Intelligent positioning method and system for protective film based on machine vision

Abstract

The present invention relates to the field of image processing. More specifically, the present invention relates to an intelligent positioning method and system for a protective film based on machine vision. The method includes: obtaining a template image of the protective film and a surface image of the protective film being conveyed in real time, respectively calculating the possibility of pixel points in the template image and the surface image being on the protective film; using the ORB algorithm to locate feature points and calculating the feature intensity of each pixel point within the neighborhood of the feature points. Based on the position of the feature points, the feature intensity of the pixel points within the neighborhood, and gradient information, it is used to calculate the feature uniqueness of each feature point. Feature points with a feature uniqueness less than the average feature uniqueness of all feature points are screened out, and the protective film is positioned for the screened feature points. The present invention improves the accuracy of ORB algorithm matching by calculating the feature uniqueness of feature points, retaining the feature points crucial for the positioning of the protective film, and removing unimportant or incorrect feature points at the same time.

Description

Technical Field

[0001] The present invention relates to the field of image processing. More specifically, the present invention relates to an intelligent positioning method and system for a protective film based on machine vision. Background Art

[0002] A protective film is a thin film material used to protect the surface of a product from scratches, contamination, and damage. It is widely used in the packaging and transportation processes of various products, especially in industries such as electronics, automotive, construction, and medical. In the process of manufacturing a mobile phone lens protective film, the protective film is usually conveyed by a vibrating disk to a vacuum suction nozzle adsorption area, and then the vacuum suction nozzle sucks the protective film and places it on the protective sticker. During this process, the conveyance by the vibrating disk is random, and the protective film may face any direction. However, the production specification requires that the protective film placed on the protective sticker needs to have the same orientation. Therefore, it is necessary to position the protective film in the vacuum suction nozzle adsorption area.

[0003] The existing Chinese patent application document with the publication number CN118310419A discloses a workpiece hole position center positioning method based on multi-scale image feature matching, which relates to the field of machine vision measurement technology. The method includes: acquiring a workpiece hole position image; cropping the workpiece hole position image to obtain a template image retaining the hole position edge information; extracting template features in the template image; performing a masking process on the template image to remove the area outside the hole position area; acquiring an image of the workpiece to be matched, and using the pyramid template feature matching algorithm to perform super-resolution multi-scale image feature matching on the template features and the image of the workpiece to be matched to obtain the hole position center position.

[0004] This application document solves the problems that in the regular geometric shape detection algorithm, special-shaped holes such as straight groove holes and square holes cannot be detected, and the hole position measurement accuracy decreases when the resolution of the camera lens decreases, thereby reducing the equipment selection cost. Currently, the ORB (Oriented Fast and Rotated Brief) algorithm is usually used for feature point matching, and the rotation angle of the protective film is calculated based on the coordinates of the matching point pairs for positioning. However, the existing technology of the ORB algorithm screens feature points through machine learning algorithms and non-maximum suppression, lacking a fine evaluation of the quality of feature points, which may cause feature points with obvious features of the protective film to be screened out, and thus may not be able to achieve high-quality positioning of the protective film. Summary of the Invention

[0005] To solve the problem that when using the ORB algorithm to screen feature points, there is a lack of a fine evaluation of the quality of feature points, which may cause feature points with obvious features of the protective film to be screened out, and thus may not be able to achieve high-quality positioning of the protective film, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a machine vision-based intelligent positioning method for a protective film includes: obtaining a template image of the protective film and a surface image of the continuously conveyed protective film in real time, and respectively calculating the possibility of pixel points in the template image and the surface image being on the protective film; using the ORB algorithm to extract feature points from the surface image to obtain a plurality of feature points and corresponding directions, and calculating the feature intensity of each pixel point within the neighborhood range of the feature points; based on the position of the feature points, the feature intensity of the pixel points within the neighborhood range, and the gradient information, for calculating the feature uniqueness of each feature point, screening out the feature points whose feature uniqueness is less than the average feature uniqueness of all feature points, and performing protective film positioning on the screened feature points; wherein, the feature intensity satisfies the following relational expression: ; where represents the feature intensity of the th pixel point within the neighborhood range of any feature point, represents the range difference of the possibility of all pixel points within the neighborhood range of any feature point being on the protective film, represents the number of times the feature value of the th pixel point within the neighborhood range of any feature point appears among the feature values of the pixel points within the neighborhood range of the feature point, represents the number of pixel points within the neighborhood range of the feature point, represents an exponential function with the natural number as the base.

