An operating panel sheet printing positioning method based on machine vision
By using machine vision-based scarcity calculation and template matching methods, the problem of inaccurate positioning in the printing of control panel sheets was solved, achieving high-precision and stable positioning results, and improving production efficiency and printing quality.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- JIANGSU KUNDA ELECTRICAL DECORATION CO LTD
- Filing Date
- 2025-07-08
- Publication Date
- 2026-06-09
AI Technical Summary
In the process of printing control panel sheets, existing technologies are prone to misidentification or positioning deviation when patterns are repetitive, symmetrical, or subject to changes in lighting. This makes it difficult to accurately and stably select positioning points, affecting printing quality and production efficiency.
A machine vision-based approach is adopted. By calculating the scarcity of any pixel in the image, the target point corresponding to the maximum scarcity is selected as the localization point. By combining corner detection and feature distribution within the window, scarce feature values are obtained, and template matching is used to achieve high-precision localization.
It improves the positioning accuracy of the control panel sheet printing, reduces human error, increases production efficiency, avoids interference from repeated patterns, and enhances recognition stability.
Smart Images
Figure CN120726128B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image processing technology. More specifically, this invention relates to a machine vision-based method for positioning and printing control panel sheets. Background Technology
[0002] With the development of modern manufacturing, especially in industries such as electronics, automobiles, and home appliances, the demand for control panels is constantly increasing. Control panels are a key component of these products, requiring not only ease of operation but also an attractive appearance that meets consumer aesthetic requirements. To ensure the quality and performance of control panels, precise printing positioning has become a crucial step in the production process. In traditional control panel sheet printing, the level of positioning accuracy directly affects the printing quality and effect. Traditional positioning methods often rely on manual operation or mechanical positioning, which are easily affected by human factors, mechanical precision, and environmental changes, causing printing deviations or errors, affecting product quality, and even reducing production efficiency. To overcome these problems, machine vision technology has emerged and shown great application potential in control panel sheet printing positioning. Machine vision systems acquire images of the control panel through cameras and analyze positioning information in real time using image processing algorithms. This allows for precise identification and correction of the control panel sheet position, greatly improving positioning accuracy, reducing human error, and increasing production efficiency.
[0003] In existing technologies, when positioning printed patterns on control panel sheets, template matching, feature point matching, or preset marker points are commonly used. These methods are prone to misidentification or positioning deviation when patterns are repetitive, symmetrical, or subject to changes in lighting. It is difficult to accurately and stably select positioning points, thereby affecting the accuracy and stability of subsequent automated processes such as printing, lamination, or inspection. Summary of the Invention
[0004] This invention provides a machine vision-based method for printing and positioning operation panel sheets, aiming to solve the problem in related technologies that misidentification or positioning deviations easily occur when patterns are repetitive, symmetrical, or subject to changes in lighting, making it difficult to accurately and stably select positioning points.
[0005] This invention provides a machine vision-based method for positioning printed operation panel sheets, comprising: acquiring an image of the operation panel sheet; taking any pixel in the image as a target point, calculating the scarcity of the target point in the image, and selecting the target point corresponding to the maximum scarcity as a positioning point, and positioning the real-time image, wherein the scarcity is positively correlated with both the first and second scarcity feature values of the target point; obtaining the first scarcity feature value, comprising: extracting connected components from the image to obtain multiple connected components, obtaining corner points on the connected components based on corner point detection, and calculating the difference in distance values from each corner point on the connected components to the target point, and taking the reciprocal of the frequency of the difference in the distance value of the target point among the differences in the distance values of all target points as the first scarcity feature value of the target point; obtaining the second scarcity feature value of the target point, comprising: establishing a target window centered on the target point, and taking the reciprocal of the frequency of the feature distribution within the target window in the image as the second scarcity feature value of the target point. The first scarcity is calculated by the difference in distance between each point on the image and the corner points of each connected region. The target window of each point is obtained by dividing the image with each point as the center. The second scarcity is calculated based on the number of times the target window appears on the image. The scarcity of each point is obtained based on the first scarcity and the second scarcity. The point with the largest scarcity is selected as the positioning point, thereby improving the accuracy of image positioning.
