least 20 microfacets are divided, and each microfacet corresponds to a local curvature value;

[0026] The geometric characteristics of the surface of the labeling area are described by a double curvature model, the double curvature model includes Gaussian curvature and mean curvature, and a spatial parameter equation of the labeling area is established:

[0027] ;

[0028] In the formula, represents the parameter coordinates of the surface and , , , represents , , the three-dimensional coordinate function of the surface and .

[0029] As a preferred scheme of the present application, in the feature processing module, the front and back distances between the main fitting point and each of the auxiliary fitting points are calculated, the main fitting point is marked as , the auxiliary fitting point is marked as , =1, 2, 3, 4; and the direct spatial distance method is used for calculation.

[0030] ;

[0031] wherein, denotes the three-dimensional space distance between the main bonding point and the secondary bonding point ;

[0032] , , denotes the coordinate of the main bonding point in the labeling area;

[0033] , , denotes the coordinate of the th secondary bonding point in the labeling area.

[0034] As a preferred scheme of the present application, wherein: further comprising calculating according to the normal projection distance method:

[0035] ; wherein, denotes the projection distance of the th secondary bonding point along the normal of the curved surface to the main bonding point;

[0036] wherein, denotes the displacement vector of any two secondary bonding points and the main bonding point;

[0037] denotes the unit normal vector of the curved surface corresponding to the main bonding point, and the calculation formula is:

[0038] ;

[0039] wherein, denotes the vector cross product, denotes the vector module length, and denotes the partial derivative vector of the space parameter equation to and , and denotes the parameter coordinates of the curved surface and .

[0040] As a preferred scheme of the present application, wherein: further comprising calculating according to the arc length distance method:

[0041] ;

[0042] wherein, denotes the a curved geodesic arc length from a sub-attachment point to a main attachment point;

[0043] represents a geodesic parameter, dimensionless;

[0044] 、 、 represents a tangent vector component of the geodesic.

[0045] As a preferred scheme of the present application, wherein: according to the calculation result, a safety threshold of the front-back distance is preset, the safety threshold is one-tenth of the front-back distance, the protrusion height of the protrusion point is obtained, if the protrusion height is less than the safety threshold, the system determines that the target label has good attachment in the labeling area and there is no warping; otherwise, it is not determined.

[0046] As a preferred scheme of the present application, wherein: in the calculation unit, according to the calculation result, if there is no distance and the protrusion height is less than the safety threshold, the system determines that the target label has good attachment in the labeling area and there is no warping; otherwise, it is not determined.

[0047] At the same time, in the calculation unit, if there is no distance and the protrusion height is less than the safety threshold, the horizontal transverse edge of the upper and lower ends of the protrusion site is obtained, the horizontal transverse edge is the horizontal transverse edge of the left and right sides of the upper and lower ends of the target label, and the edge of the upper and lower ends of the target label is obtained, the distance between the edge and the horizontal transverse edge is calculated, if the distance is zero, the system determines that the target label has good attachment in the labeling area and there is no warping; otherwise, it is not determined.

[0048] On the other hand, the present application provides a method applied to a multi-angle stereoscopic imaging label 3D topography detection system as described above, comprising the following steps:

[0049] Obtaining relevant data, the relevant data including the surface angle of the labeling area of the target object and the area of the target label;

[0050] According to the area and the surface angle, the 3D topography of the labeling area is constructed, and the labeling quality of the labeling area in the constructed 3D topography is detected, the detection including the attachment of the target label in the labeling area and the target object;

[0051] A center point is given at the center of the labeling area, the center point is marked as a main attachment point, at least four sub-attachment points are given around the main attachment point based on the main attachment point, and the front-back distance between the main attachment point and each sub-attachment point is calculated in the constructed 3D topography.

[0052] collecting a protrusion of the labeling area after the target label is covered, marking the protrusion as a protrusion point, and obtaining a shape corresponding to the protrusion point;

[0053] analyzing an edge of the shape, obtaining a related feature of the protrusion point based on a result of the analysis, and the related feature including a trend of the protrusion point;

[0054] performing a correlation calculation according to the obtained trend;

[0055] determining a quality of the fit degree according to a result of the correlation calculation.

