Visual inspection method and system for mobile phone shell silk screen printing process and mobile phone shell

Abstract

This invention relates to the fields of signal transmission, computer vision, image recognition, and printing. It provides a visual inspection method, system, and mobile phone casing for screen printing processes. The method uses a screen printing image recognition algorithm to identify the aperture spacing and mesh size of the screen printing stencil. When the aperture spacing and mesh size are compliant, it prompts the assembly of a compliant screen printing stencil to the printing press for printing the base color layer of the substrate. The substrate, after being left to stand and bake, is then transported to a CCD inspection line for CCD imaging. The system checks for jagged edges, smearing, and light leaks in the characters of the ink-printed layer on the substrate, and determines whether the brightness of the ink in the ink-printed layer is compliant. This improves the efficiency and accuracy of detecting the specifications, serial numbers, aperture spacing, and mesh size of the screen printing stencil, and enhances the efficiency and accuracy of detecting jagged edges, smearing, and light leaks in the characters of the ink-printed layer on the substrate, as well as whether the ink brightness meets requirements, thereby achieving high-quality mobile phone casing production.

Description

Technical Field

[0001] The present invention relates to the technical fields of signal transmission, computer vision, image recognition and printing, and in particular to a visual inspection method for a mobile phone shell silk screen printing process. Background Art

[0002] In the silk screen printing process of mobile phone shells, it generally includes processes such as silk screen preparation, ink silk screen printing, static, baking and laminating. When preparing the silk screen, it is necessary to detect whether the hole spacing and mesh number of the silk screen meet the requirements. After the silk screen is prepared, the configured ink is printed on the base color layer of the substrate (for example, the green film surface of the substrate) through the silk screen, and then the ink printing layer is covered with a frosted film after static and baking. In the process of configuring the ink to be printed on the base color layer of the substrate through the silk screen, it is necessary to detect whether the characters of the ink printing layer of the substrate have tooth edges, fat oil and light leakage points, and detect whether the brightness of the ink meets the requirements. In the prior art, the detection of the hole spacing and mesh number of the silk screen is manual visual inspection, and the detection of whether the characters of the ink printing layer of the substrate have tooth edges, fat oil and light leakage points, and whether the brightness of the ink meets the requirements is also manual visual inspection, which not only has low detection efficiency, but also low detection accuracy, which is not conducive to the production of high-quality mobile phone shells.

[0003] In summary, in the prior art, the detection of the hole pitch and mesh number of the silk screen printing screen, whether the characters of the ink printing layer of the substrate have jagged edges, fat oil and light leakage spots, and whether the brightness of the ink meets the requirements has technical problems such as low detection efficiency and low detection accuracy, which is not conducive to the production of high-quality mobile phone casings. Summary of the invention

[0004] In view of the deficiencies in the above-mentioned prior art, the present invention provides a visual inspection method and system for the silk-screen printing process of mobile phone casings, so as to improve the inspection efficiency and accuracy of the specifications, numbers, hole pitches and mesh counts of the silk-screen printing screens, and to improve the inspection efficiency and accuracy of whether the characters on the ink printing layer of the substrate have teeth, fat oil and light leakage spots, and whether the brightness of the ink meets the requirements, so as to realize the production of high-quality mobile phone casings.

[0005] In a first aspect, the present invention provides a visual inspection method for a mobile phone housing silk screen printing process, comprising:

[0006] Before assembling the screen printing screen to the printing machine, the screen printing screen is photographed to obtain a screen image, wherein the screen image reflects the hole pitch and mesh number of the screen printing screen;

[0007] Identify the hole pitch and mesh of the screen printing screen by using a screen image recognition algorithm, determine whether the hole pitch and mesh of the screen printing screen are compliant, and when the hole pitch and mesh of the screen printing screen are compliant, prompt to assemble the compliant screen printing screen to the printing machine to print the base color layer of the substrate;

[0008] After the base color layer of the substrate is printed with ink, the substrate after the base color layer is left to stand, and the substrate after standing to stand is baked, and the substrate after standing to stand and bake is transported to a CCD detection production line for CCD imaging to obtain an image of the ink printing layer of the substrate;

[0009] The ink printing layer of the substrate is identified by an ink printing layer identification algorithm, and it is determined whether the characters of the ink printing layer of the substrate have tooth edges, fat oil and light leakage points, and whether the brightness of the ink of the ink printing layer of the substrate is in compliance with regulations.

