A method for detecting surface defects of laser-etched label paper based on machine vision

Abstract

The present application relates to the field of image processing technology, and in particular to a method for detecting surface defects of laser-etched label paper based on machine vision. The method comprises: obtaining a set A of laser-etched label paper surface images collected within a historical etching time period; obtaining defect recognition results for each laser-etched label paper surface image in A; for the defect recognition results of any image in A, if the defect recognition results include defects, comparing the identified defects with the corresponding real defects, and obtaining defects that differ from the corresponding real defects among the identified defects; if the ratio of the number of defects with differences corresponding to all images in A to the number of real defects is greater than or equal to a preset ratio threshold, obtaining a motion speed adjustment amount of the laser etching device based on the defects with differences, and obtaining the target motion speed of the laser etching device in the current time period in combination with V0. The present invention can determine a laser etching speed that matches the parameters of the image acquisition device.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method for detecting surface defects of laser-etched label paper based on machine vision. Background Art

[0002] Laser etching technology, due to its high precision and non-contact processing characteristics, is widely used in the mass production of label surface graphics and text. Surface defects on label paper (such as incomplete etching, rough edges, and smearing) directly affect the readability and reliability of product labels, placing high demands on real-time defect detection and equipment parameter adjustment during the etching process. Traditional manual visual inspection is inefficient and highly subjective, making it difficult to adapt to high-speed automated production lines. While machine vision-based inspection technology has automated defect recognition, it faces the following issues: it focuses solely on defect recognition but fails to consider image distortion caused by the mismatch between the image acquisition device and the laser etching speed. This, in turn, results in unreliable defect detection results based on distorted images. Determining a laser etching speed that matches the parameters of the image acquisition device to avoid distortion in the images captured by the image acquisition device is an urgent issue that needs to be addressed. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for detecting surface defects of laser-etched label paper based on machine vision, so as to determine the laser etching speed that matches the parameters of the image acquisition device and avoid distortion of the image captured by the image acquisition device.

[0004] According to the present invention, a method for detecting surface defects of laser-etched label paper based on machine vision is provided, comprising the following steps:

[0005] S100, obtaining a set A of laser-etched label paper surface images collected within a historical etching time period; the historical etching time period includes several image collection moments, A includes the laser-etched label paper surface image collected at each image collection moment, and the movement speed of the laser etching device within the historical etching time period is V0.

[0006] S200 , inputting the laser-etched label paper surface images in A into the trained target defect recognition model respectively, and obtaining the defect recognition result of each laser-etched label paper surface image in A.

[0007] S300, for the defect recognition result of any laser-etched label paper surface image in A, if the defect recognition result includes defects, the defects recognized by the target defect recognition model are compared with the corresponding real defects to obtain defects that are different from the corresponding real defects among the defects recognized by the target defect recognition model; the real defects are defects existing in the corresponding laser-etched label paper surface image obtained by static shooting.

[0008] S400, if the ratio of the number of defects with differences corresponding to all laser-etched label paper surface images in A to the number of actual defects is greater than or equal to a preset ratio threshold, then the motion speed adjustment amount of the laser etching device is obtained according to the defects with differences, and the target motion speed of the laser etching device in the current time period is obtained according to the motion speed adjustment amount and V0.

[0009] Compared with the prior art, the present invention has at least the following beneficial effects:

[0010] The present invention obtains a set of images of the surface of laser-etched label paper captured during a historical etching period. These images are captured at a laser etching device speed of V0. By comparing defects in these images with defects in corresponding images of the laser-etched label paper surface captured using a static shooting method, it is possible to determine whether the images captured at the laser etching device speed of V0 are distorted. For example, if the ratio of the number of defects corresponding to all laser-etched label paper surface images in A to the number of actual defects is greater than or equal to a preset ratio threshold, then it is determined that the images captured at the laser etching device speed of V0 are distorted. In this case, a speed adjustment amount for the laser etching device is obtained based on the defects, so as to adjust V0 and obtain a speed of the laser etching device that matches the parameters of the image acquisition device. Thus, the present invention improves the compatibility between the laser etching speed and the parameters of the image acquisition device, can avoid distortion in the images captured by the image acquisition device due to mismatch between the laser etching speed and the parameters of the image acquisition device, and facilitates subsequent accurate defect detection based on the images captured by the image acquisition device. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A flow chart of a method for detecting surface defects of laser-etched label paper based on machine vision provided by an embodiment of the present invention;

