Method, system and mobile phone case for transferring texture inside mobile phone case based on visual recognition

Abstract

The invention relates to the technical fields of signal transmission, computer vision, image recognition and printing, and provides a method, system and mobile phone case for transferring texture inside a mobile phone case based on visual recognition. The method judges whether the positioning piece is inserted in place by sensing an insertion state signal of a positioning piece inserted into a positioning hole. When all the positioning holes have been inserted with the positioning piece in place, a prompt is given to apply transfer glue, and a texture mold is controlled to be aligned and fitted with a transfer surface of a mobile phone case substrate to perform internal texture transfer. The transferred transfer surface is photocured and an image is taken to obtain a texture transfer image. The texture offset state, the noise state, the scratch state, the stripe printing state and the crystal point state of the texture on the transfer surface are analyzed by a texture analysis algorithm, so that the mobile phone case substrate with qualified transfer surface texture appearance flows into the next process, thereby realizing intelligent detection of the insertion state of the positioning piece and the transfer surface texture appearance, and improving detection efficiency and accuracy.

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 method and system for transferring texture inside a mobile phone case based on visual recognition, and a mobile phone case. Background Art

[0002] In the transfer process of the inner texture of the mobile phone case, it is necessary to transfer the inner texture on the texture mold to the transfer surface of the mobile phone case substrate through UV glue, and then inspect the texture appearance on the transfer surface to determine whether there is texture offset, spots, scratches, stripes and crystal points. In order to realize the transfer of the inner texture on the texture mold, it is necessary to first align and fix the mobile phone case substrate to the substrate fixture, and then apply UV glue to the transfer surface of the mobile phone case substrate through the glue coating device, and then install the texture mold to the printing press, and control the texture mold on the printing press to transfer the inner texture to the mobile phone case substrate. In the prior art, when the mobile phone case substrate is aligned and fixed to the substrate fixture, it is necessary to insert the positioning plate into the positioning hole of the substrate and fix the substrate on the substrate fixture. After the positioning plate is inserted, it is necessary to manually visually detect the insertion status of the positioning plate, for example, whether there is a missing hole and whether the insertion depth is sufficient. This process is not only inefficient, but also prone to detection deviation. In addition, manual visual inspection of texture offset, spots, scratches, streaks and crystal points on the transfer surface usually requires multiple quality inspectors to observe and discuss together, which is not only extremely inefficient but also highly subjective. There is a greater risk of fluctuation in the texture appearance quality on the transfer surface of the mobile phone case substrate.

[0003] To sum up, in the prior art, manual visual inspection of the insertion status of the positioning sheet has technical problems such as low efficiency and easy detection deviation. Manual visual inspection of texture offset, spots, scratches, stripe prints and crystal points on the transfer surface has extremely low efficiency and high subjectivity. There is also a high risk of fluctuation in the texture appearance quality on the transfer surface of the mobile phone case substrate. Summary of the invention

[0004] In view of the shortcomings of the above-mentioned prior art, the present invention provides a method and system for transferring texture inside a mobile phone case based on visual recognition, thereby realizing intelligent detection of the insertion status of the positioning sheet, improving detection efficiency and accuracy, and realizing intelligent detection of texture offset, noise, scratches, stripe prints and crystal points on the transfer surface, thereby improving efficiency and improving the consistency of texture appearance quality detection results.

[0005] In a first aspect, the present invention provides a method for transferring texture inside a mobile phone case based on visual recognition, comprising:

[0006] An in-place sensor is arranged at a positioning hole of the substrate fixture, and an insertion state signal of the positioning piece inserted into the positioning hole is sensed by the in-place sensor;

[0007] According to the insertion state signal, it is determined whether all the positioning holes of the substrate fixture have been inserted into the positioning pieces, and when all the positioning holes have been inserted into the positioning pieces, a prompt is given to apply transfer glue on the transfer surface of the mobile phone case substrate;

[0008] Controlling the texture mold of the printing machine to be aligned with the transfer surface of the mobile phone shell substrate after the transfer glue is applied, so as to transfer the preset inner texture on the texture mold to the transfer surface of the mobile phone shell substrate through the transfer glue;

[0009] Photocuring the transfer surface of the mobile phone case substrate after the inner texture is transferred, and photographing the transfer surface of the mobile phone case substrate after the photocuring to obtain a texture transfer image;

[0010] The texture deviation state, noise state, scratch state, stripe print state and crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image are analyzed by a texture analysis algorithm. When there is no texture deviation, noise, scratches, stripe print and crystal point on the transfer surface, it is determined that the transfer surface texture appearance of the mobile phone case substrate is qualified, and the mobile phone case substrate with qualified transfer surface texture appearance is controlled to flow into the next process.