[0007] The effect is that: by respectively calculating the possibility of pixel points in the template image and the surface image being on the protective film, the method can more accurately identify the actual position of the protective film. This possibility-based evaluation is more refined than traditional positioning methods, can reduce errors, and improve the positioning accuracy. By calculating the feature uniqueness of the feature points and screening out those feature points with uniqueness lower than the average value, noise and irrelevant feature points are removed, the feature point set is optimized, and the remaining feature points are more suitable for the precise positioning of the protective film.

[0008] Preferably, the possibility of the pixel point being on the protective film includes:

[0009] Taking any pixel point as the target pixel point, after normalizing the gray value of the target pixel point, using an exponential function for attenuation to obtain the saliency of the target pixel point;

[0010] Calculating the absolute difference between the target pixel point and the average gray value of the pixel points within the eight-neighborhood of the target pixel point, and after normalizing, using an exponential function for attenuation to obtain the contribution degree of the target pixel point;

[0011] Taking the contribution degree of the target pixel as the exponential power of saliency, the possibility that the target pixel is on the protective film is obtained.

[0012] Its effect is as follows: By analyzing the pixels with important visual features in the image, the boundary and important feature areas of the protective film can be better identified; calculating the absolute difference between the average gray value of the target pixel and the pixels in its eight-neighborhood, and using the exponential function to decay after normalization can effectively enhance the subtle contrast changes in the image, which helps to capture the possible details and texture features on the surface of the protective film, thereby improving the positioning accuracy.

[0013] Preferably, the possibility that the pixel is on the protective film further includes:

[0014] Taking any pixel as the target pixel, calculating the absolute difference between the target pixel and the average gray value of the pixels in the eight-neighborhood of the target pixel, and obtaining the contribution degree of the target pixel by the ratio with the gray-scale adjustment parameter;

[0015] Using the exponential function to exponentially decay the ratio between the gradient value of the target pixel and the gradient adjustment parameter to obtain the edge intensity of the target pixel;

[0016] Taking the product of the contribution degree and the edge intensity and normalizing it as the possibility that the target pixel is on the protective film.

[0017] Preferably, the feature distinctiveness satisfies the following relational expression:

[0018] ;

[0019] In the formula, represents the feature distinctiveness of any feature point, represents the number of pixels in the neighborhood range of the feature point, represents the cosine similarity between the direction of any feature point and the gradient direction of the th pixel in the neighborhood range, represents the feature intensity of the th pixel in the neighborhood range of any feature point, represents the number of feature points in the neighborhood range of any feature point, represents the total number of feature points in the image, represents the exponential function with the natural number as the base.

[0020] Preferably, the feature distinctiveness also satisfies the following relational expression:

[0021] ;

[0022] Wherein, represents the feature uniqueness of any feature point, represents the number of pixel points within the neighborhood range of the feature point, represents the cosine similarity between the direction of any feature point and the gradient direction of the th pixel point within the neighborhood range, represents the feature intensity of the th pixel point within the neighborhood range of any feature point, represents the th pixel point, represents the distance between the th pixel point and the feature point, represents the maximum distance within the neighborhood range of the feature point, represents the number of feature points within the neighborhood range of any feature point, represents the total number of feature points in the image,

[0023] Preferably, the positioning of the protective film for the screened feature points includes:

[0024] Using the ORB algorithm to extract feature points and generate descriptors for the template image and the surface image respectively. Using the descriptors of the feature points, a feature matching method is used to find matching pairs between the feature points of the template image and the surface image, and the rotation angle of the protective film in the vacuum chuck adsorption area relative to the preset position is calculated, and the protective film is positioned based on the feature points.