[0010] In a second aspect, the present invention provides a visual inspection system for a mobile phone shell silk screen process, wherein the visual inspection system uses the above-mentioned visual inspection method for a mobile phone shell silk screen process.

[0011] In a third aspect, the present invention provides a mobile phone casing, wherein the mobile phone casing is inspected in a silk-screen printing process using the above-mentioned visual inspection method for the silk-screen printing process of a mobile phone casing.

[0012] Compared with the prior art, the present invention has the following beneficial effects:

[0013] The present invention provides a visual inspection method, system and mobile phone shell for the silk screen printing process of a mobile phone shell. Before assembling the silk screen screen to a printing machine, the silk screen screen is imaged to obtain a screen image, wherein the screen image reflects the hole pitch and mesh number of the silk screen screen, the hole pitch and mesh number of the silk screen screen are identified by a screen image recognition algorithm, and it is judged whether the hole pitch and mesh number of the silk screen screen are compliant. When the hole pitch and mesh number of the silk screen screen are compliant, a prompt is given to assemble the compliant silk screen screen to the printing machine to print the base color layer of a substrate. After the base color layer of the substrate is printed with ink, the substrate after the base color layer is printed is allowed to stand, and the base color layer is printed. The substrate is baked after standing still, and then transported to a CCD inspection production line for CCD imaging to obtain an ink printing layer image of the substrate. The ink printing layer of the substrate is identified by an ink printing layer recognition algorithm to determine whether the characters on the ink printing layer of the substrate have teeth, fat, and light leakage spots, and whether the brightness of the ink on the ink printing layer of the substrate meets the requirements, thereby improving the inspection efficiency and accuracy of the specifications, numbers, hole pitches, and mesh counts of the silk screen printing screen, and improving the inspection efficiency and accuracy of whether the characters on the ink printing layer of the substrate have teeth, fat, and light leakage spots, and whether the brightness of the ink meets the requirements, thereby achieving high-quality mobile phone casing production. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The drawings described herein are used to provide further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0015] Figure 1 It is a flow chart of a visual inspection method for a mobile phone housing silk screen printing process according to an embodiment of the present invention;

[0016] Figure 2 This is a schematic diagram of the architecture of a visual inspection system for a mobile phone housing silk screen printing process according to an embodiment of the present invention;

[0017] Figure 3 This is a schematic diagram of a process flow of a substrate from a static station to a baking station according to an embodiment of the present invention;

[0018] Figure 4 It is a schematic diagram of a process flow of a substrate from a baking station to a CCD detection production line according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0020] See also Figure 1-Figure 4 The embodiment of the present invention provides a visual inspection method for the silk-screen process of a mobile phone shell, a system and a mobile phone shell. The visual inspection system for the silk-screen process of a mobile phone shell uses the above-mentioned visual inspection method for the silk-screen process of a mobile phone shell. The mobile phone shell is inspected in the silk-screen process using the above-mentioned visual inspection method for the silk-screen process of a mobile phone shell. The visual inspection method for the silk-screen process of a mobile phone shell comprises the following steps:

[0021] S101, before assembling the screen printing screen to the printing machine, photographing the screen printing screen to obtain a screen image, wherein the screen image reflects the hole pitch and mesh number of the screen printing screen;

[0022] S102, identifying the hole pitch and mesh of the screen printing screen by a screen image recognition algorithm, determining whether the hole pitch and mesh of the screen printing screen are compliant, and when the hole pitch and mesh of the screen printing screen are compliant, prompting the user to assemble the compliant screen printing screen to a printing machine to print a base color layer of a substrate;

[0023] S103, after the base color layer of the substrate is printed with ink, the substrate after the base color layer is printed is allowed to stand, and the substrate after the standing is baked is transported to a CCD detection production line for CCD imaging to obtain an image of the ink printing layer of the substrate;

[0024] S104, identifying the ink printing layer of the substrate by using an ink printing layer identification algorithm, determining whether the characters of the ink printing layer of the substrate have tooth edges, fat oil and light leakage points, and determining whether the brightness of the ink of the ink printing layer of the substrate is in compliance with regulations.