[0011] In a second aspect, the present invention provides a system for transferring texture inside a mobile phone case based on visual recognition, wherein the visual inspection system uses the above-mentioned method for transferring texture inside a mobile phone case based on visual recognition.

[0012] In a third aspect, the present invention provides a mobile phone case, which uses the above-mentioned mobile phone case internal texture transfer method based on visual recognition to perform internal texture transfer in the transfer process.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] The present invention provides a method, system and mobile phone case for transferring the texture inside a mobile phone case based on visual recognition. The method comprises the following steps: a positioning sensor is arranged at a positioning hole of a substrate fixture, and an insertion state signal of a positioning piece inserted into the positioning hole is sensed by the positioning sensor. According to the insertion state signal, it is judged whether all the positioning holes of the substrate fixture have been inserted into the positioning piece. When all the positioning holes have been inserted into the positioning piece, a prompt is given to apply transfer glue on the transfer surface of the mobile phone case substrate, and a texture mold of a printing machine is controlled to be aligned and fitted with the transfer surface of the mobile phone case substrate after the transfer glue is applied, so that the preset inner texture on the texture mold is transferred to the transfer surface of the mobile phone case substrate through the transfer glue, and the transfer of the mobile phone case substrate after the inner texture is transferred is controlled. The surface is photocured, and the transfer surface of the mobile phone case substrate after photocuring is imaged to obtain a texture transfer image, and the texture offset state, the noise state, the scratch state, the stripe print state and the crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image are analyzed by a texture analysis algorithm. When there is no texture offset, noise, scratch, stripe print and crystal point on the transfer surface, it is determined that the transfer surface texture appearance of the mobile phone case substrate is qualified, and the mobile phone case substrate with qualified transfer surface texture appearance is controlled to flow into the next process, thereby realizing intelligent detection of the insertion state of the positioning sheet, improving detection efficiency and accuracy, realizing intelligent detection of texture offset, noise, scratch, stripe print and crystal point on the transfer surface, improving efficiency, and improving the consistency of texture appearance quality detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] 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.

[0016] Figure 1 It is a flow chart of a method for transferring texture inside a mobile phone case based on visual recognition according to an embodiment of the present invention;

[0017] Figure 2 This is a schematic diagram of the architecture of a texture transfer system in a mobile phone case based on visual recognition according to an embodiment of the present invention;

[0018] Figure 3 It is a schematic diagram of a state in which a textured mold is attached to a mobile phone case substrate in an embodiment of the present invention.

[0019] Description of reference numerals:

[0020] 1. Mobile phone case base material;

[0021] 2. Texture mold. DETAILED DESCRIPTION

[0022] 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.

[0023] See also Figure 1-Figure 3 The embodiment of the present invention provides a method for transferring texture inside a mobile phone shell based on visual recognition, a system and a mobile phone shell. The system for transferring texture inside a mobile phone shell based on visual recognition uses the above-mentioned method for transferring texture inside a mobile phone shell based on visual recognition. The mobile phone shell uses the above-mentioned method for transferring texture inside a mobile phone shell based on visual recognition to transfer the internal texture in the transfer process. The method for transferring texture inside a mobile phone shell based on visual recognition includes the following steps:

[0024] S101, setting an in-place sensor at a positioning hole of a substrate fixture, and sensing an insertion state signal of a positioning piece inserted into the positioning hole by the in-place sensor;

[0025] S102, judging whether all the positioning holes of the substrate fixture have been inserted into the positioning pieces according to the insertion status signal, and when all the positioning holes have been inserted into the positioning pieces, prompting to apply transfer glue on the transfer surface of the mobile phone case substrate;

[0026] S103, controlling the texture mold of the printing machine to align with the transfer surface of the mobile phone case substrate after the transfer adhesive is applied, so as to transfer the preset inner texture on the texture mold to the transfer surface of the mobile phone case substrate through the transfer adhesive;

[0027] S104, photocuring the transfer surface of the mobile phone shell substrate after the internal texture transfer, and photographing the transfer surface of the mobile phone shell substrate after the photocuring to obtain a texture transfer image;

[0028] S105. Analyze the texture shift state, noise state, scratch state, stripe print state and crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image through a texture analysis algorithm. When there is no texture shift, noise, scratch, stripe print and crystal point on the transfer surface, determine that the transfer surface texture appearance of the mobile phone case substrate is qualified, and control the mobile phone case substrate with qualified transfer surface texture appearance to flow into the next process.