[0013] Figure 2 A flowchart of the V0 process provided by an embodiment of the present invention;

[0014] Figure 3 A flowchart of a process for obtaining a motion speed adjustment value of a laser etching device according to defects with differences provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0016] According to this embodiment, Figure 1 As shown, a method for detecting surface defects of laser-etched label paper based on machine vision is provided, comprising the following steps:

[0017] S100, obtaining a set A of laser-etched label paper surface images collected within a historical etching time period; the historical etching time period includes several image collection moments, A includes the laser-etched label paper surface image collected at each image collection moment, and the movement speed of the laser etching device within the historical etching time period is V0.

[0018] As an optional specific embodiment, the laser etching equipment is a roll-to-roll system, the material is PET label paper with a thickness of 0.1 mm and a width of 100 mm; the image collector is a linear array camera with a line frequency of 200 kHz, a pixel resolution of 2048, a pixel size of 5 μm, and an optical magnification of 1:1.

[0019] As an optional specific implementation, during a historical period, images are collected at fixed intervals and saved as an image set A={A1, A2, ..., A j ,...,A m}; Among them, except for the different acquisition time, other variables are the same, for example, the label paper material is the same, the relevant parameters of the image acquisition device (such as exposure time, etc.) are the same, and the relevant parameters of the laser etching device (such as laser frequency and vibration conditions, etc.) are also the same.

[0020] As an optional embodiment, V0 is an empirical value; as a preferred embodiment, Figure 2 As shown, the process of obtaining V0 includes:

[0021] S110, obtaining a first condition satisfied by the movement speed of the laser etching device according to the line frequency, pixel resolution and single-line field of view width of the image collector; the first condition includes: V≤f×W / p, wherein V is the movement speed of the laser etching device, f is the line frequency of the image collector, W is the single-line field of view width, and p is the pixel resolution.

[0022] In this embodiment, the line frequency, pixel resolution and single-line field of view width of the image collector are known and do not change. Based on the above first condition, a first value range of the movement speed of the laser etching device can be obtained.

[0023] S120, obtaining a second condition satisfied by the movement speed of the laser etching device based on the exposure time, pixel size and optical magnification of the image collector; the second condition includes: V≤s / (T×M), where s is the pixel size, T is the exposure time of the image collector, and M is the optical magnification.

[0024] In this embodiment, the exposure time, pixel size and optical magnification of the image collector are known and do not change. Based on the above second condition, a second value range of the movement speed of the laser etching device can be obtained.

[0025] S130, obtaining a third condition satisfied by the movement speed of the laser etching equipment based on the minimum detectable defect size, the exposure time of the image collector, and the minimum number of pixels spanned by the defect; the third condition includes: V≤d / (T×q), where d is the minimum detectable defect size, and q is the minimum number of pixels spanned by the defect.

[0026] In this embodiment, the minimum detectable defect size, the exposure time of the image collector, and the minimum number of pixels spanned by the defect are known and do not change. Based on the above third condition, the third value range of the movement speed of the laser etching equipment can be obtained.

[0027] S140: Determine a movement speed that meets the first condition, the second condition, and the third condition as V0.

[0028] In this embodiment, V0>0, a movement speed that satisfies the first condition, the second condition, and the third condition is a value in the intersection of the first value range, the second value range, and the third value range. Optionally, V0 is the maximum movement speed that satisfies the first condition, the second condition, and the third condition.

