Paperboard warping degree detection method and device

Abstract

The invention discloses a paperboard warping degree detection method and device, and relates to the technical field of paperboard surface flatness detection. The paperboard warping degree detection method comprises the steps of image acquisition, image preprocessing, reflection feature extraction, edge detection, morphological operation, contour analysis and warping degree calculation. Image acquisition: acquiring image information of the paperboard through a light source and a camera positioned above the to-be-detected paperboard; preprocessing the image, converting the collected image information into a grey-scale map, removing noise, and performing histogram equalization to enhance the contrast; reflection feature extraction: calculating brightness gradients in horizontal and vertical directions to obtain a reflection change intensity graph, and carrying out binarization processing to obtain a reflection feature graph; edge detection: extracting a basic edge by adopting a Canny algorithm, carrying out morphological closing operation on the reflection characteristic pattern, and then carrying out bit OR operation with the basic edge to obtain an enhanced edge; and performing morphological operation, and sequentially performing corrosion operation of different structural elements on the enhanced edge to obtain an optimized edge image.

Description

TECHNICAL FIELD

[0001] The application relates to the paperboard surface flatness detection technical field, in particular to a paperboard warping degree detection method and device. BACKGROUND

[0002] In the paperboard production process, warping is a common quality defect, which directly affects the subsequent processing and product quality. The traditional paperboard warping detection mainly relies on manual visual inspection, which has problems such as low efficiency, strong subjectivity and insufficient precision.

[0003] Therefore, some devices for detecting the warping degree of hard paperboard have appeared, such as a plurality of detection terminals arranged in a row, and then the paperboard is conveyed through the detection terminals by a conveying belt. When the paperboard warps, the detection terminals will be displaced, so as to judge the surface state of the paperboard. However, this method has low detection efficiency, and the result is inaccurate due to the design of a certain gap between the detection terminals and the paperboard. SUMMARY

[0004] The application provides a paperboard warping degree detection method and device, which provides a new more sensitive and accurate paperboard warping degree detection scheme.

[0005] The application provides a paperboard warping degree detection method, which comprises image acquisition, image preprocessing, reflection feature extraction, edge detection, morphological operation, contour analysis and warping degree calculation. The image acquisition acquires image information of the paperboard through a light source and a camera located above the paperboard to be detected. The image preprocessing converts the acquired image information into a gray image, removes noise and performs histogram equalization to enhance contrast. The reflection feature extraction calculates the brightness gradient in the horizontal and vertical directions to obtain a reflection change intensity map, and obtains a reflection feature map through binarization processing. The edge detection extracts the basic edge by using the Canny algorithm, and performs a morphological closing operation on the reflection feature map, and then performs a bit or operation with the basic edge to obtain an enhanced edge. The morphological operation performs corrosion operations on the enhanced edge with different structure elements in sequence to obtain an optimized edge image. The contour analysis extracts an outer contour from the optimized edge image, selects the contour with the largest area as the paperboard contour, and then detects the main straight lines in the contour by Hough transformation. The warping degree calculation selects two longest detection straight lines as reference lines, calculates the distances of all contour points to the two reference lines, calculates the weighted distance according to the reflection intensity of the corresponding positions in the reflection feature map, takes the maximum side length of the contour boundary box as the length of the paperboard, and outputs the warping degree as the ratio of the maximum weighted distance to the length of the paperboard.

[0006] The application forms a full-link technical scheme from image to quantitative result through image acquisition, preprocessing, reflection feature extraction, edge detection, morphological operation, contour analysis and warping degree calculation, realizes high-precision and automatic detection, replaces manual subjective judgment, and can output accurate paperboard warping degree detection results.

[0007] The problems that the traditional edge detection is not sensitive to slight warping and is easily disturbed by surface texture are solved; the reflection difference caused by warping of the paperboard is captured through reflection feature extraction (based on brightness gradient), and the discrimination of edge information is enhanced; the reflection intensity is associated with geometric deviation through weighted distance calculation, so that the warping quantification is more in line with the actual physical form.

[0008] In some embodiments of the application, the result classification of the warping degree is further included, the detection result is classified into qualified, slight warping and serious warping according to the warping degree value, and the detection result is output.

[0009] The result classification of the warping degree and the output of the results of qualified, slight warping and serious warping can make the output value directly connect with the production quality standard, facilitate the production line to quickly judge the product grade, guide the subsequent sorting, rework and other operations, and improve the quality control efficiency and adapt to the industrial batch detection demand.

[0010] In some embodiments of the application, the morphological operation includes a corrosion operation of a 3*3 structure element and a dilation operation of a 5*5 structure element on the enhanced edge.

[0011] The problem of unstable edge optimization effect caused by ambiguous morphological parameters is solved; the 3*3 corrosion can accurately remove the tiny noise edge (such as paperboard surface dust and texture interference), and the 5*5 dilation can effectively connect the broken edge caused by noise or uneven illumination, balancing the "noise removal" and "edge integrity"; the fixed parameters make the method highly reproducible, facilitating standardized deployment in different scenarios.

[0012] In some embodiments of the application, the reflection feature extraction includes: calculating the brightness gradient in the horizontal and vertical directions; generating a reflection change intensity map based on the gradient amplitude; and performing threshold processing on the reflection change intensity map to highlight the significant reflection change area.