[0029] It should be noted that, in the present embodiment, an in-place sensor is provided at the positioning hole of the substrate fixture, and the in-place sensor senses an insertion status signal of the positioning piece inserted into the positioning hole. According to the insertion status signal, it is determined whether all the positioning holes of the substrate fixture have been inserted with the positioning piece. When all the positioning holes have been inserted with the positioning piece, a prompt is given to apply transfer glue on the transfer surface of the mobile phone case substrate, and the texture mold of the printing machine is controlled to be aligned and fitted with the transfer surface of the mobile phone case substrate after the transfer glue is applied, so that the preset inner texture on the texture mold is transferred to the transfer surface of the mobile phone case substrate through the transfer glue, and the transfer surface of the mobile phone case substrate after the inner texture is transferred is photocured, and the photocuring is performed. The transfer surface of the solidified mobile phone case substrate is imaged to obtain a texture transfer image, and the texture offset state, noise state, scratch state, stripe print state and crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image are analyzed by a texture analysis algorithm. When there is no texture offset, noise, scratches, stripe prints and crystal points on the transfer surface, the transfer surface texture appearance of the mobile phone case substrate is determined to be qualified, and the mobile phone case substrate with qualified transfer surface texture appearance is controlled to flow into the next process, thereby realizing intelligent detection of the insertion state of the positioning sheet, improving detection efficiency and accuracy, realizing intelligent detection of texture offset, noise, scratches, stripe prints and crystal points on the transfer surface, improving efficiency, and improving the consistency of texture appearance quality detection results.

[0030] It should be noted that, in the present embodiment, the insertion status signal sensed by the in-place sensor is transmitted to the host end, and the host end determines whether all the positioning holes of the substrate fixture tooling have been inserted into the positioning pieces according to the insertion status signal. When all the positioning holes have been inserted into the positioning pieces, it is prompted to apply transfer glue on the transfer surface of the mobile phone case substrate, and the texture mold of the printing machine is controlled to align and fit with the transfer surface of the mobile phone case substrate after the transfer glue is applied, so that the preset inner texture on the texture mold can be transferred to the transfer surface of the mobile phone case substrate through the transfer glue. In addition, image capture can be achieved through an image capture system. The image capture system captures the transfer surface of the mobile phone case substrate after light curing. After obtaining the texture transfer image, the texture transfer image is transmitted to the host end. The host end analyzes the texture offset state, noise state, scratch state, stripe print state and crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image through a texture analysis algorithm. When there is no texture offset, noise, scratches, stripe prints and crystal points on the transfer surface, it is determined that the texture appearance of the transfer surface of the mobile phone case substrate is qualified, and the mobile phone case substrate with qualified transfer surface texture appearance is controlled to flow into the next process.

[0031] In step S101, by setting a position sensor on the positioning hole of the substrate fixture, the insertion status of the positioning piece can be automatically sensed. Compared with manual visual inspection, this automated detection method can improve the intelligence of detection, eliminate the risk of manual misjudgment and deviation, and ensure that all positioning holes are accurately inserted with positioning pieces. In addition, automatic detection improves the efficiency of positioning detection, avoids the time waste of manual inspection, ensures that the mobile phone case substrate can enter the subsequent steps more quickly, and reduces the overall production time.

[0032] In step S102, it is determined whether all the positioning holes are inserted into place based on the insertion status signal. Only after it is confirmed that all the positioning pieces are in place will the transfer glue application be prompted to avoid deviations in subsequent processes due to improper insertion of the positioning pieces, thereby ensuring the accuracy and consistency of the transfer process and eliminating the need for manual confirmation of the insertion status of the positioning pieces, greatly improving efficiency and effectively avoiding subsequent transfer errors caused by missed insertion or incomplete insertion.

[0033] In step S103, the alignment of the texture mold and the substrate coated with UV glue is controlled to ensure that the inner texture of the mold can be accurately transferred to the mobile phone case substrate, reducing the risk of texture deviation. It can be understood that alignment can ensure the accuracy and consistency of the texture in the transfer process, thereby improving the transfer quality and reducing the occurrence of deviation, blur and other problems.

[0034] In step S104, after the transfer is completed, the mobile phone case substrate is light-cured to fix the transferred texture, ensure the durability and stability of the texture, and avoid texture damage caused by subsequent operations. Among them, image capture enables texture detection to be digitally analyzed through image data, providing basic data for subsequent automatic detection, avoiding the uncertainty and subjectivity of manual observation of texture appearance.

[0035] In step S105, a texture analysis algorithm is used to automatically detect the five major defects of the texture (texture offset, spots, scratches, streaks and crystal points), replacing manual visual inspection and automatically identifying defect characteristics, so as to standardize the inspection process. This not only greatly improves the inspection efficiency and reduces the inspection cost, but also avoids the inconsistency and errors caused by the subjective judgment of different inspectors. When the texture appearance is judged to be qualified, the substrate will enter the next process, thereby reducing the flow of unqualified products into subsequent production links, improving the product qualification rate and reducing the risk of rework.