[0025] The effect is that by screening out those feature points with uniqueness lower than the average value, noise and irrelevant feature points are removed, so that the feature points can calculate the rotation angle of the protective film relative to the preset position according to the matching conditions, and the position of the protective film can be accurately adjusted to ensure its correct alignment.

[0026] Preferably, the rotation angle includes:

[0027] According to the fitting alignment angle algorithm, calculate the rotation angle of the protective film in the vacuum chuck area relative to the preset orientation, and rotate the protective film to the preset orientation based on the vacuum chuck.

[0028] In a second aspect, a protective film intelligent positioning system based on machine vision includes: a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned protective film intelligent positioning method based on machine vision is implemented.

[0029] The present invention has the following effects:

[0030] 1. The present invention enhances the adaptability of the positioning process to different lighting conditions, background interference, and changes in the surface characteristics of the protective film by calculating the possibility of a pixel point being on the protective film, combining the gray value, gradient information, and the local texture features of the pixel point. Through multi-dimensional feature evaluation, the positioning system can maintain high robustness in complex environments.

[0031] 2. By introducing the calculation of feature intensity and feature uniqueness, the present invention is conducive to more accurately evaluating the quality of feature points, avoiding the problem of important feature points being mis-screened due to the lack of fine evaluation in the ORB algorithm, retaining the feature points crucial for the positioning of the protective film, while removing unimportant or incorrect feature points, and improving the accuracy of the ORB algorithm matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] By referring to the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understandable. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0033] Figure 1 is a flowchart of the method from step S1 to step S3 in the intelligent positioning method of the protective film based on machine vision according to an embodiment of the present invention.

[0034] Figure 2 is a schematic diagram of the process of placing the protective film in the intelligent positioning system of the protective film based on machine vision according to an embodiment of the present invention.

[0035] Figure 3 is a schematic diagram of the protection orientation inside the vibrating bowl in the intelligent positioning system of the protective film based on machine vision according to an embodiment of the present invention.

[0036] Figure 4 is a schematic diagram of the production specification orientation of the protective film in the intelligent positioning system of the protective film based on machine vision according to an embodiment of the present invention.

[0037] Figure 5 is a structural block diagram of the intelligent positioning system of the protective film based on machine vision according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0039] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0040] In a specific implementation scenario, during the production of the mobile phone lens protective film, the protective film is usually conveyed by a vibrating disk to the vacuum suction nozzle adsorption area, and then the vacuum suction nozzle sucks the protective film and places the protective film on the protective sticker. Refer to Figure 2 , during this process, the conveying of the vibrating disk is random, and the protective film may face any direction. Refer to Figure 3 , however, the production specification requires that the protective film placed on the protective sticker needs to have the same orientation. Refer to Figure 4 , therefore, it is necessary to position the protective film in the vacuum suction nozzle adsorption area to determine the orientation of the protective film in the vacuum suction nozzle adsorption area, and then use the vacuum suction nozzle with a rotation function to unify the orientation of the protective film.

[0041] Refer to Figure 1 , the intelligent positioning method of the protective film based on machine vision includes steps S1 - S3, specifically as follows:

[0042] S1: Obtain the template image of the protective film and the surface image of the continuously conveyed protective film, and calculate the possibility that the pixel points in the template image and the surface image are on the protective film respectively.

[0043] Taking any pixel point as the target pixel point, after normalizing the gray value of the target pixel point, use the exponential function for attenuation to obtain the saliency of the target pixel point;

[0044] Calculate the absolute difference between the target pixel point and the average gray value of the pixel points in the eight - neighborhood of the target pixel point, and after normalizing, use the exponential function for attenuation to obtain the contribution degree of the target pixel point;

[0045] Take the contribution degree of the target pixel point as the exponential power of the saliency to obtain the possibility that the target pixel point is on the protective film.