[0025] It should be noted that, in this embodiment, the screen printing screen is photographed before being assembled to the printing machine to obtain a screen image, the screen image reflects the hole pitch and mesh number of the screen, the screen image recognition algorithm is used to identify the hole pitch and mesh number of the screen, and it is determined whether the hole pitch and mesh number of the screen are compliant. When the hole pitch and mesh number of the screen are compliant, it is prompted to assemble the compliant screen to the printing machine to print the base color layer of the substrate. After the base color layer of the substrate is printed with ink, the substrate after the base color layer is printed is allowed to stand, and the substrate after standing is subjected to Baking, conveying the substrate after static baking to the CCD detection production line for CCD imaging to obtain the ink printing layer image of the substrate, identifying the ink printing layer of the substrate through the ink printing layer recognition algorithm, judging whether the characters of the ink printing layer of the substrate have teeth, fat and light leakage points, and judging whether the brightness of the ink of the ink printing layer of the substrate meets the requirements, thereby improving the detection efficiency and accuracy of the specifications, numbers, hole pitches and mesh counts of the silk screen printing screen, improving the detection efficiency and accuracy of whether the characters of the ink printing layer of the substrate have teeth, fat and light leakage points, and whether the brightness of the ink meets the requirements, and realizing high-quality mobile phone casing production.

[0026] It should be noted that, in this embodiment, the image capturing component can be used to capture the image of the screen printing screen to obtain the screen image, and the image capturing component then transmits the screen image to the host end, and the host end runs the screen image recognition algorithm to perform corresponding recognition and judgment of the hole pitch and mesh size of the screen printing screen. When the hole pitch and mesh size of the screen printing screen are compliant, it is prompted to assemble the compliant screen printing screen to the printer to print the base color layer of the substrate. In addition, the host end also runs the ink printing layer recognition algorithm to perform corresponding recognition and judgment on the characters of the ink printing layer of the substrate, and judges the brightness of the ink of the ink printing layer of the substrate, thereby achieving printing quality control.

[0027] In step S101, before the screen printing screen is assembled to the printing machine, the hole spacing and mesh number information of the screen are collected by image shooting, and the specification information of the screen printing screen is digitized to ensure that each screen parameter (such as hole spacing and mesh number) can be accurately identified, reduce the errors that may be caused by manual inspection, and avoid the assembly of unqualified screens. It can be understood that in this embodiment, automated image acquisition can provide basic data for subsequent efficient detection processes. It should be noted that the hole spacing refers to the distance between the centers of two adjacent holes, that is, the arrangement spacing of the holes on the screen printing screen. The hole spacing affects the accuracy of the screen and the clarity of the pattern. Reasonable hole spacing ensures that the ink can pass through the screen evenly to form a clear and accurate printing pattern. The size of the hole spacing is directly related to the fluidity and uniformity of the ink during printing. If the hole spacing is uneven, the distribution of the ink will be inconsistent, affecting the printing quality. The mesh refers to the number of holes per unit area, and the mesh number determines the fineness of the screen and the penetration of the ink. The stencil with high mesh count has finer holes, which is suitable for printing fine patterns, but the amount of ink passing through is less; the stencil with low mesh count is suitable for coarser patterns, and the amount of ink passing through is larger. According to the fineness of the pattern and the characteristics of the ink, the appropriate mesh count can be selected to ensure that the ink can evenly cover the pattern area according to the design requirements during printing.

[0028] In step S102, the stencil image recognition algorithm can automatically identify the hole pitch and mesh count of the stencil while quickly determining whether the stencil parameters meet the regulatory requirements, indicating the compliance or non-compliance status, avoiding manual item-by-item verification, reducing the risk of rework due to incorrect stencil assembly, and ensuring the smooth progress of subsequent processes, thereby effectively improving detection accuracy, reducing manual judgment errors, and ultimately improving production efficiency and product qualification rates.

[0029] In step S103, after the substrate is left to stand and baked, the image of the ink printing layer is collected by CCD (charge coupled device) imaging, which can achieve automated and detailed image collection.