[0013] The problem of ambiguous reflection feature extraction logic and difficult engineering implementation is solved; the brightness gradient directly reflects the angle change of the paperboard surface caused by warping (the greater the gradient, the more significant the reflection difference), the threshold processing highlights the core warping area, providing accurate basis for subsequent edge enhancement; the association between the reflection feature and the physical warping is more explicit, and the robustness of feature extraction is improved.

[0014] In some embodiments of the present application, the calculation of the warping degree is: weighted distance = basic distance + basic distance x reflection intensity + basic distance x gradient sensitivity coefficient; the reflection intensity is the reflection characteristic value normalized to the range of 0-1; the gradient sensitivity coefficient is in the range of 1.0-1.5.

[0015] The problem that the traditional geometric distance calculation is not sensitive to the "reflection abnormal area" is solved; by weighting the reflection intensity, the deviation contribution of the warping core area (rapid reflection change) is amplified, so that the quantification result of the slight warping (such as 0.1mm level) is more significant; the gradient sensitivity coefficient can be adjusted according to the paperboard material (such as surface smoothness), so as to enhance the adaptability of the method to different types of paperboard.

[0016] A paperboard warping degree detection device adopts the paperboard warping degree detection method described above to detect the paperboard warping degree, and the paperboard warping degree detection device comprises a support main body, a conveying belt, a baffle, a diffuse light source, a first support, a first image capture unit, a second support, and a second image capture unit.

[0017] The support main body has a placement area for containing the paperboard to be detected, the conveying belt is arranged on the support main body and located at one side of the placement area, and the conveying belt is used to move the paperboard to be detected along the conveying direction of the conveying belt; the baffle is arranged on the support main body and located at the conveying end of the conveying belt, and the baffle is used to shield the paperboard to be detected; the diffuse light source is arranged on the baffle, and the light of the diffuse light source is directed towards the paperboard to be detected; the first support extends perpendicular to the conveying direction of the conveying belt, and the first support is arranged at one side of the placement area and located above the conveying belt; the first image capture unit is arranged on the first support; the second support extends along the conveying direction of the conveying belt, and the second support is arranged at one side of the conveying belt and located above the conveying belt; the second image capture unit is arranged on the second support and spaced apart from the first image capture unit projected on the surface of the paperboard to be detected along the direction perpendicular to the conveying direction of the conveying belt.

[0018] The conveying belt realizes continuous conveying of the paperboard, which is suitable for pipeline operation; the double supports and double image capture elements realize multi-angle image acquisition in the "vertical conveying direction" and "along the conveying direction", cover the edge and middle area of the paperboard, and avoid missing detection at a single angle; the diffuse light source provides uniform illumination, reduces the interference of specular reflection, and cooperates with the baffle to fix the position of the paperboard, thereby ensuring the stability of image acquisition; the paperboard warping degree detection device can adapt to and adapt to the production line operation, and improve the accuracy and efficiency of paperboard detection.

[0019] In some embodiments of the present application, the paperboard warping degree detection device further comprises a paperboard blocking plate, the paperboard blocking plate extends along a direction perpendicular to a plane on which the conveying belt is located, the paperboard blocking plate is located on one side of the placement area, the paperboard blocking plate is located above the conveying belt, and the paperboard blocking plate is located on a side of the first support away from the blocking plate; a gap is arranged between the paperboard blocking plate and the conveying belt, the gap between the paperboard blocking plate and the conveying belt is greater than the thickness of one paperboard to be detected and less than the thickness of two paperboards to be detected.

[0020] The paperboard blocking plate can force only one paperboard to pass at a time, ensure the uniqueness of the image acquisition object, avoid edge confusion in the overlapping area, and improve the accuracy of profile extraction; the structure is simple, and single separation can be achieved without complex sensors, thereby reducing the cost of the device.

[0021] In some embodiments of the present application, the paperboard warping degree detection device further comprises two vertical sliding rails, the two vertical sliding rails are arranged on the support main body, the two vertical sliding rails are located on both sides of the conveying belt, and the two ends of the blocking plate are slidingly arranged on the two vertical sliding rails; when the paperboard to be detected is detected to be qualified, the vertical sliding rails drive the blocking plate to slide upward, and the paperboard to be detected slides out of the conveying belt.

[0022] The cooperation of the vertical sliding rails and the blocking plate can realize automatic output and solve the efficiency problem of manual sorting after detection; qualified paperboards automatically flow out, and unqualified paperboards are temporarily stored, thereby reducing manual intervention and improving the degree of automation of the production line; the sliding rail driving response is fast, can match a high-speed conveying belt (such as 60 meters / minute), and is suitable for high-capacity requirements.

[0023] In some embodiments of the present application, a space for passing the paperboard to be detected is arranged on the second support, the conveying belt is provided in a plurality of forms, and the plurality of conveying belts are parallel and spaced apart; the paperboard warping degree detection device further comprises a plurality of discharge rollers, the discharge rollers are arranged on one side of the blocking plate, the plurality of discharge rollers are respectively arranged between the plurality of conveying belts, the rotation direction of the discharge rollers is perpendicular to the conveying direction of the conveying belt, and the discharge rollers can send the unqualified paperboard to be detected out of the space in the second support.