[0036] In some preferred embodiments, the texture transfer method in a mobile phone case based on visual recognition further includes: a light indicator device is provided next to each positioning hole on the substrate fixture, and the light indicator device lights up to indicate when the corresponding positioning hole is inserted into the positioning piece. It should be noted that the light indicator device next to each positioning hole can light up when the positioning piece is inserted into place, providing direct visual feedback. The operator does not need to check the insertion status of the positioning piece one by one, and can see the light status of each hole at a glance, which greatly improves the detection efficiency and helps to quickly identify unqualified positioning holes. If the indicator light of a certain hole is not on, the operator can immediately realize that the positioning piece at that position is not inserted into place, so as to quickly take corrective measures to help the operator promptly discover and deal with abnormal situations, and avoid subsequent errors or process problems caused by improper insertion of positioning pieces.

[0037] In some preferred embodiments, when there is a positioning hole where the positioning piece is not inserted into place, a prompt is given to check the insertion status of the positioning piece into the positioning hole. When the inspector checks the insertion status of the positioning piece into the positioning hole, the inspection is performed through the lighting status of the light indicator device. It should be noted that when it is detected that there is a positioning piece that has not been inserted into the positioning hole, a prompt is immediately issued to remind the inspector to check, to ensure that the abnormal situation of the insertion status of the positioning piece is discovered and handled in time, to avoid problems in the subsequent process due to inaccurate positioning. Through the lighting status of the light indicator device, the inspector can quickly and intuitively confirm the insertion status of the positioning piece. The light is on, which means that it is inserted into place, and the light is not on, which means that it is not inserted into place. The inspector does not need to measure or carefully check the positioning holes one by one, but only needs to check the light status to make a judgment, which can effectively reduce the risk of misjudgment or missed detection in manual inspection, optimize the inspection process, and make positioning detection and correction more efficient, thereby improving the overall efficiency of the production line.

[0038] In some further preferred embodiments, when judging whether all the positioning holes of the substrate fixture have been inserted into the positioning piece according to the insertion status signal, the method includes: assuming that there are N positioning holes on the fixture, the insertion status of each positioning hole is represented by a sensor signal, and the signal value 1 indicates that the insertion is in place, and 0 indicates that the insertion is not in place; assuming that the insertion status signal of the i-th positioning hole among the N positioning holes is , Indicates that the insertion is in place. Indicates that it is not in place; let the overall insertion status judgment result be , obtain the insertion status signal of each positioning hole from the sensor to form a status signal vector ; Use the calculation formula to determine whether all holes have been inserted into place, the calculation formula is:

[0039] ; 1 means all positioning pieces are in place. 0 means at least one positioning piece is not in place. It should be noted that in this embodiment, by representing the sensor signal of each positioning hole as a binary state (1 means in place, 0 means not in place), the states of all positioning holes can be uniformly represented as a state signal vector , by calculating the product, it is possible to quickly determine whether all hole positions are in place without having to inspect them one by one, thus simplifying the overall inspection process.

[0040] In some preferred embodiments, when controlling the texture mold of the printing machine to align with the transfer surface of the mobile phone shell substrate after the transfer glue is applied, it includes: presetting the alignment reference points on the mobile phone shell substrate and the texture mold; capturing the transfer surface of the mobile phone shell substrate and the reference point images of the texture mold; identifying the coordinate position of the reference point in the reference point image by an image processing algorithm, and calculating the relative position difference between the mobile phone shell substrate and the texture mold; performing alignment adjustment according to the relative position difference to control the texture mold of the printing machine to align with the transfer surface of the mobile phone shell substrate after the transfer glue is applied. It should be noted that in this embodiment, the setting of the reference point provides an accurate alignment reference, and the relative position of the mobile phone shell substrate and the texture mold can be obtained by capturing the reference point image. Accurate alignment reference can ensure the position accuracy of the texture during the transfer process, and avoid problems such as transfer misalignment, blurred or uneven texture caused by position deviation. The image processing algorithm can quickly and automatically identify the coordinate position of the reference point, and calculate the relative position difference between the mobile phone shell substrate and the texture mold. The automated image recognition and position calculation process greatly improves the efficiency of alignment, which is suitable for the automation requirements of large-scale production lines.