[0046] Specifically, the possibility that the pixel point is on the protective film satisfies the following relational expression:

[0047] ;

[0048] In the formula, represents the possibility that the th pixel point in the image is on the protective film, represents the gray value of the th pixel point, represents the average gray value of the pixel points in the eight - neighborhood of the th pixel point, represents the exponential function with the natural number as the base.

[0049] That is to say, represents the relative value of the gray value of any pixel point in the image. Since the protective film border is black with a low gray value, while the image background is bright with a high gray value, therefore, generally, the larger it is, the more likely the pixel point is on the image background, and the less likely the pixel point is on the protective film. The smaller it is, the more likely the pixel point is on the protective film in the image, and the greater the possibility that the pixel point is on the protective film;

[0050] It should be noted that there are texture lines generated during the production process on the protective film border. The texture lines will reflect light under illumination and also form bright areas in the image, which affects the judgment of whether the pixel point is on the protective film;

[0051] Furthermore, represents the difference between the gray value of any pixel point in the image and the gray values of its neighboring pixel points. The larger this value is, the more likely the pixel point is a pixel point in the reflective area of the texture lines on the protective film in the image. The smaller this value is, the more likely the pixel point is a pixel point in the normal area or background area inside the protective film in the image. Therefore, by using the form of gamma transformation to make corrections, the larger it is, then the smaller it is. At this time, the correction effect on is greater, making the pixel points in the reflective area of the texture lines on the protective film more likely to be on the protective film. the smaller it is, then the larger it is. At this time, the correction effect on is smaller, avoiding the possibility that the pixel points in the normal area on the protective film in the image are overly affected.

[0052] In addition, in another embodiment, it further includes:

[0053] Taking any pixel point as the target pixel point, calculating the absolute difference between the target pixel point and the average gray value of the pixel points in the eight-neighborhood of the target pixel point, and obtaining the contribution degree of the target pixel point by the ratio with the gray adjustment parameter;

[0054] Using an exponential function to perform exponential decay on the ratio between the gradient value of the target pixel point and the gradient adjustment parameter to obtain the edge strength of the target pixel point;

[0055] Taking the product of the contribution degree and the edge strength and performing normalization processing as the possibility that the target pixel point is on the protective film.

[0056] Specifically, the possibility that a pixel point is on the protective film satisfies the following relational expression:

[0057] ;

[0058] In the formula, represents the possibility that the -th pixel point in the image is on the protective film, represents the gray value of the -th pixel point, represents the average gray value of the pixel points within the eight-neighborhood of the -th pixel point, represents the gray-scale adjustment parameter, represents the gradient value of the -th pixel point, represents the gradient adjustment parameter, represents the exponential function with the natural number as the base.

[0059] That is to say, , , and implementers can adjust according to specific situations.

[0060] Furthermore, although the orientation position of the protective film is random when it is in the vacuum suction nozzle adsorption area, the shape of the protective film is fixed, and the background of image acquisition is also fixed. Then, when the protective film faces different positions, the features within the neighborhood range of each pixel point on the protective film in the image are also relatively fixed. Feature points can be screened based on this characteristic. Therefore, in the present invention, the feature intensity of each pixel point within the neighborhood range of a feature point is calculated according to the possibility that the pixel points within the neighborhood range of the feature point are on the protective film and the feature values of the pixel points within the neighborhood range of the feature point.

[0061] S2: Use the ORB algorithm to extract feature points from the surface image, obtain multiple feature points and corresponding directions, and calculate the feature intensity of each pixel point within the neighborhood range of the feature points.

[0062] The feature intensity includes:

[0063] Taking any feature point as the center and a preset range as the neighborhood range of the feature point, use the local binary pattern algorithm with eight-point sampling to obtain the feature values of each pixel point in the image, count the number of occurrences of the feature values of each pixel point within the neighborhood range and the total number of pixel points within the neighborhood range, and obtain the range of the possibilities that all pixel points within the neighborhood range of any feature point are on the protective film;

[0064] Exemplarily, the preset range is , and implementers can adjust according to the actual situation.