[0056] Advantages:

[0057] 1. The application can accurately describe the surface geometry characteristics by dividing the differential surface element and adopting the double curvature model (Gaussian curvature and mean curvature) to construct the 3D topography of the labeling area, and avoid the error of the traditional two-dimensional detection. For example, at least 20 differential surface elements are divided based on the surface angle to ensure the accuracy of the local curvature value; the fit degree of the target label and the target object is directly detected in the 3D topography, including the edge fit state and the local protrusion analysis, which breaks through the limitation of the traditional detection which can only judge “fit / not fit”;

[0058] 2. At least four auxiliary fit points are set around the main fit point, and the fit degree is calculated by the three-dimensional space distance method, the normal projection distance method and the arc length distance method to cover different scene requirements:

[0059] For the direct space distance method, the physical interval can be evaluated, and it is suitable for preliminary screening;

[0060] For the normal projection distance method, the influence of the curved surface can be eliminated to reflect the vertical fit state, which is the core quality index;

[0061] And for the arc length distance method, the deformation of the label on the curved surface can be quantified to adapt to the detection of flexible materials;

[0062] And according to the calculation result, a safety threshold (such as one tenth of the front and rear interval) is preset, and the double judgment of the protrusion height and the edge interval can avoid the misjudgment of a single index;

[0063] 3. By collecting the shape, edge trend and related features of the protrusion part, combining the calculation unit to analyze the interval between the label edge and the horizontal edge, the small warping (such as the fit degree is good when the protrusion height is less than the safety threshold) can be detected, and in the case of no interval and the protrusion height meeting the standard, the interval between the upper and lower edges of the target label and the horizontal edge is further verified to ensure that the edge is completely fit to eliminate local gaps. BRIEF DESCRIPTION OF DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings. Among them:

[0065] Fig. 1 It is a modular structure schematic diagram of the label 3D topography detection system of the multi-angle stereoscopic imaging of the embodiments of the present application.

[0066] Fig. 2 It is a method flow schematic diagram of the embodiments of the present application.

[0067] The figure label: 110-data acquisition module; 120-data construction module; 130-feature processing module; 140-data fusion processing module; 1401-acquisition unit; 1402-analysis unit; 1403-computation unit; 1404-judgment unit. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be described clearly and completely in the following with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the described embodiments of the present application, all other embodiments obtained by those skilled in the art belong to the scope of protection of the present application.

[0069] Because the prior art is prone to have a shielding blind area when detecting the label on a curved surface object, the detection precision of the complex curved surface label is insufficient, and the three-dimensional defects such as label warping and wrinkling cannot be detected.

[0070] Based on this, the present application provides a label 3D topography detection system and method of multi-angle stereoscopic imaging, which realizes the leap of label detection from "two-dimensional plane" to "three-dimensional stereo" by constructing 3D topography, introducing multi-dimensional fitting degree analysis method, accurately identifying local defects, and adapting to complex curved surface scene, improves the detection precision, and is compatible with industrial diversity demand.

[0071] The present application will be further specifically described by embodiments and in combination with the drawings.

[0072] Reference Figs. 1-2 For an embodiment of the present application, the embodiment provides a label 3D topography detection system of multi-angle stereoscopic imaging, comprising:

[0073] The data acquisition module 110 is configured to acquire relevant data, including a surface angle of a labeling area of a target object and an area of a target label.

[0074] In this embodiment, the surface angle of the labeling area of the target object and the area of the target label are acquired by a sensor or an image processing technology. The surface angle reflects the geometric characteristics (such as a cylindrical surface or a spherical surface) of the surface of the object, and the label area is used for subsequent size adaptation of the 3D topography.

[0075] For example, a laser triangulation sensor emits a laser beam to the surface of the object, and calculates the surface angle by using the position change of the reflected light spot on the detector.

[0076] Or a structured light 3D sensor projects a stripe or a coded grating to the surface of the object, and calculates the three-dimensional coordinates of the surface by capturing the deformed grating pattern by a camera.

[0077] The image processing technology, for example, multi-view stereo vision (MVS), captures the object from different angles by multiple cameras, and restores the three-dimensional topography by using the parallax principle.

[0078] The accurate acquisition of the surface angle and the label area is a prerequisite for the construction of the 3D topography, which avoids the distortion of subsequent analysis caused by input data errors.