[0041] In some preferred embodiments, the coordinate position of the reference point in the reference point image is identified by an image processing algorithm, and the relative position difference between the mobile phone case substrate and the texture mold is calculated, including: defining the reference point image on the mobile phone case substrate as the substrate image Image base, and the reference point image on the texture mold as the mold image Image template; using the ORB algorithm to detect feature points on the substrate image Image base and the mold image Image template respectively, and using the BRIEF descriptor of the ORB algorithm to calculate the feature vector of each feature point, so as to perform feature matching between the substrate and the mold image; using the BRIEF descriptor, matching the feature points of the substrate image Image base and the mold image Image template by a brute force matching method, finding similar feature point pairs, and using a random sampling consistency algorithm to remove erroneous matching points to obtain a set of accurate matching feature point pairs; from the accurate matching feature point pairs, selecting a preset reference point pair (such as a predetermined corner point or center point in the substrate image Image base and the mold image Image template) to obtain the coordinates of the reference points on the substrate and the mold, and the coordinates of the reference points on the substrate are recorded as , the coordinates of the reference point on the mold are Based on the coordinates of the reference point, the relative position difference between the substrate and the mold is calculated, and the relative position difference includes the lateral (ie, X-axis) displacement difference , longitudinal (ie Y-axis) displacement difference and the rotation angle difference ; Rotation angle difference Estimated by affine transformation between matching points; ; . It should be noted that the ORB algorithm is a lightweight algorithm with fast algorithm speed and low computational cost, and is suitable for application in real-time industrial detection. In addition, the ORB algorithm has good robustness to image scale and rotation, and can accurately identify feature points, ensuring that alignment can be performed even when there is a slight position deviation between the substrate and the mold. In this embodiment, the ORB algorithm is used to perform feature point detection, descriptor calculation, and feature matching, so as to accurately identify the coordinates of the reference points in the substrate and mold images, and calculate the relative position difference, providing fast and reliable alignment adjustment to ensure high-precision alignment of the substrate and the mold.

[0042] In some further preferred embodiments, the rotation angle difference When the estimation is performed by affine transformation between matching points, it includes: defining a feature point coordinate set, the feature point coordinate set including a feature point coordinate set in the substrate image in the matching feature point pair , the corresponding feature point coordinate set in the mold image ;in, , is the number of matched feature point pairs; an affine transformation model is established, and the affine transformation model is:

[0043] ;

[0044] in, Represents the parameters of rotation and scaling, Represents the parameters of translation, rotation and scaling And the parameters of translation Denoted as affine transformation parameters;

[0045] Substitute the feature point coordinate set into the affine transformation model to construct a linear equation system:

[0046] ;

[0047] Solving the affine transformation parameters using the least square method to minimize the error of the linear equations for the matching point pairs;

[0048] Calculate the rotation angle difference according to the obtained affine transformation parameters ;

[0049] .

[0050] It should be noted that, in this embodiment, by calculating the rotation angle difference, the rotation angle of the mold or substrate can be accurately adjusted to ensure that the angle error in the alignment process is compensated. The affine transformation model can handle multiple transformations such as translation, rotation, and scaling, and is suitable for complex deviations that may occur in actual production. The calculation of affine transformation parameters can be used to achieve real-time and accurate automatic alignment.

[0051] In some preferred embodiments, when the transfer surface of the mobile phone shell substrate after the inner texture is transferred is photocured, it includes: firstly performing LED photocuring on the transfer surface of the mobile phone shell substrate after the inner texture is transferred, and when performing LED photocuring, the LED energy of the energy meter is 1000±100MJ; after LED photocuring, the transfer surface after LED photocuring is mercury lamp cured, and when performing mercury lamp curing, the mercury lamp energy of the energy meter is selected between 1000MJ-1200MJ. It should be noted that in this embodiment, a dual curing process of first LED photocuring and then mercury lamp curing is adopted, the purpose of which is to cure the transfer texture at different depths in stages. LED photocuring can quickly cure the surface layer of UV glue, while mercury lamp curing can further cure the internal structure of UV glue to ensure the depth and firmness of curing. Among them, LED photocuring is characterized by low heat and rapid reaction, which can cure the surface in a short time without generating too much heat, thereby reducing the thermal impact on the substrate, especially for heat-sensitive mobile phone shell materials, which can avoid deformation or material aging caused by excessive temperature. During the LED light curing process, the surface of the UV glue layer is quickly cured to form a preliminary protective layer, which prevents dirt or dust from invading the texture surface and helps maintain the clarity and integrity of the transferred texture. Mercury lamp curing can produce a wide spectrum of ultraviolet light, including deep ultraviolet bands, with a high energy density, which can penetrate the surface of the UV glue and further cure the internal structure of the glue layer. The strong energy of the mercury lamp can ensure the deep curing of the UV glue, so that the texture is firmly attached to the substrate and improves durability. Mercury lamp curing can deeply cure the UV glue layer, ensure the uniformity of curing, increase the adhesion and wear resistance of the texture, and is suitable as a means of further strengthening after LED initial curing. In addition, the LED energy is selected within the range of 1000±100 mJ, which can quickly complete the surface curing in a short time, while controlling the energy to avoid excessive thermal effects. Within this energy range, LED curing can effectively complete the curing of the surface of the UV glue, forming a preliminary protective layer, while avoiding damage or deformation of the substrate caused by high temperature. After LED light curing, the energy of the mercury lamp is set between 1000-1200 mJ to ensure that the internal structure of the UV glue is deeply cured. The high energy density of the mercury lamp can penetrate the surface layer and further stimulate the deep area of ​​the UV glue to ensure uniformity and durability of curing. The range of 1000-1200 mJ is sufficient to complete deep curing without damaging the substrate.