[0030] CCD detection has extremely high resolution and can clearly capture the edges and details of the ink layer. It is particularly suitable for detecting tiny defects such as tooth edges, fat oil and light leakage points. In this embodiment, the use of CCD imaging detection can avoid the omission of defects that are difficult to detect with the naked eye, improve the detection accuracy, and reduce the time of manual inspection, thereby improving efficiency. It should be noted that the ink layer that has just been printed is usually in a wet state and has fluidity. By standing for a period of time, the ink is fully settled and evenly distributed on the surface of the substrate, which can reduce the unevenness of the printed layer and prevent the pattern from being blurred or the contour edge from being unclear due to flow. In addition, standing can also reduce problems such as bubbles, particles and printing defects caused by the instability of the ink, making the ink layer smoother, and providing a more ideal foundation for subsequent baking and imaging. The ink needs to be cured by baking after standing to ensure that it is firmly attached and achieves the final color and brightness effect. Baking can accelerate the volatilization of the solvent in the ink, promote drying and curing, thereby improving the durability, gloss and touch of the ink layer. The ink layer after baking is more stable and is not easily affected by external forces such as friction and collision, which prevents the printed layer from being scratched, discolored, and other problems in subsequent operations. After baking, the color, brightness, and edge of the ink layer have reached the final effect, which is suitable for high-precision detection. The optical properties of the ink layer at this time are more stable, which is convenient for the CCD imaging system to accurately capture details. Through CCD imaging, defects in the ink layer, such as tooth edges, fat oil, light leakage points, etc., can be clearly detected to ensure the accuracy and reliability of the detection. Undried ink may reflect light and flow, affecting the imaging quality and causing detection errors. In this embodiment, standing, baking, and CCD detection are carried out step by step, so that each step of the processing can achieve the best effect, ensure the consistency of the ink layer in deposition, curing, and imaging, improve the stability of the production line and the accuracy of the detection results, and facilitate the production of high-quality mobile phone shells.

[0031] In step S104, the ink printing layer recognition algorithm is used to automatically detect the integrity and accuracy of the characters, and to identify defects such as tooth edges, fat oil, and light leakage points, which can greatly improve the speed and accuracy of detection. The algorithm can also automatically identify whether the brightness of the ink is compliant through image data, avoiding subjective errors caused by manual judgment of brightness. In this embodiment, through the automated image recognition algorithm, it can ensure that all printing layer defects are identified with high precision, and can quickly screen out qualified or unqualified products, significantly improving the overall quality control level of the production line.

[0032] It should be noted that, in the present embodiment, the tooth edge refers to the irregular state, uneven state, and the presence of fine burrs or ripples on the edge of the characters of the ink printing layer. Fat refers to the excessive accumulation of ink in the character part of the ink layer, and its defect characteristics usually include character deformation and ink overflowing the character outline. Fat is usually caused by excessive ink or excessive printing pressure. The fat phenomenon will affect the clarity and overall appearance of the characters. Light leakage points refer to the state where small gaps or uncovered areas appear on the characters. Light leakage points are usually caused by blockage of holes on the screen, uneven ink, etc. Ink brightness refers to the reflectivity or glossiness of the ink printing layer, which is the degree of light and dark displayed by the ink after printing. The brightness will affect the visual effect of the pattern or character. The brightness is usually determined by factors such as the ink formula, baking temperature and time. Inappropriate brightness will make the printed layer look too dim or insufficiently reflective, thereby reducing the appearance quality requirements.

[0033] In some preferred embodiments, when the screen image recognition algorithm identifies that the hole pitch and mesh of the screen printing screen are not compliant, it indicates that the screen printing screen has defects, and controls the display terminal to display the types of defects in the screen printing screen. It should be noted that when the screen is identified as non-compliant, the defect is immediately prompted, and the defect type is displayed in detail on the display terminal, which can provide instant feedback to the operator and prevent non-compliant screens from entering the next process. At this time, prompting the screen for defects ensures that only compliant screens are used on the production line, which helps reduce rework or waste caused by incorrect assembly and improves production efficiency.