[0079] And the detection of non-planar objects expands the limitations of traditional label detection systems that can only process planar objects.

[0080] The data construction module 120 is configured to construct a 3D topography of the labeling area according to the area and the surface angle, and detect the labeling quality of the labeling area in the constructed 3D topography, including the adhesion of the target label to the target object in the labeling area.

[0081] In this embodiment, the 3D topography of the labeling area is constructed according to the area and the surface angle, and the steps include:

[0082] At least 20 microfacets are divided based on the surface angle of the labeling area, and each microfacet corresponds to a local curvature value.

[0083] The geometric characteristics of the surface of the labeling area are described by a double-curvature model, which includes Gaussian curvature and mean curvature, and a spatial parameter equation of the labeling area is established:

[0084] ;

[0085] In the formula, x and y represent the parameter coordinates of the surface of the labeling area, and and , and , , represent 、 、 In the curved surface and Three-dimensional coordinate function;

[0086] Need to explain, and The parametric coordinates of the curved surface (local coordinates or curved surface coordinates) are used to locate the points on the curved surface in the two-dimensional parameter domain;

[0087] Similar to the longitude and latitude on the map, and The three-dimensional curved surface is "unfolded" into a two-dimensional parameter domain (usually a rectangular area, such as );

[0088] By adjusting and The values can traverse all points on the curved surface;

[0089] For example, on the cylindrical surface, The coordinates along the generatrix direction can be represented by The angular coordinate around the axis;

[0090] At the same time, 、 、 Together describe the position of any point on the curved surface in the global coordinate system, and the geometry of the curved surface (such as bending, twisting) is explicitly expressed through the parametric equation;

[0091] Based on the curved surface angle, the labeling area is divided into at least 20 differential surface elements, each corresponding to a local curvature value (Gaussian curvature, mean curvature);

[0092] Combined with the curvature value of the differential surface element and the double curvature model, a three-dimensional model of the labeling area is generated, and the label fitting degree is detected in the model;

[0093] The differential surface element and the double curvature model can accurately restore the curved surface geometry and avoid the distortion problem of traditional two-dimensional projection;

[0094] And in the 3D topography, the contact state of the label and the curved surface is directly detected, realizing the quantitative evaluation from "whether it fits" to "fitting degree".

[0095] The feature processing module 130, the feature processing module responds to the detection of the labeling quality, for giving a fitting point at the center of the labeling area, marking the fitting point as the main fitting point, giving at least four vice fitting points around the main fitting point based on the main fitting point, and calculating the front and rear spacing of the main fitting point and each vice fitting point in the constructed 3D topography;

[0096] The data fusion processing module 140 is responsive to the feature processing module, and is configured to analyze and judge the adhesion degree according to the calculation result. The data fusion processing module 140 comprises a collection unit 1401, an analysis unit 1402, a calculation unit 1403 and a judgment unit 1404.

[0097] The collection unit 1401 is configured to collect the protrusion of the labeling area after the target label is covered, mark the protrusion as a protrusion point, and obtain the shape corresponding to the protrusion point.

[0098] The analysis unit 1402 is configured to analyze the edge of the shape, and obtain the related features of the protrusion point based on the analysis result. The related features include the trend of the protrusion point.

[0099] The calculation unit 1403 is configured to perform related calculation according to the obtained trend.

[0100] The judgment unit 1404 is responsive to the calculation unit, and is configured to determine the quality of the adhesion degree according to the result of the related calculation.

[0101] In the calculation unit, the steps of the related calculation are as follows:

[0102] Based on the shape, the center of the protrusion point is obtained, and one protrusion part farthest from the center is collected;

[0103] The edge of the target label closest to the protrusion part is obtained;

[0104] The highest point and the lowest point of the edge are obtained, and the highest point and the lowest point are two points horizontally and vertically on the edge;

[0105] A horizontal edge is generated based on the highest point and the lowest point, and the distance between the edge of the target label and the horizontal edge is calculated;

[0106] According to the calculation result, if the distance exists and the edge of the target label is closer to one side of the protrusion part, the system determines that the adhesion degree has a warping; otherwise, it is not determined.