[0052] In some preferred embodiments, when the texture shift state, the noise state, the scratch state, the stripe print state and the crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image are analyzed by a texture analysis algorithm, it includes: when detecting the texture shift state of the texture, using Mask R-CNN to perform image segmentation, calculate the coordinates of the center point of the transfer texture, and use the texture center point of the preset standard template as a comparison standard; by image alignment and coordinate offset calculation, detect the distance difference between the center point of the transfer texture and the center point of the standard template; when the center offset distance exceeds the preset tolerance (for example, ±0.5 mm), it is determined that there is texture shift; when detecting the noise state of the texture, set the area and brightness standards of the noise, such as the area is less than a certain threshold (for example, 5 pixels), and the brightness is significantly higher or lower than the brightness of the surrounding texture; when the size, shape and brightness of the noise are different from the preset standard texture, it is determined that there are noise in the texture; when detecting the scratch state of the texture, using Mask R-CNN performs image segmentation to detect linear or strip-shaped areas; uses edge detection algorithms (such as Sobel or Canny edge detection) combined with shape analysis to determine whether there are linear scratch areas in the texture transfer image whose length exceeds a set threshold; when the length and width of the scratch exceed the preset standard (such as the length is greater than 2 mm and the width is between 0.1-0.5 mm), the texture is judged to be scratched; when detecting the stripe print state of the texture, Mask R-CNN is used for segmentation detection, and the periodic stripe structure in the image is analyzed in combination with Fourier transform to identify the stripe structure that appears repeatedly in the texture transfer image, and determine its direction and spacing; when the repetition frequency and spacing of the stripes reach a preset range (such as the stripe width is between 1-3 mm and the spacing is less than 5 mm), the texture is judged to have stripe prints; using Mask R-CNN to detect the stripe print state of the texture, and combine the periodic stripe structure in the image with the Fourier transform to identify the stripe structure that appears repeatedly in the texture transfer image, and determine its direction and spacing; when the repetition frequency and spacing of the stripes reach a preset range (such as the stripe width is between 1-3 mm and the spacing is less than 5 mm), the texture is judged to have stripe prints; R-CNN performs image segmentation and detects small areas of high brightness; combined with brightness threshold and regional shape analysis, the highlight points of the small areas of high brightness in the texture transfer image are identified. The highlight points are reflected as higher brightness values ​​than the surrounding areas, and when the area is within a preset range (such as an area less than 10 pixels and a brightness higher than the surrounding areas), it is determined that there are crystal points in the texture. It should be noted that in this embodiment, by comparing the coordinates of the center point, the offset of the texture position can be accurately identified, which helps to ensure the accurate positioning of the texture on the substrate. Miscellaneous point detection ensures that there are no foreign matter or impurities on the surface of the product texture, and improves the visual cleanliness. In this embodiment, by setting the area and brightness thresholds, small miscellaneous points that affect the visual effect can be accurately identified, and unnecessary detection errors can be eliminated. In addition, scratch detection ensures that the surface is smooth and intact, avoiding scratches that affect the feel and appearance. In this embodiment, the length and width standards can exclude minor scratches that do not affect the appearance, effectively focus on more serious scratch defects, accurately identify linear defects through edge detection, improve the accuracy of scratch detection, and ensure surface quality.In addition, stripe print detection ensures that there is no periodic texture interference on the product surface, preventing the product from having uneven stripes visually. In this embodiment, the frequency characteristics of the stripes are analyzed by Fourier transform, which can quickly identify periodic stripe prints and improve the sensitivity and reliability of detection. In addition, crystal point detection ensures uniform surface brightness and avoids sudden changes in brightness that affect the overall appearance. In this embodiment, by setting the brightness threshold, it is possible to identify bright spots that affect visual consistency, effectively preventing overly bright crystal points from affecting visual quality.