[0065] Calculate the frequency of occurrence of the eigenvalue of each pixel point within the neighborhood range, and use exponential decay to obtain the particularity of the feature point. Calculate the product between the particularity and the range difference to obtain the feature intensity of the pixel points within the neighborhood range of any feature point.

[0066] Specifically, the feature intensity satisfies the following relational expression:

[0067] ;

[0068] Among them, represents the feature intensity of the th pixel point within the neighborhood range of any feature point, represents the range difference of the possibility that all pixel points within the neighborhood range of any feature point are on the protective film, represents the number of times the eigenvalue of the th pixel point within the neighborhood range of any feature point appears among the eigenvalues of the pixel points within the neighborhood range of the feature point, represents the number of pixel points within the neighborhood range of the feature point, represents the exponential function with the natural number as the base.

[0069] That is to say, the frequency of occurrence of the eigenvalue of the th pixel point within the neighborhood range of the feature point. The larger this value is, the more likely it is that the th pixel point within the neighborhood range of the feature point has the same characteristics as more pixel points, then the feature of this pixel point is more common, and then the feature intensity of the th pixel point within the neighborhood range of the feature point is smaller. The smaller this value is, the more likely it is that the th pixel point within the neighborhood range of the feature point has the same characteristics as fewer pixel points, then the feature of this pixel point is less common, and then the feature intensity of the th pixel point within the neighborhood range of the feature point is larger.

[0070] Furthermore, it shows that when the range difference of the possibility that all pixel points within the neighborhood range of the feature point are on the protective film is larger, it means that the feature point is more likely to be on the boundary between the protective film and the image background in the image. Since the outer contour of the protective film has obvious boundary features, the pixel points within the neighborhood range of the feature point are more likely to have unique features. Therefore, when it is larger, the feature intensity of the th pixel point within the neighborhood range of the feature point is larger, The smaller it is, the more likely it indicates that the feature point is inside the protective film material in the image or in the image background. Since the pixel points within the neighborhood of the feature point inside the protective film material or in the image background are similar, the pixel points within the neighborhood of this feature point may not have unique features. Therefore When it is smaller, the th pixel point within the neighborhood of the feature point has a smaller feature intensity.

[0071] S3: Based on the position of the feature point, the feature intensity of the pixel points within the neighborhood, and the gradient information, for calculating the feature uniqueness of each feature point, screen out the feature points whose feature uniqueness is less than the average feature uniqueness of all feature points, and perform protective film positioning on the screened feature points.

[0072] Specifically, the feature uniqueness satisfies the following relational expression:

[0073] ;

[0074] In the formula, represents the feature uniqueness of any feature point, represents the number of pixel points within the neighborhood of the feature point, represents the cosine similarity between the direction of any feature point and the gradient direction of the th pixel point within the neighborhood, represents the feature intensity of the th pixel point within the neighborhood of any feature point, represents the number of feature points within the neighborhood of any feature point, represents the total number of feature points in the image, represents the exponential function with the natural number as the base.

[0075] That is to say, represents the consistency between the pixel points within the neighborhood of the feature point and the direction of the feature point. The larger it is, the higher the consistency between the direction of the feature point and the gradient direction of the pixel points within the neighborhood, the more likely the feature point has unique features, and the greater the feature uniqueness of the feature point. The smaller it is, the lower the consistency between the direction of the feature point and the gradient direction of the pixel points within the neighborhood, the less likely there are unique features within the neighborhood of the feature point, and the smaller the feature uniqueness of the feature point. For the convenience of calculation, is normalized by adding 1 and dividing by 2 in the formula.

[0076] It should be noted that the feature intensity of the th pixel point within the neighborhood of the feature point The larger it is, the more likely the feature points containing this pixel point in the neighborhood range have unique features, and the greater the feature uniqueness of the feature points containing this pixel point in the neighborhood range. The smaller it is, the less prominent the features of the feature points containing this pixel point in the neighborhood range are likely to be, and the smaller the feature uniqueness of the feature points containing this pixel point in the neighborhood range.