[0006] Furthermore, the method for obtaining the target window includes: establishing windows of different sizes centered on the target point; determining the optimal window for the target point based on the contour coefficient method, and using it as the target window for the target point. By establishing windows of different sizes centered on the target point and adaptively selecting the optimal window size based on the contour coefficient method, the region that best expresses the local features of the target point can be dynamically captured, improving the discriminative power and stability of the second scarce feature value.
[0007] Furthermore, the method for obtaining the target window includes: setting the size of the target window to... , It is a constant.
[0008] Furthermore, real-time image localization includes: acquiring an image of the target window corresponding to the localization point as a template to be matched; and performing template matching on the real-time image according to the template to be matched to complete the localization.
[0009] Furthermore, the frequency of occurrence of the feature distribution within the target window in the image includes: calculating the number of matches of the target window in the image using a squared difference matching algorithm, and using the number of matches as the frequency of occurrence of the feature distribution within the target window in the image. Calculating the number of matches of the target window in the entire image using the squared difference matching algorithm can efficiently and stably evaluate the repetition frequency of features within the window in the image.
[0010] Furthermore, the difference in distance values from the corner points of each connected component to the target point is calculated, including: for any connected component, calculating the mean of the distance values from all corner points of that component to the target point, and calculating the mean of the distance values from all connected components to the target point. The variance of all means is used as the difference in distance values from the corner points of each connected component to the target point. By calculating the mean of the distance values from the corner points of each connected component to the target point in the image, and using the mean-variance as an index of the spatial variability of the target point, this method effectively quantifies the uniqueness of the target point relative to the overall structural distribution.
[0011] Furthermore, the differences in distances from corner points to the target point within each connected component are calculated. This includes: for any connected component, calculating the distances from each corner point within that component to the target point, and then using these distances as a set. The differences between these sets are then used as the overall difference in distances from corner points to the target point across all connected components. By using the distances from corner points to the target point within each connected component as a set and comparing the differences between these sets to measure the spatial structure distribution, this method more comprehensively preserves the geometric information of the structure in the image than the traditional mean-variance method, thus more accurately assessing the scarcity of target points.
[0012] Furthermore, the differences in distances from the upper corner points of each connected component to the target point are calculated using the following formula: In the formula, Indicates the distance from each corner point of a connected domain to the first... The difference in distance values between target points Represents the distance from each corner point on all connected domains to the first... The number of unique distance values in the union of all sets of a target point Represents the distance from each corner point on all connected domains to the first... The number of common distance values in the intersection of all sets of a target point.
[0013] Furthermore, obtaining corner points on the connected domain based on corner detection includes: obtaining corner points on the connected domain using Harris corner detection.
[0014] Furthermore, acquiring an image of the operation panel sheet also includes: performing noise reduction processing on the image using Gaussian filtering.
[0015] Beneficial effects: By introducing the concept of scarcity and combining corner distribution features and the frequency of image features within the window, the most unique positioning point in the image can be effectively selected to achieve high-precision positioning. This method can avoid interference from repeated patterns, improve recognition stability, and significantly improve positioning accuracy. Attached Figure Description
[0016] Figure 1This is a flowchart illustrating the positioning of an operation panel sheet according to an embodiment of the present invention.
[0017] Figure 2 This is a flowchart illustrating the calculation of a second scarce feature value according to an embodiment of the present invention. Detailed Implementation
[0018] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
[0019] like Figure 1 As shown, S101: Acquires images of the operation panel sheet.
[0020] In one embodiment, an industrial camera is used to capture images of the control panel sheet. During shooting, light sources are positioned on both sides of the control panel sheet and turned on to illuminate the sheet. To achieve the best illumination effect, a ring-shaped shadowless light source, a strip light source, or a backlight can be used to obtain a clear, high-contrast, and detail-rich image, improving the accuracy of subsequent image processing. Then, Gaussian filtering is used to denoise the captured image, removing interference noise caused by tiny dust particles, surface scratches, static electricity particles, etc., further improving the accuracy of subsequent image processing.
[0021] S102: Calculate the scarcity of the target point.
[0022] In one embodiment, the control panel sheet typically features panels with identical shapes. Therefore, to avoid positioning errors caused by similar shapes, this embodiment determines the positioning location based on the scarcity of target points at different positions in the image. The point with the highest scarcity in the image is used as the positioning point, avoiding misjudgment or misalignment due to identical shapes on the panel. For example, if a control panel has multiple buttons, each with the same shape and size, but only a logo or numbered box in the upper left corner, the logo can be selected as the positioning reference point to improve positioning accuracy. The scarcity of the target point is positively correlated with both its first and second scarcity feature values.