[0029] Based on S110-S140, the movement speed of the laser etching device that is more closely matched with the parameters of the image collector can be obtained, which is beneficial to reducing the workload of the subsequent adjustment process of the movement speed of the laser etching device.

[0030] S200 , inputting the laser-etched label paper surface images in A into the trained target defect recognition model respectively, and obtaining the defect recognition result of each laser-etched label paper surface image in A.

[0031] As an optional specific implementation, the target defect recognition model adopts an existing target detection model to detect the following defects in the image: excessive ablation (area abnormality), tail marks (length abnormality) and edge deformation (shape abnormality). The output of the target defect recognition model (i.e., the recognition result) includes the category of each defect, the location coordinates of each defect, the area of each defect (number of pixels), the tail length of each defect and the contour circumference of each defect.

[0032] Those skilled in the art know that the process of training the target detection model is an existing technology and will not be described in detail here.

[0033] S300, for the defect recognition result of any laser-etched label paper surface image in A, if the defect recognition result includes defects, the defects recognized by the target defect recognition model are compared with the corresponding real defects to obtain defects that are different from the corresponding real defects among the defects recognized by the target defect recognition model; the real defects are defects existing in the corresponding laser-etched label paper surface image obtained by static shooting.

[0034] As an optional specific implementation method, for any laser-etched label paper surface image in A (i.e., a dynamically captured image), its corresponding static captured image is used as the real defect reference (the label paper is stationary during shooting, and the movement speed of the laser etching equipment is 0, which is the same as the object captured during dynamic shooting). There is no distortion problem in the static captured image. The dynamically captured image is aligned with the corresponding static reference through the image registration algorithm, and the geometric difference between the identified defect and the corresponding real defect is calculated.

[0035] As an optional specific implementation, the geometric difference includes area difference, tail length difference and image deformation, wherein the area difference is |identified area-real area| / real area×100%. If the area difference is greater than or equal to the preset area ratio threshold (for example, 5%), it is determined that there is an area difference; the tail length difference is the difference between the extension length of the defect identified in the dynamic image along the etching direction and the extension length of the real defect along the etching direction. If the difference is greater than or equal to the preset tail length difference threshold (for example, 2 pixels), it is determined that there is a tail length difference; the image deformation amount is |identified contour circumference and real contour circumference| / real contour circumference×100%. If the graphic deformation amount is greater than or equal to the preset image deformation threshold (for example, 2%), it is determined that there is image deformation.

[0036] As an optional specific implementation, if there is no area difference, tail length difference and image deformation between the detected defect and the corresponding real defect, it is determined that there is no difference between the identified defect and the corresponding real defect; otherwise, it is determined that there is a difference between the identified defect and the corresponding real defect.

[0037] S400, if the ratio of the number of defects with differences corresponding to all laser-etched label paper surface images in A to the number of actual defects is greater than or equal to a preset ratio threshold, then the motion speed adjustment amount of the laser etching device is obtained according to the defects with differences, and the target motion speed of the laser etching device in the current time period is obtained according to the motion speed adjustment amount and V0.

[0038] Optionally, the preset ratio threshold is an empirical value, for example, the preset ratio threshold is 10%; if the ratio of the number of defects with differences corresponding to all laser-etched label paper surface images in A to the number of actual defects is less than the preset ratio threshold, the movement speed of the laser etching device will no longer be adjusted, and it is determined that the laser etching speed matches the parameters of the image acquisition device.

[0039] As a preferred embodiment, Figure 3 As shown, obtaining the motion speed adjustment amount of the laser etching equipment according to the defects with differences includes:

[0040] S410, for the i-th defect D with difference i , get D i Compared with the area difference of the corresponding real defect, the area difference is substituted into the first preset model to obtain D i The corresponding first motion speed adjustment amount ΔV i,1 ; The first preset model includes the relationship between the area difference and the movement speed adjustment amount; the value range of i is 1 to n, where n is the number of defects with differences.

[0041] In this embodiment, n is the number of different defects corresponding to all laser-etched label paper surface images in A.