[0024] The plurality of conveying belts are spaced apart, the discharge rollers are located in the gap and have a rotation direction perpendicular to the conveying direction, and are used to discharge unqualified paperboards. The discharge rollers are perpendicular to the conveying belt, can discharge unqualified paperboards from the side, separate from the forward flow of qualified paperboards, realize fully automatic sorting, and avoid paperboard deformation in conveying through the design of the plurality of conveying belts, thereby ensuring the stability of the shape during detection.

[0025] In some embodiments of the present application, the distance between the diffuse light source and the upper surface of the paperboard to be detected in a direction perpendicular to the conveying plane of the conveying belt is greater than 10 mm, the distance between the first image capture unit and the diffuse light source is greater than 10 mm, and the distance between the second image capture unit and the diffuse light source is greater than 10 mm.

[0026] The distance between the diffuse light source and the paperboard and the distance between the image capturing element and the light source are greater than 10 mm, so that the diffuse light source light uniformly covers the paperboard surface through sufficient spacing, avoiding local overexposure or underexposure; at the same time, the image capturing element (camera) ensures that the range of view is complete, the focus is clear, and the edge distortion is reduced; provides clear parameters for device installation, ensures the consistency of different batches of devices, and reduces the debugging difficulty. BRIEF DESCRIPTION OF DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the technical solutions of the present application, and constitute a part of the specification, and are used together with the embodiments of the present application to explain the technical solutions of the present application, and do not constitute a limitation on the technical solutions of the present application.

[0028] Figure 1 A flowchart of a paperboard warping degree detection method provided by the embodiments of the present application.

[0029] Figure 2 A structural schematic diagram of a paperboard warping degree detection device provided by the embodiments of the present application.

[0030] The accompanying drawings are used to provide a further understanding of the technical solutions of the present application, and constitute a part of the specification, and are used together with the embodiments of the present application to explain the technical solutions of the present application, and do not constitute a limitation on the technical solutions of the present application. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0032] It should be noted that all directionality indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship, movement condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directionality indications also change accordingly.

[0033] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0034] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "connected" and "linked" should be interpreted broadly, for example, as a fixed connection, a detachable connection, or an integral connection. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, when describing pipelines, the terms "connected" and "linked" as used in this application have the meaning of establishing electrical connection. The specific meaning needs to be understood in conjunction with the context.

[0035] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0036] Warping is a common quality defect in cardboard production, directly affecting subsequent processing and product quality. Traditional cardboard warping detection mainly relies on manual visual inspection, which suffers from low efficiency, high subjectivity, and insufficient accuracy.

[0037] Therefore, some devices for detecting the warpage of cardboard have emerged. For example, multiple detection terminals are arranged in a row, and the cardboard is passed through the detection terminals by a conveyor belt. When the cardboard warps, the detection terminals will be displaced, thereby judging the surface condition of the cardboard. However, this method has low detection efficiency, and the results are inaccurate because a certain gap is usually designed between the detection terminals and the cardboard.

[0038] Therefore, please refer to Figure 1 This application provides a method for detecting the warpage of cardboard, including image acquisition 01, image preprocessing 02, reflection feature extraction 03, edge detection 04, morphological operation 05, contour analysis 06, and warpage calculation 07.

[0039] Image Acquisition 01: Image information of the cardboard is acquired using a light source and camera located above the cardboard to be inspected.

[0040] Image preprocessing 02 converts the acquired image information into a grayscale image, removes noise, and then performs histogram equalization to enhance contrast.

[0041] Reflection feature extraction 03, calculate the brightness gradient in horizontal and vertical directions to get the reflection change intensity map, and get the reflection feature map through binarization processing.

[0042] Edge detection 04, extract the basic edge using Canny algorithm, and perform morphological closing operation on the reflection feature map, then perform bit or operation with the basic edge to get the enhanced edge. Morphological operation 05, perform corrosion operation on the enhanced edge with different structure elements to get the optimized edge image.

[0043] Contour analysis 06, extract the outer contour from the optimized edge image, select the largest area contour as the paperboard contour, and then detect the main straight line in the contour through Hough transform.

[0044] Warping degree calculation 07, select the two longest detected straight lines as reference lines, calculate the distance of all contour points to the two reference lines, calculate the weighted distance according to the reflection intensity of the corresponding position in the reflection feature map, and the paperboard length is the maximum side length of the contour bounding box. The warping degree output is the ratio of the maximum weighted distance to the paperboard length.

[0045] Please refer to Figure 1 The present application forms a "from image to quantitative result" full-link technical solution through image acquisition 01, preprocessing, reflection feature extraction 03, edge detection 04, morphological operation 05, contour analysis 06, and warping degree calculation 07, realizes high-precision and automatic detection, replaces manual subjective judgment, and can output accurate paperboard warping degree detection results.

[0046] Solves the problem that traditional edge detection 04 is not sensitive to subtle warping and is easily disturbed by surface texture; through reflection feature extraction 03 (based on brightness gradient) to capture the reflection difference of paperboard caused by warping, and enhance the discrimination of edge information; weighted distance calculation relates the reflection intensity and geometric deviation, so that the warping quantization is more in line with the actual physical form.