[0107] In the feature processing module, the front and back distances between the main adhesion point and each auxiliary adhesion point are calculated, the main adhesion point is marked as , and the auxiliary adhesion point is marked as , =1, 2, 3, 4; the following formula is obtained according to the direct spatial distance method:

[0108]

[0109] In the formula, d represents the three-dimensional spatial distance between the main adhesion point P and the auxiliary adhesion point P ; unit: mm, typical value (depends on the size of the label);​

[0110] 、 、 denotes the main bonding point coordinates in the labeling area; unit: mm;

[0111] 、 、 denotes the coordinates of the th sub-bonding point in the labeling area; unit: mm;

[0112] This formula directly calculates the straight-line distance between two points in three-dimensional space, reflecting the actual physical spacing between the label edge and the center. It is suitable for quickly evaluating the overall size adaptability of the label, but cannot distinguish the distance changes caused by curved surfaces.

[0113] Also includes according to the normal projection distance method:

[0114] ; wherein, denotes the projection distance of the th sub-bonding point along the normal of the curved surface to the main bonding point; unit: mm, typically 0-0.5 (within the thickness range of the target label);

[0115] In the formula, denotes the displacement vector of any two sub-bonding points and the main bonding point; unit: mm, dependent on the size of the target label.

[0116] denotes the unit normal vector of the curved surface corresponding to the main bonding point (the direction is perpendicular to the tangent plane of the curved surface), and the normalized modulus is 1, and the calculation formula is:

[0117] ;

[0118] In the formula, denotes the vector cross product, denotes the vector modulus, and denote the partial derivative vectors of the spatial parameter equation to and ; mm / parameter unit; and denote the parameter coordinates of the curved surface and ; the definition domain is usually [0, 1];

[0119] This formula eliminates the influence of curved surface bending and only reflects the bonding deviation of the label edge and center in the direction perpendicular to the curved surface. If exceeds the thickness of the target label (such as 0.2 mm), it indicates that there is a gap or excessive compression locally.

[0120] Also included is the calculation according to the arc length method:

[0121] ;

[0122] In the formula, represents the geodesic arc length of the secondary bonding point to the primary bonding point, unit: mm;

[0123] represents the geodesic line parameter, dimensionless;

[0124] , , represents the tangent vector component of the geodesic line, mm / parameter unit;

[0125] The formula reflects the actual stretching or compression of the label on the curved surface. For example, on a cylindrical surface, the geodesic arc length between two points will be greater than the plane distance due to the curvature of the curved surface. If is significantly greater than , it indicates that the label is over-stretched in this area.

[0126] According to the calculation result, a safety threshold of the front-back distance is preset, the safety threshold is one-tenth of the front-back distance, the protrusion height of the protrusion point is obtained, if the protrusion height is less than the safety threshold, the system determines that the target label has good adhesion in the labeling area and there is no warping; otherwise, it is not determined.

[0127] In the calculation unit, according to the calculation result, if there is no distance and the protrusion height is less than the safety threshold, the system determines that the target label has good adhesion in the labeling area and there is no warping; otherwise, it is not determined;

[0128] It should be noted that through the protrusion point feature and the edge distance analysis, small warping can be detected to avoid missed detection;

[0129] And combined with the distance and protrusion height conditions, false positives are reduced and detection reliability is improved.

[0130] At the same time, in the calculation unit, if there is no distance and the protrusion height is less than the safety threshold, the horizontal transverse edge of the upper and lower ends of the protrusion part is obtained, the horizontal transverse edge is the horizontal transverse edge of the left and right sides of the upper and lower ends of the target label, and the edge of the upper and lower ends of the target label is obtained, the distance between the edge and the horizontal transverse edge is calculated, if the distance is zero, the system determines that the target label has good adhesion in the labeling area and there is no warping; otherwise, it is not determined;

[0131] It should be emphasized that the safety threshold is set according to one-tenth of the front-back distance, which is adaptive to different label sizes; ​

[0132] When the distance is zero and less than the safety threshold, further verify whether the distance between the upper and lower edges of the label and the horizontal transverse edge is zero, to ensure that the edges are completely matched;

[0133] This dynamic threshold makes the system suitable for labels of different specifications, improving versatility;

[0134] At the same time, this edge matching verification can eliminate local gaps and meet the high-precision industrial demand (such as electronic component labeling).