[0042] In this embodiment, if the movement speed of the laser etching device does not match the parameters of the image collector, then D i There is an area difference between the area of the corresponding real defect. As an optional specific implementation, the first preset model is obtained based on experience; as an optional specific implementation, the acquisition process of the first preset model includes:

[0043] S411, obtaining the areas of defects in the laser-etched label paper surface images corresponding to different sample speeds.

[0044] Optionally, the number of different sample speeds is large enough to cover the full speed range of the laser etching equipment, so as to improve the generalization and accuracy of the first preset model obtained subsequently; and except for the different speeds, other variables are the same as the variables corresponding to the above-mentioned historical etching time period, for example, the label paper material is the same, the relevant parameters of the image acquisition device (such as exposure time, etc.) are the same, and the relevant parameters of the laser etching equipment (such as laser frequency and vibration conditions, etc.) are also the same.

[0045] S412, obtaining the areas of defects in the surface images of the laser-etched label paper obtained by static shooting corresponding to different sample speeds.

[0046] S413, obtaining the area difference of the defects in the laser etched label paper surface image and the corresponding speed adjustment amount corresponding to different sample speeds; the area difference of the defects in the laser etched label paper surface image corresponding to the adjusted speed corresponding to any sample speed is less than or equal to a preset area difference threshold, and the adjusted speed corresponding to any sample speed is the sum of the sample speed and the corresponding speed adjustment amount.

[0047] In this embodiment, the area difference of any defect in the laser-etched label paper surface image corresponding to any sample speed is |area of the defect - actual area| / actual area × 100%, where the actual area is the area of the corresponding defect in the laser-etched label paper surface image obtained by static shooting corresponding to the sample speed.

[0048] Optionally, the speed adjustment amounts corresponding to different sample speeds are obtained through experiments.

[0049] S414 , obtaining a relationship between the area difference of the defect and the speed adjustment amount by data fitting, and obtaining a first preset model.

[0050] Those skilled in the art know that the process of data fitting is an existing technology and will not be described in detail here.

[0051] Based on S411 - S414 , a relatively accurate second preset model can be obtained.

[0052] S420, obtain D i Compared with the tail length difference of the corresponding real defect, the tail length difference is substituted into the second preset model to obtain D i The corresponding second motion speed adjustment amount ΔV i,2 ; The second preset model includes the relationship between the tail length difference and the motion speed adjustment amount.

[0053] In this embodiment, if the movement speed of the laser etching device does not match the parameters of the image collector, then D iThere is a difference in tail length between the corresponding real defect and the real defect. As an optional specific implementation, the second preset model is obtained based on experience; as an optional specific implementation, the acquisition process of the second preset model includes:

[0054] S421, obtaining the tail length of the defect in the laser-etched label paper surface image corresponding to different sample speeds.

[0055] Optionally, the number of different sample speeds is large enough to cover the full speed range of the laser etching equipment, so as to improve the generalization and accuracy of the first preset model obtained subsequently; and except for the different speeds, other variables are the same as the variables corresponding to the above-mentioned historical etching time period, for example, the label paper material is the same, the relevant parameters of the image acquisition device (such as exposure time, etc.) are the same, and the relevant parameters of the laser etching equipment (such as laser frequency and vibration conditions, etc.) are also the same.

[0056] S422, obtaining tail lengths of defects in the image of the surface of the laser-etched label paper obtained by static shooting corresponding to different sample speeds.

[0057] S423, obtaining the difference in the tail length of the defect in the laser-etched label paper surface image corresponding to different sample speeds and the corresponding speed adjustment amount; the difference in the tail length of the defect in the laser-etched label paper surface image corresponding to any sample speed is the difference between the tail length of the defect in the laser-etched label paper surface image corresponding to the sample speed and the tail length of the defect in the corresponding laser-etched label paper surface image obtained by static shooting; the difference in the tail length of the defect in the laser-etched label paper surface image corresponding to the adjusted speed corresponding to any sample speed is less than or equal to a preset tail length difference threshold, and the adjusted speed corresponding to any sample speed is the sum of the sample speed and the corresponding speed adjustment amount.