[0047] For example, image acquisition 01 is to collect image information of the paperboard through a light source and a camera located above the paperboard to be detected. A diffuse light source is used to provide uniform illumination to avoid glare interference caused by direct strong light; the camera is vertically or obliquely photographed to ensure complete capture of the paperboard surface features.

[0048] For example, a 45° angle arranged ring LED diffuse light source 4 is used, combined with a 200 million pixel industrial camera, to collect 1920x1080 resolution images at a working distance of 50 cm.

[0049] For example, the image preprocessing 02 is to convert the collected image information into a grayscale image, remove noise, and then perform histogram equalization to enhance the contrast. Grayscale reduces the amount of calculation, Gaussian filter removes high-frequency noise, and histogram equalization enhances the detail difference between light and dark areas, laying a foundation for subsequent feature extraction.

[0050] For example, a 5x5 Gaussian kernel is used to filter the color image, and the cv2.equalizeHist() function is used to enhance the contrast of the grayscale image, so that the grayscale difference between the paperboard edge and the background is increased from 10 to more than 30.

[0051] For example, the reflection feature extraction 03 is to calculate the brightness gradient in the horizontal and vertical directions to obtain a reflection change intensity map, and through binarization processing, a reflection feature map is obtained. The brightness gradient reflects the change of the reflection angle of the paperboard surface due to warping, and the greater the gradient, the more obvious the warping; binarization highlights the significant reflection change area (warping core area).

[0052] For example, the Sobel operator is used to calculate the horizontal (Gx) and vertical (Gy) gradients, and the gradient amplitude map is generated by √(Gx²+Gy²), and the threshold value is set to 30 for binarization to obtain the reflection feature map.

[0053] For example, the edge detection 04 is to extract the basic edge using the Canny algorithm, and perform morphological closing operation on the reflection feature map, and then perform bit or operation with the basic edge to obtain the enhanced edge. The Canny algorithm extracts the complete basic edge, and the reflection feature map is fused with the basic edge after morphological closing (filling small gaps), and the edge signal of the warping area is strengthened.

[0054] For example, the Canny algorithm sets the low threshold value to 50 and the high threshold value to 150, and performs closing operation on the reflection feature map with a 3x3 structure element, and the fusion of the edge continuity is improved by 40%.

[0055] For example, the morphological operation 05 is to perform erosion and dilation operations on the enhanced edge with different structure elements in sequence to obtain the optimized edge image. Erosion removes small noise edges (such as surface texture), and dilation connects the broken effective edges, balancing the purity and integrity of the edges.

[0056] For example, first, a 3x3 rectangular structure element is used for erosion 1 time to remove noise with a diameter <3 pixels; then a 5x5 structure element is used for dilation 1 time to connect edges with a gap <5 pixels.

[0057] For example, the contour analysis 06 extracts the outer contour from the optimized edge image, selects the contour with the largest area as the paperboard contour, and then detects the main straight line in the contour through Hough transform. The outer contour reflects the overall shape of the paperboard, and the largest area contour filters out background interference; Hough transform identifies the ideal straight line (flat paperboard edge), which provides a reference for the warpage calculation 07.

[0058] For example, the contour is extracted using the cv2.findContours() function, and the contour with an area greater than 10000 pixels is selected; the Hough transform sets the minimum line length to 100 pixels to detect the two long side straight lines of the paperboard.

[0059] For example, the warpage calculation 07 selects the two longest detected straight lines as reference lines, calculates the distance of all contour points to the two reference lines, calculates the weighted distance according to the reflection intensity of the corresponding position in the reflection feature map, and outputs the warpage as the ratio of the maximum weighted distance to the length of the paperboard. The weighted distance amplifies the deviation contribution of the area where the reflection is obvious (the warpage core), making it easier to quantify subtle warpage; the warpage is expressed as a relative value, eliminating the influence of paperboard size differences.

[0060] For example, the basic distance of a certain contour point is 0.5 mm, the corresponding reflection intensity is 0.8 (after normalization), and the gradient sensitivity coefficient is 1.2. The weighted distance = 0.5 + 0.5 x 0.8 + 0.5 x 1.2 = 1.5 mm; the length of the paperboard is 100 mm, and the warpage = 1.5 / 100 = 1.5%.

[0061] For example, the result classification 08 can classify the detection results into qualified, slightly warped, and severely warped according to the warpage value, and output the detection results. This is connected with the production standard, which facilitates quick determination of product grade and guides the sorting process.

[0062] For example, set the threshold values 0.01 and 0.03, and the warpage < 0.01 is qualified, 0.01-0.03 is slightly warped, and ≥ 0.03 is severely warped.

[0063] Please refer to Figure 1 In some examples, the result classification 08 of the warpage is also included, which classifies the detection results into qualified, slightly warped, and severely warped according to the warpage value, and outputs the detection results.

[0064] The result classification 08 of the warpage and the output of the qualified, slightly warped, and severely warped results can make the output value directly connect with the production quality standard, which facilitates the quick determination of product grade on the production line and guides the subsequent sorting, rework, and other operations; at the same time, it improves the quality control efficiency and adapts to the industrial batch detection demand.