[0077] It should also be noted that represents the relative value of the number of feature points in the neighborhood range of the feature point. Since the shape mutation area of the protective film in the image is more likely to generate feature points, and these feature points play a greater role in the positioning of the protective film, and when it is larger, it indicates that the feature is more likely to be in the shape mutation area of the protective film. In order to ensure that these feature points can be well retained, it is corrected through the form of gamma transformation by for to be corrected. When it is larger, it is smaller, for the greater the correction effect on When it is smaller, it is larger, for the smaller the positive effect on

[0078] In addition, in another embodiment, it further includes:

[0079] The feature uniqueness also satisfies the following relational expression:

[0080] ;

[0081] In the formula, represents the feature uniqueness of any feature point, represents the number of pixel points in the neighborhood range of the feature point, represents the cosine similarity between the direction of any feature point and the gradient direction of the th pixel point in the neighborhood range, represents the feature intensity of the th pixel point in the neighborhood range of any feature point, represents the th pixel point and the distance between the feature point, represents the maximum distance in the neighborhood range of the feature point, represents the number of feature points in the neighborhood range of any feature point, represents the total number of feature points in the image, represents the exponential function with the natural number as the base.

[0082] That is to say, represents the relative magnitude of the distance between the th pixel point within the neighborhood of the feature point and the feature point. The larger it is, the greater the influence of the th pixel point within the neighborhood of the feature point on the feature point. The smaller it is, the relatively smaller the influence of the th pixel point within the neighborhood of the feature point on the feature point. Therefore, by performing correction on When it is smaller, the correction degree for is greater, making the pixels closer to the feature point have a greater influence on the feature uniqueness of the feature point. When it is larger, the

[0083] Use the ORB algorithm to perform feature point extraction and descriptor generation on the template image and the surface image respectively. Utilize the descriptors of the feature points, use the feature matching method to find matching pairs between the feature points of the template image and the surface image, and calculate the rotation angle of the protective film in the vacuum suction nozzle adsorption area relative to the preset position, and perform protective film positioning based on the feature points.

[0084] It should be noted that in this embodiment, the feature matching method is the nearest neighbor matching, which is a well-known technology to those skilled in the art and will not be described in detail.

[0085] The rotation angle includes:

[0086] According to the fitting alignment angle algorithm, calculate the rotation angle of the protective film in the vacuum suction nozzle area relative to the preset orientation, and rotate the protective film to the preset orientation based on the vacuum suction nozzle.

[0087] Finally, it should be noted that in the protective film positioning scenario, the height of the vacuum suction nozzle adsorption area is fixed, the size of the protective film is fixed, and the image acquisition angle is fixed. That is to say, the size of the protective film in the image is the same. Therefore, in order to improve the protective film positioning efficiency, the present invention cancels the process of constructing an image pyramid in the existing ORB algorithm to achieve feature point scale invariance.

[0088] The present invention also provides a smart protective film positioning system based on machine vision. As Figure 5As shown, the system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the intelligent positioning method of the protective film based on machine vision according to the first aspect of the present invention.

[0089] The system also includes other components well-known to those skilled in the art, such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be elaborated here.

[0090] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or apparatus. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device. Any application or module described in the present invention can be implemented using computer-readable / executable instructions that can be stored or otherwise held by such a computer-readable medium.

[0091] In the description of this specification, the meanings of "a plurality" and "several" are at least two, such as two, three, or more, etc., unless otherwise specifically and clearly defined.