[0058] Optionally, the speed adjustment amounts corresponding to different sample speeds are obtained through experiments.

[0059] S424 , obtaining the relationship between the defect tail length difference and the speed adjustment amount by data fitting, and obtaining a second preset model.

[0060] Those skilled in the art know that the process of data fitting is an existing technology and will not be described in detail here.

[0061] Based on S421-S424, a relatively accurate second preset model can be obtained.

[0062] S430, obtain D i Compared with the image deformation of the corresponding real defect, the image deformation is substituted into the third preset model to obtain D i The corresponding third motion speed adjustment amount ΔV i,3; The third preset model includes the relationship between the image deformation amount and the motion speed adjustment amount.

[0063] In this embodiment, if the movement speed of the laser etching device does not match the parameters of the image collector, then D i There is an image deformation amount corresponding to the real defect. As an optional specific implementation, the third preset model is obtained based on experience; as an optional specific implementation, the acquisition process of the third preset model includes:

[0064] S431, obtaining image deformation amounts of the laser-etched label paper surface images corresponding to different sample speeds compared with the corresponding laser-etched label paper surface images obtained by adopting a static shooting method.

[0065] Optionally, the number of different sample speeds is large enough to cover the full speed range of the laser etching equipment, so as to improve the generalization and accuracy of the first preset model obtained subsequently; and except for the different speeds, other variables are the same as the variables corresponding to the above-mentioned historical etching time period, for example, the label paper material is the same, the relevant parameters of the image acquisition device (such as exposure time, etc.) are the same, and the relevant parameters of the laser etching equipment (such as laser frequency and vibration conditions, etc.) are also the same.

[0066] S432, obtaining speed adjustment amounts corresponding to different sample speeds; the image deformation amount of the laser-etched label paper surface image corresponding to the adjusted speed corresponding to any sample speed compared to the corresponding laser-etched label paper surface image obtained by static shooting is less than or equal to a preset image deformation amount threshold, and the adjusted speed corresponding to any sample speed is the sum of the sample speed and the corresponding speed adjustment amount.

[0067] Optionally, the speed adjustment amounts corresponding to different sample speeds are obtained through experiments.

[0068] S433: Obtain the relationship between the image deformation difference and the speed adjustment amount by data fitting to obtain a third preset model.

[0069] Those skilled in the art know that the process of data fitting is an existing technology and will not be described in detail here.

[0070] Based on S431 - S433 , a relatively accurate third preset model can be obtained.

[0071] S440, according to ΔV i,1 , ΔV i,2 and ΔV i,3 Get D i The corresponding movement speed adjustment ΔV i .

[0072] Optionally, ΔV is obtained by weighted summation formulai , where ΔV i,1 , ΔV i,2 and ΔV i,3 The corresponding weight is the experience value, optional, ΔV i,1 , ΔV i,2 and ΔV i,3 The corresponding weights are all 1 / 3, or ΔV i,2 The corresponding weight is 1 / 2, ΔV i,1 and ΔV i,3 The corresponding weights are all 1 / 4.

[0073] S450, according to {ΔV1, ΔV2,…, ΔV i ,…,ΔV n}Obtain the motion speed adjustment value ΔV of the laser etching equipment.

[0074] Optionally, ΔV=∑ n i=1 ΔV i / n.

[0075] Based on S410 - S450 , a relatively accurate motion speed adjustment value ΔV of the laser etching device can be obtained.

[0076] In this embodiment, after obtaining the target movement speed of the laser etching device in the current time period, the movement speed of the laser etching device is updated, and a method similar to S100-S200 is used to determine again whether there is a difference between the defects identified by the updated target defect recognition model and the corresponding real defects. If the ratio of the number of defects with differences to the number of corresponding real defects is greater than or equal to a preset ratio threshold, the movement speed of the laser etching device is continued to be updated until the ratio of the number of defects with differences to the number of corresponding real defects is less than the preset ratio threshold, thereby achieving matching of the laser etching speed with the parameters of the image acquisition device.