[0034] In some preferred embodiments, before the control display terminal is used to display the defect type existing in the silk screen screen, the causes of different types of defects of the silk screen screen are intelligently judged, and when the control display terminal is used to display the defect type existing in the silk screen screen, the display terminal is simultaneously controlled to display the causes of the defects of the silk screen screen. It should be noted that in this embodiment, by displaying the causes of the defects, the operator can clearly understand the causes of the specific defects, so as to directly take targeted adjustment measures. Compared with simply displaying the defect type, displaying the causes is more helpful in quickly locating the source of the problem, accelerating the problem solving process, and reducing production delays caused by improper troubleshooting.

[0035] In some preferred embodiments, when the screen image recognition algorithm determines whether the hole spacing of the silk screen screen is compliant, it includes: comparing the identified hole spacing of the silk screen screen with the standard hole spacing of the silk screen screen in the compliance database. If the identified hole spacing of the silk screen screen matches the standard hole spacing of the silk screen screen in the compliance database, then the identified hole spacing of the silk screen screen is judged to be compliant, otherwise it is not compliant. It should be noted that in this embodiment, using the standard hole spacing in the compliance database as the basis for judgment can ensure clear compliance standards during the detection process. Standardized comparison methods can reduce the subjectivity of manual judgment and improve the accuracy and consistency of detection. Automatically comparing the identified hole spacing with the standard value in the database can realize automated compliance detection and reduce errors in manual judgment.

[0036] In some preferred embodiments, when the screen image recognition algorithm determines whether the mesh of the silk screen screen is compliant, it includes: comparing the mesh of the silk screen screen identified with the standard mesh of the silk screen screen in the compliance database, if the mesh of the silk screen screen identified matches the standard mesh in the compliance database, then the mesh of the silk screen identified is judged to be compliant, otherwise it is not compliant. It should be noted that using the standard mesh in the compliance database as a basis for judgment can provide a unified benchmark and improve detection accuracy. The standardized comparison method eliminates subjectivity and inconsistency in manual judgment, ensuring that the results of mesh detection are accurate and reliable. By automatically comparing and identifying the mesh with the standard mesh, rapid automated detection can be achieved, reducing errors caused by human judgment, making the entire process more efficient and suitable for the needs of large-scale production lines.

[0037] In some improved embodiments, when judging whether the hole spacing and mesh of the silk screen screen are compliant by the screen image recognition algorithm, first judge whether the hole spacing of the silk screen screen is compliant, and when the hole spacing of the silk screen screen is compliant, then judge whether the mesh of the silk screen screen is compliant. It should be noted that the hole spacing of the silk screen screen determines the arrangement spacing of the holes on the screen, which is a basic parameter of the screen structure. If the hole spacing is not compliant, the hole layout of the entire screen will be inaccurate, which will cause the distribution of the mesh to be unbalanced. Therefore, first detecting the hole spacing can ensure that the basic structure of the screen is compliant, providing a reliable foundation for subsequent mesh detection. If the hole spacing of the screen is already non-compliant, the result of the mesh detection is meaningless even if it is compliant. Therefore, first detecting the hole spacing can quickly screen out unqualified screens and avoid wasting resources for further mesh detection. This can reduce unnecessary detection steps and improve detection efficiency.