[0023] In one embodiment, a formula for calculating the scarcity of a target point is provided, and the formula is as follows: , Indicates the first The scarcity of each target point Indicated by An exponential function with base 0. This represents the first scarce feature value of the target point. This represents the second scarce feature value of the target point.
[0024] In one embodiment, the first scarce feature value reflects the difference between the target point and the corner points of each connected domain. Since the shapes of each connected domain are the same, the difference in distance between the target point and the corner points of each connected domain is relatively large in most locations. However, the target point located at the center of the image has a smaller difference in distance between it and the corner points of each connected domain because the distance to each connected domain is the same. Therefore, the smaller the value of the first scarce feature value, the less scarce the target point is. The second scarce feature value reflects the number of times the features around the target point appear in the image. The fewer the number of times they appear, the greater the scarcity of the target point. Using the target point with the greatest scarcity value will result in more accurate image localization.
[0025] like Figure 2 As shown, S1021: Calculate the first scarce feature value of the target point.
[0026] In one embodiment, calculating the first scarce feature value includes: extracting connected components from the image to obtain multiple connected components; obtaining corner points on the connected components based on corner detection, wherein the corner detection is Harris corner detection; calculating the difference in distance values from each corner point on each connected component to the target point; and using the reciprocal of the frequency of the difference in the target point distance value among all the differences in the target point distance values as the first scarce feature value of the target point. The first scarce feature value of the target point reflects the scarcity of the relationship between the target point and the corner points of each connected component. Therefore, the larger the first scarce feature value of the target point, the greater its scarcity, and the more accurate the image localization using the target point.
[0027] In one embodiment, a method is provided for calculating the difference in distance values from the corner points of each connected domain to the target point, comprising: for any connected domain, calculating the mean of the distance values from all corner points of the connected domain to the target point, and calculating the mean of the distance values from all connected domains to the target point, and using the variance of all the means as the difference in distance values from the corner points of each connected domain to the target point.
[0028] Following the above, a formula is provided for calculating the difference in distances from the upper corner points of each connected component to the target point. The formula is as follows: ; Indicates the distance from each corner point of a connected domain to the first... The difference in distance values between target points Indicates the first All corners on the connected component to the first... The mean of the distances to each target point. The mean distance from each corner point in all connected domains to the target point. This indicates the number of connected components.
[0029] In another embodiment, a method for calculating the difference in distance values from the upper corner points of each connected domain to the target point is also provided, comprising: for any connected domain, calculating the distance values from each corner point of the connected domain to the target point respectively, and taking the distance values from each corner point of the connected domain to the target point as a set, calculating the difference between the sets as the difference in distance values from the upper corner points of each connected domain to the target point.
[0030] Following the above, a formula is also provided for calculating the difference in distances from the upper corner points of each connected component to the target point. The formula is as follows: In the formula, Indicates the distance from each corner point of a connected domain to the first... The difference in distance values between target points Represents the distance from each corner point on all connected domains to the first... The number of unique distance values in the union of all sets of a target point Represents the distance from each corner point on all connected domains to the first... The intersection of all sets of a target point yields the number of common distance values in the set. The intersection-union ratio reflects the consistency of the distance distribution between corner points and the target point in different connected components, and is used to measure the spatial structural differences between connected components.
[0031] In summary, the differences between all points in the image that are target points are calculated, the frequency of the difference of any target point among all differences is counted, and then the reciprocal of the frequency is used as the first scarce feature value of the target point.
[0032] S1022: Calculate the second scarce feature value of the target point.
[0033] In one embodiment, a target window is established centered on the target point, and the reciprocal of the frequency of occurrence of the feature distribution within the target window in the image is used as the second scarce feature value of the target point. Specifically, the number of matches of the target window in the image is calculated using a squared difference matching algorithm, and this number of matches is used as the frequency of occurrence of the feature distribution within the target window in the image. This method allows us to obtain the scarcity of the target point's features across the entire image; the larger the second scarce feature value of the target point, the scarcer the target point.
[0034] In one embodiment, the method for obtaining the target window includes: setting the size of the target window to... , It is a constant. In this embodiment, The value of is 3. In other embodiments, The value can also be 5, and can be adjusted according to the implementation situation.