[0135] Based on the above, the present application realizes the leap from "two-dimensional plane" to "three-dimensional" in label detection through the whole process innovation of "data acquisition-3D topography-analysis-determination", significantly improving the detection accuracy and industrial applicability.

[0136] The embodiment combines the above-mentioned multi-angle three-dimensional imaging label 3D topography detection system, and further proposes a working method applied to the system, as follows:

[0137] S10: Obtain relevant data, including the curved surface angle of the labeling area of the target object and the area of the target label;

[0138] S20: Construct the 3D topography of the labeling area according to the area and the curved surface angle, and detect the labeling quality of the labeling area in the constructed 3D topography, including the adhesion of the target label to the target object in the labeling area;

[0139] S30: Give a matching point at the center of the labeling area, mark the matching point as a main matching point, give at least four vice matching points around the main matching point based on the main matching point, and calculate the front and back distances between the main matching point and each vice matching point in the constructed 3D topography;

[0140] S40: Collect the protrusion of the labeling area after the target label is covered, mark the protrusion as a protrusion point, and obtain the shape corresponding to the protrusion point;

[0141] S50: Analyze the edge of the shape, obtain the relevant features of the protrusion point based on the analysis result, and the relevant features include the trend of the protrusion point;

[0142] S60: Perform relevant calculations according to the obtained trend;

[0143] S70: Determine the quality of the adhesion degree according to the result of the relevant calculation.

[0144] To sum up, by constructing high-precision 3D topography, introducing multi-dimensional fitting degree analysis method, accurately identifying local defects, and adapting to complex curved surface scenes, the label detection is realized from 'two-dimensional plane' to 'three-dimensional', which not only improves the detection accuracy, but also is compatible with industrial diversity demand and strengthens quality control.

[0145] It should be noted that the above examples are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A multi-angle stereoscopic imaging label 3D topography detection system, characterized in that, The application relates to a label adhesion quality detection system, which comprises the following parts: a data acquisition module, which is used for acquiring relevant data, including the surface angle of a label adhesion area of a target object and the area of a target label; a data construction module, which is used for constructing the 3D topography of the label adhesion area according to the area and the surface angle, and detecting the label adhesion quality of the label adhesion area in the constructed 3D topography, wherein the detection includes the adhesion degree of the target label to the target object in the label adhesion area; a feature processing module, which is used for giving a fitting point at the center of the label adhesion area and marking the fitting point as a main fitting point, giving at least four sub-fitting points around the main fitting point based on the main fitting point, and calculating the front and back distances between the main fitting point and each of the sub-fitting points in the constructed 3D topography in response to the detection of the label adhesion quality; a data fusion processing module, which is used for analyzing and judging the adhesion degree according to the calculation result in response to the feature processing module, and comprises an acquisition unit, an analysis unit, a calculation unit and a judgment unit; the acquisition unit is used for acquiring a protrusion after the label adhesion area is covered by the target label, marking the protrusion as a protrusion point, and acquiring the shape corresponding to the protrusion point; the analysis unit is used for analyzing the edge of the shape, acquiring the relevant features of the protrusion point based on the analysis result, and the relevant features include the trend of the protrusion point; the calculation unit is used for performing relevant calculation according to the acquired trend; the judgment unit is used for judging the quality of the adhesion degree according to the result of the relevant calculation in response to the calculation unit.

2. The multi-angle stereoscopic imaging label 3D topography detection system according to claim 1, wherein, In the calculation unit, the steps of the relevant calculation are as follows: the center of the protrusion point is acquired based on the shape, and a protrusion part farthest from the center is acquired; the edge of a target label closest to the protrusion part is acquired; the highest point and the lowest point of the edge are acquired, and the highest point and the lowest point are two points horizontally and vertically on the edge; a horizontal edge is generated based on the highest point and the lowest point, and the distance between the edge of the target label and the horizontal edge is calculated; according to the calculation result, if there is a distance and the edge of the target label is closer to one side of the protrusion part, the system judges that the adhesion degree has a curling phenomenon; otherwise, no judgment is made.