[0038] In some further preferred embodiments, the hole pitch and mesh of the screen printing screen are identified by the screen image recognition algorithm, and when judging whether the hole pitch and mesh of the screen printing screen are compliant, it includes: extracting the hole edge features in the screen image by the convolution layer of the CNN in the CNN detection model, and the hole edge features include the position and arrangement pattern of the holes; using the target detection algorithm (such as YOLO) to identify the position of each hole and calculate the hole pitch of adjacent holes; comparing the calculated hole pitch with the standard hole pitch in the database, if all the identified hole pitches meet the standard, the hole pitch is determined to be compliant; otherwise, the hole pitch is determined to be non-compliant; when the hole pitch is compliant, the screen image is divided into multiple small areas (for example, a grid of a certain area), and the CNN detection model is used to count the number of holes in each area; the number of holes in each area is summarized, and the total number of holes per unit area is calculated to determine the mesh number; comparing the identified mesh number with the value in the standard database, if the mesh number meets the standard, the mesh is determined to be compliant, otherwise it is determined to be non-compliant. It should be noted that in this embodiment, an orderly screening process can be achieved by first detecting the hole pitch and then detecting the mesh number. Hole spacing is the basic parameter of the stencil structure. If the hole spacing is not compliant, the detection of the number of meshes will be meaningless. First determine whether the hole spacing is compliant. If it is not compliant, directly remove the stencil, thereby reducing unnecessary mesh detection and optimizing the use of detection resources. In addition, the convolution layer of CNN can effectively extract the edge features of the stencil holes, such as the position and arrangement pattern of the holes. This level of feature extraction helps to identify the exact position and distance of the holes, reduce the detection error caused by image quality or noise, and ensure the accuracy and reliability of the detection. Object detection algorithms such as YOLO can quickly identify the position of each hole, which is suitable for real-time detection needs on the production line. Under the premise of compliance with the hole spacing, the image is divided into multiple small areas and the number of holes in each area is counted. By summarizing and calculating the total number of holes per unit area, regional detection is achieved, which helps to reduce local errors when calculating the number of meshes, ensure the accuracy and consistency of the detection results, and avoid statistical deviations caused by uneven hole distribution. In addition, by comparing the standard values ​​in the database, it can be ensured that each detection has a clear compliance basis. The standardized comparison method can eliminate the error of subjective judgment, ensure the stability and uniformity of detection, and improve the quality control level on the production line.

[0039] In some preferred embodiments, the visual inspection method for the silk-screen printing process of the mobile phone shell further includes: setting a first conveyor belt mechanism and a first suction cup mechanism between the static station and the baking station of the substrate, the first suction cup mechanism sucks the substrate after being static at the static station to the first conveyor belt mechanism, the first conveyor belt mechanism transfers the static substrate to the baking station, and the baking device of the baking station bakes the static substrate. It should be noted that in this embodiment, the use of the first conveyor belt and the suction cup mechanism can automatically transfer the static substrate from the static station to the baking station, eliminating manual handling operations, reducing human intervention and operation time, not only improving production efficiency, but also reducing errors caused by manual operations and the risk of substrate damage.

[0040] In some preferred embodiments, the visual inspection method for the screen printing process of the mobile phone shell further includes: a second suction cup mechanism is set between the baking station and the CCD detection production line, and the second suction cup mechanism sucks the substrate output from the substrate output end of the baking device to the CCD detection production line for CCD imaging. It should be noted that the ink layer on the surface of the substrate after baking has been fixed, but it is still relatively fragile and is easily damaged by external forces or surface defects. The second suction cup mechanism can realize non-contact transportation, avoiding scratches, deformation or surface contamination of the ink layer caused by direct contact during the transmission of the substrate, thereby protecting the surface quality after baking. In addition, the second suction cup mechanism can steadily absorb and move the substrate to avoid affecting the clarity and integrity of the pattern due to shaking or displacement during the transportation process. Stable transmission puts the substrate in an ideal position during CCD detection, improves the accuracy of CCD imaging, and ensures the accuracy and reliability of subsequent detection. In addition, in this embodiment, the baking station and the CCD detection production line are automatically connected by the second suction cup mechanism, reducing the waiting time of the substrate between different stations, speeding up the overall production rhythm, and improving the production line efficiency, which is particularly suitable for mass production needs.

[0041] In some preferred embodiments, the first conveyor belt mechanism includes a negative pressure suction component and a conveyor belt with holes; the negative pressure suction component is arranged below the conveyor belt with holes, and is used to limit the substrate after being stationary during transmission. It should be noted that in this embodiment, the substrate on the conveyor belt with holes is adsorbed and limited by the negative pressure suction component, which can prevent the substrate from shifting due to vibration, sliding or air flow during transportation. In addition, the use of negative pressure adsorption can avoid squeezing, scratching or contamination of the substrate surface or ink layer caused by physical clamping and other methods. The cooperation between the conveyor belt with holes and the negative pressure suction component can effectively control the stability of the substrate without affecting the transmission speed.