[0035] S103: Positioning of the printed control panel sheet.
[0036] In one embodiment, the scarcity of all target points in the image is calculated, the target point with the highest scarcity is selected as the positioning point, the image of the target window corresponding to the positioning point is obtained as the template to be matched, and finally, the real-time image is matched according to the template to be matched to complete the positioning.
[0037] In another embodiment, a different method for obtaining the target window is provided, including: establishing windows of different sizes centered on the target point; and determining the optimal window for the target point based on the silhouette coefficient method, which is then used as the target window for the target point. Specifically, for the target point, the silhouette coefficient method is used to evaluate the clustering effect of candidate windows at different scales or locations, and the window with the largest silhouette coefficient is selected to determine its optimal target window. The silhouette coefficient method is prior art and will not be described in detail here.
[0038] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.
Claims
1. A machine vision-based method for printing and positioning operation panel sheets, characterized in that, The positioning method includes: acquiring images of the operation panel sheet; Taking any pixel in the image as the target point, the scarcity of the target point in the image is calculated, and the target point corresponding to the maximum scarcity is selected as the positioning point for real-time image positioning. The scarcity is positively correlated with the first and second scarcity feature values of the target point. Obtaining the first scarce feature value includes: extracting connected components from the image to obtain multiple connected components, obtaining corner points on the connected components based on corner point detection, calculating the difference in distance values from each corner point on the connected components to the target point, and taking the reciprocal of the frequency of the difference in the distance value of the target point among all the differences in the distance values of the target point as the first scarce feature value of the target point. Obtaining the second scarce feature value of the target point includes: establishing a target window centered on the target point, and using the reciprocal of the frequency of occurrence of the feature distribution within the target window in the image as the second scarce feature value of the target point.
2. The machine vision-based operation panel sheet printing and positioning method according to claim 1, characterized in that, Methods for obtaining the target window include: Establish windows of different sizes centered on the target point; The optimal window for the target point is determined based on the contour coefficient method, and is used as the target window for the target point.
3. The machine vision-based operation panel sheet printing and positioning method according to claim 1, characterized in that, Methods for obtaining the target window include: Set the size of the target window to , It is a constant.
4. The machine vision-based operation panel sheet printing positioning method according to claim 1, characterized in that, Real-time image localization includes: Obtain the image of the target window corresponding to the positioning point as the template to be matched; The real-time image is located by performing template matching based on the template to be matched.
5. The machine vision-based operation panel sheet printing and positioning method according to claim 1, characterized in that, The frequency of occurrence of the feature distribution within the target window in the image includes: The number of times the target window is matched in the image is calculated using the squared difference matching algorithm, and the number of matches is used as the frequency of the feature distribution within the target window in the image.
6. The machine vision-based operation panel sheet printing and positioning method according to claim 1, characterized in that, Calculate the differences in distances from the upper corners of each connected component to the target point, including: For any connected component, calculate the mean of the distances from all corner points in the connected component to the target point, and calculate the mean of the distances from all connected components to the target point. Use the variance of all the means as the difference in the distances from the corner points in each connected component to the target point.
7. The machine vision-based operation panel sheet printing positioning method according to claim 1, characterized in that, Calculate the differences in distances from the upper corners of each connected component to the target point, including: For any connected component, calculate the distance from each corner point in the connected component to the target point, and take the distances from each corner point in the connected component to the target point as a set. Calculate the difference between the sets and use it as the difference in the distances from each corner point in the connected component to the target point.
8. The machine vision-based operation panel sheet printing positioning method according to claim 7, characterized in that, The difference in distance from the upper corner of each connected component to the target point is calculated using the following formula: ； In the formula, Indicates the distance from each corner point of a connected domain to the first... The difference in distance values between target points Represents the distance from each corner point on all connected domains to the first... The number of unique distance values in the union of all sets of a target point Represents the distance from each corner point on all connected domains to the first... The number of common distance values in the intersection of all sets of a target point.
9. The machine vision-based operation panel sheet printing positioning method according to claim 1, characterized in that, Corner points on the connected component are obtained based on corner detection, including: The corner points on the connected domain are obtained using Harris corner detection.
10. The machine vision-based operation panel sheet printing positioning method according to claim 1, characterized in that, Acquiring images of the control panel sheet also includes: The image is denoised using Gaussian filtering.