[0077] This embodiment acquires a set of laser-etched label paper surface images captured during a historical etching period. These images were captured at a laser etching device speed of V0. By comparing defects in these images with defects in corresponding images of the laser-etched label paper surface captured using a static shooting method, it is possible to determine whether the images captured at the laser etching device speed of V0 are distorted. For example, if the ratio of the number of discrepant defects corresponding to all laser-etched label paper surface images in A to the number of actual defects is greater than or equal to a preset ratio threshold, then it is determined that the images captured at the laser etching device speed of V0 are distorted. In this case, a laser etching device speed adjustment amount is obtained based on the discrepant defects to adjust V0 to obtain a laser etching device speed that matches the parameters of the image acquisition device. Thus, this embodiment improves the compatibility between the laser etching speed and the parameters of the image acquisition device, avoiding distortion in images captured by the image acquisition device due to a mismatch between the laser etching speed and the parameters of the image acquisition device, and facilitating subsequent accurate defect detection based on the images captured by the image acquisition device.

[0078] Although some specific embodiments of the present invention have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A method for detecting surface defects of laser-etched label paper based on machine vision, characterized in that: The following steps are involved: S100, obtaining a set A of laser-etched label paper surface images collected during a historical etching period; the historical etching period includes a number of image collection moments, A includes a laser-etched label paper surface image collected at each image collection moment, and a movement speed of the laser etching device during the historical etching period is V0; S200, inputting the laser-etched label paper surface images in A into the trained target defect recognition model to obtain a defect recognition result for each laser-etched label paper surface image in A; S300, for any defect recognition result of the laser-etched label paper surface image in A, if the defect recognition result includes a defect, comparing the defect recognized by the target defect recognition model with the corresponding real defect to obtain defects that are different from the corresponding real defect; the real defect is a defect existing in the corresponding laser-etched label paper surface image obtained using a static shooting method; S400: If the ratio of the number of different defects corresponding to all laser-etched label paper surface images in A to the number of actual defects is greater than or equal to a preset ratio threshold, a motion speed adjustment amount for the laser etching device is obtained based on the different defects, and a target motion speed of the laser etching device in the current time period is obtained based on the motion speed adjustment amount and V0; The speed adjustment of the laser etching equipment is obtained based on the defects with different characteristics, including: S410, for the i-th defect D with difference i , get D i Compared with the area difference of the corresponding real defect, the area difference is substituted into the first preset model to obtain D i The corresponding first motion speed adjustment amount ΔV i,1 The first preset model includes the relationship between the area difference and the movement speed adjustment amount; the value range of i is 1 to n, where n is the number of defects with differences; S420, obtain D i Compared with the tail length difference of the corresponding real defect, the tail length difference is substituted into the second preset model to obtain D i The corresponding second motion speed adjustment amount ΔV i,2 ; The second preset model includes the relationship between the tail length difference and the motion speed adjustment amount; S430, obtain D i Compared with the image deformation of the corresponding real defect, the image deformation is substituted into the third preset model to obtain D i The corresponding third motion speed adjustment amount ΔV i,3 The third preset model includes the relationship between the image deformation amount and the motion speed adjustment amount; S440, according to ΔV i,1 , ΔV i,2 and ΔV i,3 Get D i The corresponding movement speed adjustment ΔV i ; S450, according to {ΔV1, ΔV2,…, ΔV i ,…,ΔV n Obtaining the motion speed adjustment value ΔV of the laser etching device; The process of obtaining V0 includes: S110, obtaining a first condition satisfied by the movement speed of the laser etching device based on the line frequency, pixel resolution, and single-line field of view width of the image collector; the first condition comprising: V ≤ f × W / p, where V is the movement speed of the laser etching device, f is the line frequency of the image collector, W is the single-line field of view width, and p is the pixel resolution; S120, obtaining a second condition satisfied by the movement speed of the laser etching device based on the exposure time, pixel size, and optical magnification of the image collector; the second condition includes: V ≤ s / (T × M), where s is the pixel size, T is the exposure time of the image collector, and M is the optical magnification; S130, obtaining a third condition for the movement speed of the laser etching device based on the minimum detectable defect size, the exposure time of the image collector, and the minimum number of pixels spanned by the defect; the third condition includes: V ≤ d / (T × q), where d is the minimum detectable defect size and q is the minimum number of pixels spanned by the defect; S140: Determine a movement speed that meets the first condition, the second condition, and the third condition as V0.