[0042] In some further preferred embodiments, the ink printing layer of the substrate is identified by an ink printing layer recognition algorithm, and whether the characters of the ink printing layer of the substrate have jagged edges, fat and leakage spots is determined, including: using an image segmentation algorithm (such as UNet) to extract the edge and contour features of the printed characters of the ink printing layer of the substrate, and generating a fine edge segmentation mask of the characters; using a target detection algorithm (such as YOLO) to detect the smoothness or irregularity of the character edges of the printed characters, and determine whether the printed characters have jagged edges; if the character edges of the printed characters are not smooth and have irregular sawtooth and corrugations, then the printed characters are determined to have jagged edges; using a target detection algorithm (such as YOLO) to detect whether the contour of the printed characters is irregular or deformed, and if the contour of the printed characters is irregular or deformed, then the characters of the ink printing layer are determined to have fat; using an image segmentation algorithm (such as UNet) to detect defective areas inside the printed characters, and if there are defective areas inside the printed characters, then the printed characters are determined to have leakage spots. It should be noted that in this embodiment, by combining image segmentation and target detection algorithms, the edges, contours and internal defective areas of characters can be identified. Image segmentation algorithms (such as UNet) are good at detail recognition and can generate fine edge segmentation masks for characters; target detection algorithms (such as YOLO) have the ability to quickly identify edge smoothness and contour deformation. The combination of multiple algorithms ensures the detailed control of defects in the detection process and improves the accuracy and comprehensiveness of recognition. It can be understood that in this embodiment, UNet is used to generate fine edge masks for characters, and the edge morphology of characters can be accurately extracted through pixel-level segmentation, while YOLO can quickly detect edge smoothness and irregularity. The jagged or wavy defects of characters will be prominent in edge smoothness detection. Combined with the accuracy of the segmentation mask, it is easier to identify tooth edge defects, thereby significantly improving the accuracy of tooth edge detection and avoiding missed detection. In addition, the fine edge mask generated by UNet can accurately extract the contour area of ​​the character, while YOLO quickly determines the regularity of the contour and whether there is overflow. Through YOLO's deformation recognition function, it is possible to quickly determine the area where the ink is excessively accumulated, thereby determining whether there is a fat oil defect. In addition, using segmentation algorithms such as UNet to detect the interior of the character can generate a complete structural area of ​​the character, thereby identifying whether there is a defective area inside the character. Light leakage point detection mainly relies on the integrity judgment of the internal area of ​​the character. UNet's pixel-level segmentation can accurately capture small gaps or spaces within the character, which helps to accurately identify light leakage points and improve the precision of defect detection.

[0043] In some further preferred embodiments, when judging whether the characters of the ink printing layer of the substrate have tooth edges, fat and leakage points, first judge whether the characters of the ink printing layer of the substrate have tooth edges, and after the characters of the ink printing layer of the substrate have no tooth edges, judge whether the characters of the ink printing layer of the substrate have leakage points, and after the characters of the ink printing layer of the substrate have no leakage points, judge whether the characters of the ink printing layer of the substrate have fat and oil. It should be noted that the tooth edge detection is mainly to identify the smoothness and consistency of the edge of the character, and to judge whether there are defects such as jagged burrs or ripples. The detection processing speed is fast, and only the edge of the character needs to be analyzed, and no complex analysis of the interior or overall contour of the character is involved. If the tooth edge detection of the substrate fails, the substrate can be immediately judged as unqualified, and there is no need to continue to detect other defects, saving detection resources and improving detection efficiency. On the premise of determining that the edge of the character is qualified, entering the internal area of ​​the character to detect the leakage points can further eliminate the problem of internal defects. Leakage points refer to small gaps or uncovered areas that appear inside the character. This defect is usually caused by uneven ink distribution or pore blockage. If there is a light leakage point defect, the substrate can be directly judged as unqualified, avoiding unnecessary subsequent inspections and improving efficiency and accuracy. Fat and oil detection mainly focuses on the outline integrity of the characters and ink overflow, which is usually caused by excessive ink or improper printing pressure. This defect detection requires a more comprehensive inspection of the overall outline of the characters, which is relatively time-consuming, so it is placed in the last step to ensure that the edges and internal integrity of the characters are qualified before performing detailed analysis of the outline. Only after confirming that there are no major defects on the edges and inside of the characters is fat and oil detection performed, thus avoiding repeated inspections and saving resources.