[0092] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

Claims

1. An intelligent positioning method for protective films based on machine vision, characterized in that Including: Obtain the template image of the protective film and the real-time surface image of the conveyed protective film, and calculate the possibility of the pixel points in the template image and the surface image being on the protective film respectively; Taking any pixel point as the target pixel point, after normalizing the gray value of the target pixel point, use the exponential function for attenuation to obtain the saliency of the target pixel point; Calculate the absolute difference between the target pixel point and the average gray value of the pixel points in the eight-neighborhood of the target pixel point, and after normalization, use the exponential function for attenuation to obtain the contribution degree of the target pixel point; Take the contribution degree of the target pixel point as the exponential power of the saliency to obtain the possibility of the target pixel point being on the protective film; Use the ORB algorithm to extract feature points from the surface image, obtain multiple feature points and corresponding directions, and calculate the feature intensity of each pixel point within the neighborhood range of the feature points; Based on the position of the feature points, the feature intensity of the pixel points within the neighborhood range, and the gradient information, to calculate the feature uniqueness of each feature point, screen out the feature points whose feature uniqueness is less than the average feature uniqueness of all feature points, and perform protective film positioning on the screened feature points; The uniqueness of the feature satisfies the following relationship: ; where represents the uniqueness of any feature point, represents the number of pixel points within the neighborhood range of the feature point, represents the cosine similarity between the direction of any feature point and the gradient direction of the th pixel point within the neighborhood range, represents the intensity of the th pixel point within the neighborhood range of any feature point, represents the number of feature points within the neighborhood range of any feature point, represents the total number of feature points in the image, represents the exponential function with the natural number as the base; Among them, the intensity of the feature satisfies the following relational expression: ; in the formula, represents the intensity of the th pixel point within the neighborhood range of any feature point, represents the range difference of the probabilities of all pixel points within the neighborhood range of any feature point being on the protective film, represents the number of times the feature value of the th pixel point within the neighborhood range of any feature point appears among the feature values of the pixel points within the neighborhood range of the feature point, represents the number of pixel points within the neighborhood range of the feature point, represents the exponential function with the natural number as the base.

2. The intelligent positioning method of the protective film based on machine vision according to claim 1, wherein The possibility that the pixel point is on the protective film can also be replaced by: Taking any pixel point as the target pixel point, calculate the absolute difference between the target pixel point and the average gray value of the pixel points in the eight-neighborhood of the target pixel point, and obtain the contribution degree of the target pixel point by the ratio with the gray value adjustment parameter; Use the exponential function to perform exponential attenuation on the ratio between the gradient value of the target pixel point and the gradient adjustment parameter to obtain the edge intensity of the target pixel point; Take the product of the contribution degree and the edge intensity and perform normalization processing as the possibility that the target pixel point is on the protective film.

3. The intelligent positioning method of the protective film based on machine vision according to claim 1, wherein, The feature uniqueness can also be replaced by the following relational expression: ； Wherein, represents the feature uniqueness of any feature point, represents the number of pixel points within the neighborhood range of the feature point, represents the cosine similarity between the direction of any feature point and the gradient direction of the th pixel point within the neighborhood range, represents the feature intensity of the th pixel point within the neighborhood range of any feature point, represents the th pixel point and the distance between the feature points, represents the maximum distance within the neighborhood range of the feature point, represents the number of feature points within the neighborhood range of any feature point, represents the total number of feature points in the image, represents the exponential function with the natural number as the base.

4. The intelligent positioning method of the protective film based on machine vision according to claim 1, characterized in that The performing protective film positioning on the screened feature points includes: Use the ORB algorithm to extract feature points and generate descriptors of the template image, use the descriptors of the feature points, use the feature matching method to find matching pairs between the feature points of the template image and the screened feature points of the surface image, and calculate the rotation angle of the protective film in the vacuum chuck adsorption area relative to the preset position, and perform protective film positioning based on the feature points.

5. The intelligent positioning method of the protective film based on machine vision according to claim 4, wherein The rotation angle includes: According to the fitting alignment angle algorithm, calculate the rotation angle of the protective film in the vacuum chuck area relative to the preset orientation, and rotate the protective film to the preset orientation based on the vacuum chuck.

6. The intelligent positioning system for protective films based on machine vision is characterized in that, Including: A processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent positioning method of the protective film based on machine vision according to any one of claims 1-5 is realized.

Images