3. The multi-angle stereoscopic imaging label 3D topography detection system of claim 1, wherein, In the data construction module, the 3D topography of the label adhesion area is constructed according to the area and the surface angle, and the steps include: at least 20 micro-facets are divided based on the surface angle of the label adhesion area, and each micro-facet corresponds to a local curvature value; the geometric features of the curved surface of the label adhesion area are described through a double-curvature model, the double-curvature model includes Gaussian curvature and mean curvature, and a spatial parameter equation of the label adhesion area is established; ； wherein denotes a surface and the parametric coordinates of , , denotes , , the three-dimensional coordinate function of the surface and .

4. The multi-angle stereoscopic imaging label 3D topography detection system of claim 1, wherein, In the feature processing module, the front-back distance between the main attachment point and each of the sub-attachment points is calculated, the main attachment point is marked as , the sub-attachment points are marked as , = 1, 2, 3, 4; and the direct spatial distance method is used to calculate ； In the formula, represents the three-dimensional spatial distance of the main bonding point and the secondary bonding point ​ , , denotes the main attachment point coordinates in the labeling area; , , denotes the coordinates of the nth secondary attachment point in the labeling area. ​ 5. A multi-angle stereoscopic imaging label 3D topography detection system according to claim 4, wherein, the following is calculated according to the normal projection distance method: ; wherein, represents the projection distance of the i-th sub-attachment point along the normal of the curved surface to the main attachment point; and represents the projection distance of the i-th sub-attachment point along the normal of the curved surface to the main attachment point. wherein denotes the displacement vector of any two sub-attachment points from the main attachment point; a unit normal vector of a curved surface corresponding to the main fitting point, and the calculation formula is: ； wherein denotes the vector cross product, denotes the vector length, and denotes the partial derivative vector of the spatial parametric equation with respect to and and denotes the parametric coordinates of the surface and .​ 6. A multi-angle stereoscopic imaging label 3D topography detection system according to claim 4, wherein, the following is calculated according to the arc length distance method: ； In the formula, denotes the geodesic arc length of the curved surface from the secondary bonding point to the primary bonding point; represents a geodesic parameter, dimensionless; , , denotes the tangential vector component of the geodesic.

7. A multi-angle stereoscopic imaging label 3D topography detection system according to any one of claims 4-6, characterized in that, According to the result of the calculation, a safety threshold of the front-back distance is preset, the safety threshold being one-tenth of the front-back distance, a protrusion height of the protrusion point is obtained, and if the protrusion height is less than the safety threshold, the system determines that the target label has good adhesion to the labeling area and does not have warping; otherwise, the determination is not made.

8. The multi-angle stereoscopic imaging label 3D topography detection system of claim 7, wherein, In the calculation unit, according to the result of the calculation, if there is no distance and the protrusion height is less than the safety threshold, the system determines that the target label has good adhesion to the labeling area and does not have warping; otherwise, the determination is not made. Meanwhile, in the calculation unit, if there is no distance and the protrusion height is less than the safety threshold, the horizontal lateral edges of the upper and lower ends of the protrusion site are obtained, the horizontal lateral edges being the horizontal lateral edges of the left and right sides of the upper and lower ends of the target label, and the edges of the upper and lower ends of the target label are obtained, the distance between the edges and the horizontal lateral edges is calculated, and if the distance is zero, the system determines that the target label has good adhesion to the labeling area and does not have warping; otherwise, the determination is not made.

9. The method applied to the multi-angle stereoscopic imaging label 3D topography detection system of claim 1, characterized in that, The method comprises the following steps: Obtaining relevant data, the relevant data including the surface angle of the labeling area of the target object and the area of the target label; According to the area and the surface angle, the 3D topography of the labeling area is constructed, and the labeling quality of the labeling area is detected in the constructed 3D topography, the detection including the adhesion of the target label to the target object in the labeling area; A fitting point is given at the center of the labeling area, the fitting point is marked as a main fitting point, at least four sub-fitting points are given around the main fitting point based on the main fitting point, and the front-back distance between the main fitting point and each of the sub-fitting points is calculated in the constructed 3D topography; The protrusion of the labeling area after being covered by the target label is collected, the protrusion is marked as a protrusion point, and the shape corresponding to the protrusion point is obtained; The edge of the shape is analyzed, and relevant features of the protrusion point are obtained based on the result of the analysis, the relevant features including the trend of the protrusion point; According to the obtained trend, relevant calculations are performed; According to the result of the relevant calculations, the quality of the adhesion is determined.

Images