[0044] It should be pointed out that the above embodiments are only preferred specific implementation modes of the present invention, and the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A visual inspection method for mobile phone housing silk screen printing process, characterized in that: include: Before assembling the screen printing screen to the printing machine, the screen printing screen is photographed to obtain a screen image, wherein the screen image reflects the hole pitch and mesh number of the screen printing screen; Identify the hole pitch and mesh of the screen printing screen by using a screen image recognition algorithm, determine whether the hole pitch and mesh of the screen printing screen are compliant, and when the hole pitch and mesh of the screen printing screen are compliant, prompt to assemble the compliant screen printing screen to the printing machine to print the base color layer of the substrate; After the base color layer of the substrate is printed with ink, the substrate after the base color layer is left to stand, and the substrate after standing to stand is baked, and the substrate after standing to stand and bake is transported to a CCD detection production line for CCD imaging to obtain an image of the ink printing layer of the substrate; Identify the ink printing layer of the substrate by using an ink printing layer identification algorithm, determine whether the characters of the ink printing layer of the substrate have tooth edges, fat oil and light leakage points, and determine whether the brightness of the ink of the ink printing layer of the substrate is in compliance with regulations; The hole pitch and mesh of the silk screen printing screen are identified by a screen image recognition algorithm, and whether the hole pitch and mesh of the silk screen printing screen are compliant is determined, including: extracting hole edge features in the screen image by the convolution layer of CNN in the CNN detection model, the hole edge features including the position and arrangement pattern of the holes; using the target detection algorithm to identify the position of each hole, and calculating the hole pitch of adjacent holes; comparing the calculated hole pitch with the standard hole pitch in the compliance database, if all the identified hole pitches meet the standard, the hole pitch is determined to be compliant; otherwise, the hole pitch is determined to be non-compliant; when the hole pitch is compliant, the screen image is divided into multiple small areas, and the CNN detection model is used to count the number of holes in each area; the number of holes in each area is summarized, and the total number of holes per unit area is calculated to determine the mesh number; comparing the identified mesh number with the standard mesh in the compliance database, if the mesh number meets the standard, the mesh is determined to be compliant, otherwise, it is determined to be non-compliant.

2. The visual inspection method for mobile phone housing silk screen printing process according to claim 1, characterized in that: When the screen image recognition algorithm recognizes that the hole pitch and mesh of the screen printing screen are not in compliance, it is prompted that the screen printing screen has defects, and the display terminal is controlled to display the types of defects in the screen printing screen.

3. The visual inspection method for mobile phone housing silk screen printing process as claimed in claim 2, characterized in that: Before controlling the display terminal to display the defect types existing in the silk screen screen, an intelligent judgment is made on the causes of different types of defects of the silk screen screen. When controlling the display terminal to display the defect types existing in the silk screen screen, the display terminal is simultaneously controlled to display the causes of the defects of the silk screen screen.

4. The visual inspection method for mobile phone housing silk screen printing process according to claim 1, characterized in that: include: A first conveyor belt mechanism and a first suction cup mechanism are arranged between a static station and a baking station of the substrate. The first suction cup mechanism sucks the substrate after being statically placed at the static station to the first conveyor belt mechanism. The first conveyor belt mechanism transfers the statically placed substrate to the baking station. The baking device at the baking station bakes the statically placed substrate.

5. The visual inspection method for mobile phone housing silk screen printing process as claimed in claim 4, characterized in that: include: A second suction cup mechanism is arranged between the baking station and the CCD detection production line, and the second suction cup mechanism sucks the substrate output from the substrate output end of the baking device to the CCD detection production line for CCD imaging.

6. The visual inspection method for mobile phone housing silk screen printing process according to claim 5, characterized in that: The first conveyor belt mechanism includes a negative pressure suction component and a conveyor belt with holes; the negative pressure suction component is arranged below the conveyor belt with holes and is used to limit the substrate after being stationary during transportation.

7. A visual inspection system for mobile phone shell silk screen printing process, characterized in that: The visual inspection system includes a host end, and the host end runs the screen image recognition algorithm and the ink printing layer recognition algorithm in the visual inspection method for the mobile phone shell silk screen printing process as described in any one of claims 1 to 6.

Images