2. The method for detecting surface defects of laser-etched label paper based on machine vision according to claim 1, characterized in that: The process of obtaining the first preset model includes: S411, obtaining the area of defects in the laser-etched label paper surface image corresponding to different sample speeds; S412, obtaining the area of defects in the image of the surface of the laser-etched label paper obtained by static shooting corresponding to different sample speeds; S413, obtaining the area difference of the defect in the laser-etched label paper surface image corresponding to different sample speeds and the corresponding speed adjustment amount; if the area difference of the defect in the laser-etched label paper surface image corresponding to the adjusted speed corresponding to any sample speed is less than or equal to a preset area difference threshold, the adjusted speed corresponding to any sample speed is the sum of the sample speed and the corresponding speed adjustment amount; S414 , obtaining a relationship between the area difference of the defect and the speed adjustment amount by data fitting, and obtaining a first preset model.

3. The method for detecting surface defects of laser-etched label paper based on machine vision according to claim 1, characterized in that: The process of obtaining the second preset model includes: S421, obtaining the tail length of the defect in the laser-etched label paper surface image corresponding to different sample speeds; S422, obtaining tail lengths of defects in the image of the laser-etched label paper surface obtained by static shooting corresponding to different sample speeds; S423, obtaining the difference in the tail length of the defect in the laser-etched label paper surface image corresponding to different sample speeds and the corresponding speed adjustment amount; the difference in the tail length of the defect in the laser-etched label paper surface image corresponding to any sample speed is the difference between the tail length of the defect in the laser-etched label paper surface image corresponding to the sample speed and the tail length of the defect in the corresponding laser-etched label paper surface image obtained using a static shooting method; the difference in the tail length of the defect in the laser-etched label paper surface image corresponding to the adjusted speed corresponding to any sample speed is less than or equal to a preset tail length difference threshold, and the adjusted speed corresponding to any sample speed is the sum of the sample speed and the corresponding speed adjustment amount; S424 , obtaining the relationship between the defect tail length difference and the speed adjustment amount by data fitting, and obtaining a second preset model.

4. The method for detecting surface defects of laser-etched label paper based on machine vision according to claim 1, characterized in that: The process of obtaining the third preset model includes: S431, obtaining image deformation amounts of the laser-etched label paper surface images corresponding to different sample speeds compared with corresponding laser-etched label paper surface images obtained by static shooting; S432, obtaining speed adjustment amounts corresponding to different sample speeds; the image deformation amount of the laser-etched label paper surface image corresponding to the adjusted speed for any sample speed compared to the corresponding image of the laser-etched label paper surface obtained using a static shooting method is less than or equal to a preset image deformation amount threshold, and the adjusted speed corresponding to any sample speed is the sum of the sample speed and the corresponding speed adjustment amount; S433: Obtain the relationship between the image deformation difference and the speed adjustment amount by data fitting to obtain a third preset model.

5. The method for detecting surface defects of laser-etched label paper based on machine vision according to claim 1, characterized in that: Obtain ΔV using the weighted summation formula i .

6. The method for detecting surface defects of laser-etched label paper based on machine vision according to claim 1, characterized in that: ΔV=∑ n i=1 ΔV i / n.

Images