[0065] Please refer to Figure 1In some examples, the morphological operation 05 includes a 3x3 structure element erosion operation and a 5x5 structure element dilation operation on the enhanced edge.

[0066] The problem of unstable edge optimization effect caused by ambiguous morphological parameters is solved. The 3x3 erosion operation can accurately remove small noise edges (such as dust and texture interference on the surface of the paperboard), and the 5x5 dilation operation can effectively connect broken edges caused by noise or uneven illumination, balancing "noise removal" and "edge integrity". The fixed parameters make the method highly reproducible and facilitate standardized deployment in different scenarios.

[0067] In some examples, the reflection feature extraction 03 includes: calculating the brightness gradient in the horizontal and vertical directions; generating a reflection change intensity map based on the gradient amplitude; and performing threshold processing on the reflection change intensity map to highlight the significant reflection change area.

[0068] The problem of ambiguous logic and difficult engineering implementation of the reflection feature extraction 03 is solved. The brightness gradient directly reflects the angle change caused by warping on the surface of the paperboard (the greater the gradient, the more significant the reflection difference), and the threshold processing highlights the core warping area, providing accurate basis for subsequent edge enhancement. The association between the reflection feature and the physical warping is more explicit, and the robustness of feature extraction is improved.

[0069] In some examples, the calculation of the weighted distance in the warping degree calculation 07 is: weighted distance = base distance + base distance x reflection intensity + base distance x gradient sensitivity coefficient; the reflection intensity is the reflection feature value normalized to the range of 0-1; and the gradient sensitivity coefficient has a value range of 1.0-1.5.

[0070] The problem that the traditional geometric distance calculation is not sensitive to "abnormal reflection areas" is solved. By weighting the reflection intensity, the deviation contribution of the core warping area (with dramatic reflection change) is amplified, so that the quantification result of subtle warping (such as 0.1mm level) is more significant. The gradient sensitivity coefficient can be adjusted according to the paperboard material (such as surface smoothness), enhancing the adaptability of the method to different types of paperboard.

[0071] An embodiment one of the paperboard warping degree detection method is provided in the present application, which is suitable for ordinary kraft paper.

[0072] Image acquisition 01: A 45° ring-shaped diffuse light source 4 (800 lux) is used, and a 16 million pixel camera is used to capture images at a distance of 60 cm.

[0073] Preprocessing: 5x5 Gaussian filtering and histogram equalization to expand the gray scale range to 0-255.

[0074] Reflection feature extraction 03: Sobel operator is used to calculate the gradient, and a threshold of 25 is used to generate a reflection feature map.

[0075] Edge detection 04: Canny threshold (40, 120), after 3x3 close operation, the reflection feature map is fused with the basic edge.

[0076] Morphological operation 05: 3x3 erosion 1 time, 5x5 expansion 1 time.

[0077] Contour analysis 06: extract the maximum contour, and the Hough transform is used to detect the longest two straight lines (length > 200 pixels).

[0078] Warpage calculation 07: gradient sensitivity coefficient 1.1, paperboard length is the long side of the bounding box, and the warpage is output.

[0079] Classification: threshold 0.01 and 0.03, output qualified / slight / severe warpage result.

[0080] The present application provides a second embodiment, a paperboard warpage detection method, which is suitable for corrugated paperboard.

[0081] Image acquisition 01: increase the light source intensity to 1500 lux, shorten the camera exposure time to 10 ms, and reduce the texture interference.

[0082] Preprocessing: 7x7 Gaussian filter, remove corrugated texture noise.

[0083] Reflection feature extraction 03: increase the threshold to 40, filter the false reflection changes caused by texture.

[0084] Edge detection 04: Canny threshold (60, 180), enhance the edge anti-interference ability.

[0085] Morphological operation 05: 3x3 erosion 2 times, remove more texture edges.

[0086] Contour analysis 06: the minimum line length of the Hough transform is set to 300 pixels, which is suitable for the large size of the corrugated paperboard.

[0087] Warpage calculation 07: the gradient sensitivity coefficient is increased to 1.4, and the real warpage signal is amplified.

[0088] Classification: same as embodiment one, add "recheck" level (warpage 0.025-0.03).

[0089] In order to facilitate industrial production and improve the efficiency and accuracy of paperboard warpage detection, the present application also provides a paperboard warpage detection device, which uses the above-mentioned paperboard warpage detection method for paperboard warpage detection. Please refer to Figure 2 , the paperboard warpage detection device comprises a support main body 1, a conveying belt 2, a baffle 3, a diffuse light source 4, a first support 5, a first image capture unit 51, a second support 6, and a second image capture unit 61.

[0090] Please refer toFigure 2 The support body 1 has a placing area 11 for containing the paperboard to be detected, and the conveying belt 2 is arranged on the support body 1 and located at one side of the placing area 11, and the conveying belt 2 is used to move the paperboard to be detected along the conveying direction of the conveying belt 2.

[0091] The baffle 3 is arranged on the support body 1 and located at the conveying end of the conveying belt 2, and the baffle 3 is used to shield the paperboard to be detected.

[0092] The diffuse light source 4 is arranged on the baffle 3, and the light of the diffuse light source 4 is directed towards the paperboard to be detected.

[0093] The first support 5 extends along the direction perpendicular to the conveying direction of the conveying belt 2, and the first support 5 is arranged at one side of the placing area 11 and located above the conveying belt 2.

[0094] The first image capturing unit 51 is arranged on the first support 5.

[0095] The second support 6 extends along the conveying direction of the conveying belt 2, and the second support 6 is arranged at one side of the conveying belt 2 and located above the conveying belt 2.

[0096] The second image capturing unit 61 is arranged on the second support 6, and the first image capturing unit 51 and the second image capturing unit 61 are spaced apart on the surface of the paperboard to be detected along the direction perpendicular to the conveying direction of the conveying belt 2.

[0097] Please refer to Figure 2 The conveying belt 2 can realize continuous conveying of the paperboard and adapt to the flow line operation; the double support + double image capturing element realizes multi-angle image acquisition 01 in the direction perpendicular to the conveying direction and along the conveying direction, covers the edge and middle area of the paperboard, and avoids single-angle missing detection; the diffuse light source 4 provides uniform illumination, reduces mirror reflection interference, cooperates with the baffle 3 to fix the position of the paperboard, and ensures the stability of the image acquisition 01; and the paperboard warping degree detection device can adapt to and adapt to the production line operation, and improve the accuracy and detection efficiency of the paperboard detection.

[0098] For example, the support body 1 can have a placing area 11 for containing the paperboard to be detected. A stable mounting base is provided, and the placing area 11 ensures the position fixed during paperboard detection. For example, an aluminum alloy frame is adopted, the placing area 11 is a flat table of 500*500mm, and the surface is pasted with anti-skid rubber.

[0099] For example, the conveying belt 2 is arranged on the support body 1 and located at one side of the placing area 11, and is used to move the paperboard to be detected along the conveying direction. The conveying belt 2 can realize continuous conveying of the paperboard and adapt to the flow line operation of the production line. For example, the conveying belt 2 can adopt three synchronous belts with a width of 50mm, a spacing of 100mm, and a running speed of 30m / min.

[0100] For example, the baffle 3 is arranged on the support body 1 and located at the end of the conveying belt 2, for shielding the paperboard to be detected. The baffle 3 temporarily blocks the paperboard 7, so as to ensure the stability of the paperboard position during detection. For example, the baffle 3 can be made of acrylic plate, with a height of 100 mm and a gap of 5 mm between the bottom and the surface of the conveying belt 2.

[0101] For example, the diffuse light source 4 is arranged on the baffle 3 and the light is directed towards the paperboard to be detected. The diffuse light source 4 is used to provide uniform diffuse reflection lighting, so as to enhance the difference in surface reflection characteristics of the paperboard. For example, the diffuse light source 4 can be a ring-shaped LED diffuse light source 4 with a color temperature of 6500K and a light intensity of 1000 lux, and a diameter of 150 mm.

[0102] For example, the first support 5 and the first image capturing element are arranged on one side of the placement area 11 and above the conveying belt 2 in a direction perpendicular to the conveying belt 2. The first image capturing element is arranged on the first support 5. The first support 5 and the first image capturing element can realize image acquisition 01 in a direction perpendicular to the conveying direction, covering the edge features in the width direction of the paperboard.

[0103] For example, the first support 5 can be an aluminum profile with a length of 800 mm, and the first image capturing element is a 12 million pixel camera with an 8 mm lens and a shooting direction perpendicular to the conveying belt 2.

[0104] For example, the second support 6 and the second image capturing unit 61 are arranged on one side of the conveying belt 2 and above the conveying belt 2 in a direction along the conveying belt 2. The second image capturing unit 61 is arranged on the second support 6 and is spaced apart from the first image capturing element in a direction perpendicular to the conveying direction. The second support 6 and the second image capturing unit 61 can realize image acquisition 01 in a direction along the conveying direction, covering the edge features in the length direction of the paperboard, and forming multi-angle acquisition with the first camera.

[0105] For example, the second support 6 has a length of 1000 mm, and the second image capturing unit 61 is a camera of the same type as the first camera and is spaced apart from the first camera by 300 mm in the width direction.

[0106] For example, the paperboard blocking plate 7 extends in a direction perpendicular to the plane of the conveying belt 2, is located on one side of the placement area 11 and on the side of the first support 5 away from the baffle 3, and has a gap with the conveying belt 2 greater than the thickness of one paperboard and less than the thickness of two paperboards. The paperboard blocking plate 7 can prevent multiple paperboards from stacking into the detection area, and ensure that the object of single detection is unique.

[0107] For example, the paperboard blocking plate 7 is a stainless steel plate, and the gap between the paperboard blocking plate 7 and the conveying belt 2 is 0.5 mm (when the thickness of the paperboard is 0.3 mm).

[0108] Exemplarily, two vertical sliding rails 8 are arranged on the support body 1 and located on both sides of the conveying belt 2, and the baffle 3 is slidingly arranged on the sliding rails; when qualified, the sliding rails drive the baffle 3 to rise and release. The vertical sliding rails 8 can realize automatic lifting of the baffle 3 and complete automatic release of the qualified paperboard in cooperation with the detection result. For example, the vertical sliding rails 8 can be selected to have a sliding rail stroke of 50 mm, driven by a servo motor, and the response time is <0.5 s.

[0109] Exemplarily, the discharge roller 9 is arranged on one side of the baffle 3 and located between the plurality of conveying belts 2, and the rotation direction is perpendicular to the conveying belt 2, which is used to discharge the unqualified paperboard. The discharge roller 9 can discharge the unqualified paperboard from the side to realize automatic sorting. For example, three rubber discharge rollers 9 can be selected, with a diameter of 30 mm and a rotating speed of 100 r / min, and the gap between the discharge roller 9 and the conveying belt 2 is 1 mm.

[0110] Exemplarily, the distance between the diffuse light source 4 and the upper surface of the paperboard is >10 mm, and the distance between the image capturing element and the light source is >10 mm. Among them, the first image capturing element and the second image capturing element can be the same height or different heights. The height difference can avoid uneven lighting or image blur caused by too close distance, and ensure the collection quality. For example, the diffuse light source 4 is 10 mm away from the paperboard, and the camera is 10 mm away from the light source.

[0111] Please refer to Figure 2 In some examples, the paperboard warping degree detection device further comprises a paperboard blocking plate 7, the paperboard blocking plate 7 extends along a plane perpendicular to the plane on which the conveying belt 2 is located, the paperboard blocking plate 7 is located on one side of the placement area 11, the paperboard blocking plate 7 is located above the conveying belt 2, and the paperboard blocking plate 7 is located on the side of the first support 5 away from the baffle 3; a gap is provided between the paperboard blocking plate 7 and the conveying belt 2, and the gap between the paperboard blocking plate 7 and the conveying belt 2 is greater than the thickness of one paperboard to be detected and less than the thickness of two paperboards to be detected.

[0112] The paperboard blocking plate 7 can force only one paperboard to pass at a time, ensure that the image acquisition 01 object is unique, avoid edge confusion in the overlapping area, and improve the accuracy of profile extraction; the structure is simple, and single separation can be realized without complex sensors, thereby reducing the cost of the device.

[0113] Please refer to Figure 2 In some examples, the paperboard warping degree detection device further comprises two vertical sliding rails 8, the two vertical sliding rails 8 are arranged on the support body 1, the two vertical sliding rails 8 are located on both sides of the conveying belt 2, and the two ends of the baffle 3 are slidingly arranged on the two vertical sliding rails 8; when the paperboard to be detected is qualified, the vertical sliding rails 8 drive the baffle 3 to slide upward, and the paperboard to be detected slides out of the conveying belt 2.

[0114] The combination of vertical slide rail 8 and baffle 3 enables automatic output, solving the efficiency problem of manual sorting after inspection; qualified cardboard flows out automatically, while unqualified cardboard is temporarily stored, reducing manual intervention and improving the automation level of the production line; the slide rail drive has a fast response and can be matched with high-speed conveyor belt 2 (such as 60 meters / minute) to meet high capacity requirements.

[0115] Please refer to Figure 2 In some examples, the second support 6 is provided with space for the passage of the cardboard to be tested. Figure 2 (As shown in area A), multiple conveyor belts 2 are configured, and the multiple conveyor belts 2 are parallel and spaced apart; the cardboard warpage detection device also includes multiple rollers 9 ( Figure 2 (The number 9 represents the installation area only and is not shown in the specific structural diagram). The roller 9 is set on one side of the baffle 3. Multiple rollers 9 are respectively set between multiple conveyor belts 2. The rotation direction of the roller 9 is perpendicular to the conveying direction of the conveyor belt 2. The roller 9 can send out the unqualified paperboard to be tested from the space inside the second bracket 6 when it is blocked in front of the baffle 3.

[0116] Multiple conveyor belts 2 are spaced apart, with rollers 9 located in the gaps and rotating perpendicular to the conveying direction to discharge defective cardboard. The rollers 9 are perpendicular to the direction of the conveyor belts 2, which can discharge defective cardboard from the side, separating it from the forward flow of qualified cardboard, thus achieving fully automatic sorting; the design of multiple conveyor belts 2 avoids deformation of cardboard during conveying, ensuring morphological stability during inspection.

[0117] Please refer to Figure 2 In some examples, along the direction perpendicular to the conveying plane of conveyor belt 2, the distance between the diffuse light source 4 and the upper surface of the cardboard to be inspected is greater than 10 mm, the distance between the first image capturing unit 51 and the diffuse light source 4 is greater than 10 mm, and the distance between the second image capturing unit 61 and the diffuse light source 4 is greater than 10 mm. Figure 2 The positions of the first image capture unit 51 and the second image capture unit 61 are shown in the diagram. The actual height needs to be adjusted during installation to obtain the required image and light information.

[0118] The distance between the diffuse light source 4 and the cardboard, and the distance between the image capturing element and the light source, are both greater than 10mm. This ensures that the light from the diffuse light source 4 evenly covers the surface of the cardboard, avoiding local overexposure or underexposure. At the same time, it ensures that the field of view of the image capturing element (camera) is complete, the focus is clear, and edge distortion is reduced. It also provides clear parameters for device installation, ensures the consistency of different batches of devices, and reduces the difficulty of debugging.

[0119] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0120] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any change or replacement within the technical scope disclosed by the present application should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for detecting the warpage of cardboard, characterized in that, include: Image acquisition involves capturing image information of the cardboard using a light source and camera positioned above it. Image preprocessing converts the acquired image information into a grayscale image, removes noise, and then performs histogram equalization to enhance contrast. Reflection feature extraction: Calculate the brightness gradient in the horizontal and vertical directions to obtain the reflection intensity map, and then perform binarization to obtain the reflection feature map; Edge detection uses the Canny algorithm to extract basic edges, performs morphological closure operation on the reflection feature map, and then performs a bitwise OR operation with the basic edges to obtain enhanced edges; Morphological operations are performed on the enhanced edges by erosion operations with different structural elements to obtain the optimized edge image; Contour analysis extracts the outer contour from the optimized edge image, selects the contour with the largest area as the cardboard contour, and then detects the main straight lines in the contour through Hough transform. For warpage calculation, the two longest detection lines are selected as reference lines. The distance from all contour points to these two reference lines is calculated. The weighted distance is calculated based on the reflection intensity at the corresponding position in the reflection feature map. The cardboard length is taken as the maximum side length of the contour bounding box. The warpage output is the ratio of the maximum weighted distance to the cardboard length.

2. The method for detecting cardboard warpage according to claim 1, characterized in that, It also includes classifying the warp results, classifying the test results into qualified, slight warp, and severe warp based on the warp value, and outputting the test results.

3. The method for detecting cardboard warpage according to claim 1 or 2, characterized in that, The morphological operations include erosion of the enhanced edges using 3×3 structuring elements and dilation of the edges using 5×5 structuring elements.

4. The method for detecting cardboard warpage according to claim 1 or 2, characterized in that, The reflection feature extraction includes: Calculate the brightness gradient in the horizontal and vertical directions; Generate a reflection intensity map based on gradient amplitude; Thresholding is applied to the reflection intensity map to highlight areas with significant reflection changes.

5. The method for detecting cardboard warpage according to claim 1 or 2, characterized in that, The weighted distance in the warpage calculation is: Weighted distance = Base distance + Base distance × Reflection intensity + Base distance × Gradient sensitivity coefficient; The reflection intensity is a reflection characteristic value normalized to the range of 0-1; The gradient sensitivity coefficient ranges from 1.0 to 1.

5.

6. A cardboard warpage detection device, characterized in that, The paperboard warpage detection method according to any one of claims 1 to 5 is used to detect the paperboard warpage, wherein the paperboard warpage detection device comprises: The supporting body has a placement area for holding the cardboard to be tested. A conveyor belt is disposed on the support body and located on one side of the placement area. The conveyor belt is used to move the cardboard to be tested along the transport direction of the conveyor belt. A baffle is disposed on the supporting body and located at the transport end of the conveyor belt. The baffle is used to cover the cardboard to be inspected. A diffuse light source is disposed on the baffle, and the light from the diffuse light source is directed toward the cardboard to be tested; A first support extends along a direction perpendicular to the conveyor belt, and is disposed on one side of the placement area and above the conveyor belt; The first image capture unit is mounted on the first bracket; The second support extends along the conveying direction of the conveyor belt, and is disposed on one side of the conveyor belt and above the conveyor belt; The second image capture unit is disposed on the second bracket and is distributed at intervals along the conveying direction perpendicular to the conveyor belt, projecting onto the surface of the cardboard to be inspected.

7. The cardboard warpage detection device according to claim 6, characterized in that, The cardboard warpage detection device further includes a baffle plate, which extends along a plane perpendicular to the conveyor belt, is located on one side of the placement area, is located above the conveyor belt, and is located on the side of the first support away from the baffle plate; A gap is provided between the baffle plate and the conveyor belt. The gap between the baffle plate and the conveyor belt is greater than the thickness of one piece of paperboard to be tested and less than the thickness of two pieces of paperboard to be tested.

8. The cardboard warpage detection device according to claim 5, characterized in that, The cardboard warpage detection device also includes two vertical slide rails, which are mounted on the support body and located on both sides of the conveyor belt. The two ends of the baffle are slidably mounted on the two vertical slide rails. When the paperboard to be tested passes the inspection, the vertical slide rail drives the baffle to slide upward, and the paperboard to be tested slides out from the conveyor belt.

9. The cardboard warpage detection device according to claim 8, characterized in that, The second support is provided with a space for the paperboard to be inspected to pass through, and the multiple conveyor belts are arranged in parallel and spaced apart; The cardboard warpage detection device also includes multiple rollers, which are disposed on one side of the baffle and are respectively disposed between multiple conveyor belts. The rotation direction of the rollers is perpendicular to the conveying direction of the conveyor belts. The rollers can send out the unqualified cardboard to be tested from the space inside the second support when it is blocked in front of the baffle.

10. The cardboard warpage detection device according to claim 6, characterized in that, Along the direction perpendicular to the conveying plane of the conveyor belt, the distance between the diffuse light source and the upper surface of the cardboard to be inspected is greater than 10 mm, the distance between the first image capturing unit and the diffuse light source is greater than 10 mm, and the distance between the second image capturing unit and the diffuse light source is greater than 10 mm.

Images