[0053] In some further preferred embodiments, when analyzing the texture offset state, the state of the spots, the state of the scratches, the state of the stripes and the state of the crystal points of the texture on the transfer surface, the texture offset state is first analyzed, and when there is no texture offset on the texture on the transfer surface, the state of the spots of the texture is analyzed, when there are no spots on the texture on the transfer surface, the state of the scratches of the texture is analyzed, when there are no scratches on the texture on the transfer surface, the state of the stripes of the texture is analyzed, and when there are no stripes on the texture on the transfer surface, the state of the crystal points of the texture is analyzed. It should be noted that the texture offset analysis is a macroscopic detection at the position level, with low computational complexity and fast detection speed. First, check whether the texture position is correct, which helps to quickly confirm the overall alignment of the pattern and avoid the offset of the subsequent detection benchmark. If the texture position is significantly offset, subsequent fine detection (such as spots, scratches, etc.) will no longer be necessary. Therefore, first detecting the texture offset can reduce the need for other resource-intensive analysis of images with obvious offsets, thereby saving computing resources. Once the texture offset is confirmed to be qualified, the benchmark points for subsequent detection are also more accurate, which can avoid misjudgment of other defects caused by offset. Noise analysis focuses on small-area anomalies and is a refined processing of image quality. The inspection of the noise state ensures the surface cleanliness on the basis of accurate alignment and improves the accuracy of subsequent inspections. The computational complexity of the noise detection algorithm is relatively low. Detecting impurities in the low-complexity detection stage can avoid more complex resource consumption. Noise detection lays the foundation for subsequent large-area defect detection such as scratches and streaks. By removing the noise first, unnecessary detection interference can be reduced to ensure a more accurate and efficient overall inspection process. Scratches are usually large-area or long linear defects with significant visual impact. Scratches analysis is performed on the basis of alignment and surface cleanliness to reduce the possibility of misjudgment. If linear scratches are discovered in subsequent steps, repeated analysis will be required to consume resources, while early inspection can reduce repeated calculations. Scratches analysis is an algorithm based on edge and shape analysis. Placing it after noise can ensure accuracy on clean surfaces, thereby avoiding redundant detection and waste of resources. Streak marks are a periodic defect. When Fourier transform is used to detect periodic stripe structures in images, more efficient spectrum analysis can be achieved if there is no interference from offsets, noise or scratches. Streak detection is relatively complex and has high computational costs. Streak analysis requires clear and continuous texture images. When noise and scratches have been eliminated, the accuracy of spectrum analysis can be avoided from being affected by superimposed defects, thereby improving the accuracy of overall detection. Crystal points usually appear as small bright spots with abnormal brightness and are relatively minor defects. Crystal point analysis is a resource-intensive analysis. Performing crystal point analysis after other large-area defects have been eliminated can ensure that resources are focused on identifying significant brightness unevenness that has little impact on the final appearance quality, maintaining brightness uniformity, and ensuring the visual quality of the product.

[0054] 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 method for transferring texture inside a mobile phone case based on visual recognition, characterized in that: include: An in-place sensor is arranged at a positioning hole of the substrate fixture, and an insertion state signal of the positioning piece inserted into the positioning hole is sensed by the in-place sensor; According to the insertion state signal, it is determined whether all the positioning holes of the substrate fixture have been inserted into the positioning pieces, and when all the positioning holes have been inserted into the positioning pieces, a prompt is given to apply transfer glue on the transfer surface of the mobile phone case substrate; Controlling the texture mold of the printing machine to be aligned with the transfer surface of the mobile phone shell substrate after the transfer glue is applied, so as to transfer the preset inner texture on the texture mold to the transfer surface of the mobile phone shell substrate through the transfer glue; Photocuring the transfer surface of the mobile phone case substrate after the inner texture is transferred, and photographing the transfer surface of the mobile phone case substrate after the photocuring to obtain a texture transfer image; Analyzing the texture shift state, the state of the noise point, the state of the scratch, the state of the stripe print and the state of the crystal point of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image by a texture analysis algorithm, and judging that the transfer surface texture appearance of the mobile phone case substrate is qualified when there is no texture shift, noise point, scratch, stripe print and crystal point on the transfer surface, and controlling the mobile phone case substrate with qualified transfer surface texture appearance to flow into the next process; When the transfer surface of the mobile phone case substrate after the inner texture is transferred is subjected to light curing, the method includes: firstly performing LED light curing on the transfer surface of the mobile phone case substrate after the inner texture is transferred, and when performing LED light curing, the LED energy of the energy meter is 1000±100MJ; after LED light curing, performing mercury lamp curing on the transfer surface after LED light curing, and when performing mercury lamp curing, the mercury lamp energy of the energy meter is selected between 1000MJ-1200MJ; When analyzing the texture shift state, the noise state, the scratch state, the stripe print state and the crystal point state of the texture on the transfer surface of the mobile phone case substrate in the texture transfer image by a texture analysis algorithm, the method includes: when detecting the texture shift state of the texture, using Mask R-CNN to perform image segmentation, calculating the coordinates of the center point of the transfer texture, and using the texture center point of a preset standard template as a comparison standard; detecting the distance difference between the center point of the transfer texture and the center point of the standard template by image registration and coordinate offset calculation; when the distance of the center offset exceeds a preset tolerance, determining that there is texture shift; when detecting the noise state of the texture, setting the area and brightness standards of the noise; when the size, shape and brightness of the noise are different from those of the preset standard texture, determining that there are noises in the texture; when detecting the scratch state of the texture, using Mask R-CNN to perform image segmentation, calculating the coordinates of the center point of the transfer texture, and using the texture center point of the preset standard template as a comparison standard; detecting the distance difference between the center point of the transfer texture and the center point of the standard template by image registration and coordinate offset calculation; when the center offset distance exceeds a preset tolerance, determining that there is texture shift; when detecting the noise state of the texture, setting the area and brightness standards of the noise; when the size, shape and brightness of the noise are different from those of the preset standard texture, determining that there are noises in the texture; when detecting the scratch state of the texture, using Mask R-CNN to perform image segmentation, calculating the distance difference between the center point of the transfer texture and the center point of the standard template; when detecting the noise ... R-CNN performs image segmentation to detect linear or strip-shaped areas; uses an edge detection algorithm combined with shape analysis to determine whether there are linear scratch areas in the texture transfer image whose length exceeds a set threshold; when the length and width of the scratch exceed the preset standard, the texture is judged to be scratched; when detecting the stripe printing state of the texture, Mask R-CNN is used for segmentation detection, and combined with Fourier transform analysis of the periodic stripe structure in the image, the stripe structure that appears repeatedly in the texture transfer image is identified, and its direction and spacing are determined; when the repetition frequency and spacing of the stripes reach a preset range, the texture is judged to have stripe printing; Mask R-CNN is used for image segmentation to detect small areas of high brightness; combined with brightness threshold and regional shape analysis, the highlight points of the small areas of high brightness in the texture transfer image are identified, and the highlight points are reflected by the brightness value being higher than the surrounding area, and when the area is within the preset range, the texture is judged to have crystal points.

2. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 1, characterized in that: Also includes: A light indicating device is arranged beside each positioning hole on the substrate fixture tooling, and the light indicating device lights up to indicate when the corresponding positioning hole is inserted into the positioning sheet.

3. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 2, characterized in that: When the positioning piece is not inserted into the positioning hole, a prompt is given to check the insertion status of the positioning piece into the positioning hole. When the inspector checks the insertion status of the positioning piece into the positioning hole, the inspection is performed through the lighting status of the light indication device.

4. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 1, characterized in that: When judging whether all the positioning holes of the substrate fixture have been inserted into the positioning piece according to the insertion state signal, the method includes: assuming that there are N positioning holes on the fixture, the insertion state of each positioning hole is represented by a sensor signal, and the signal value 1 indicates that the insertion is in place, and 0 indicates that the insertion is not in place; assuming that the insertion state signal of the i-th positioning hole among the N positioning holes is , Indicates that the insertion is in place. Indicates that it is not in place; let the overall insertion status judgment result be , obtain the insertion status signal of each positioning hole from the sensor to form a status signal vector ; Use the calculation formula to determine whether all holes have been inserted into place, the calculation formula is: ; 1 means all positioning pieces are in place. A value of 0 indicates that at least one positioning piece is not in place.

5. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 1, characterized in that: When controlling the texture mold of the printing machine to align and fit with the transfer surface of the mobile phone case substrate after the transfer glue is applied, it includes: presetting alignment reference points on the mobile phone case substrate and the texture mold; capturing the reference point images of the transfer surface of the mobile phone case substrate and the texture mold; identifying the coordinate position of the reference point in the reference point image by an image processing algorithm, and calculating the relative position difference between the mobile phone case substrate and the texture mold; and performing alignment adjustment according to the relative position difference to control the texture mold of the printing machine to align and fit with the transfer surface of the mobile phone case substrate after the transfer glue is applied.

6. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 5, characterized in that: The method comprises the following steps: identifying the coordinate position of the reference point in the reference point image by an image processing algorithm, and calculating the relative position difference between the mobile phone case substrate and the texture mold; defining the reference point image on the mobile phone case substrate as the substrate image Image base, and defining the reference point image on the texture mold as the mold image Image template; detecting feature points on the substrate image Image base and the mold image Image template by using an ORB algorithm, respectively, and calculating the feature vector of each feature point by using a BRIEF descriptor of the ORB algorithm, so as to perform feature matching between the substrate image and the mold image; matching the feature points of the substrate image Image base and the mold image Image template by using a brute force matching method using the BRIEF descriptor, finding similar feature point pairs, and removing erroneous matching points by using a random sampling consistency algorithm to obtain a set of accurate matching feature point pairs; selecting a preset reference point pair from the accurate matching feature point pairs, obtaining the coordinates of the reference points on the substrate and the mold, and recording the coordinates of the reference points on the substrate as , the coordinates of the reference point on the mold are Based on the coordinates of the reference point, the relative position difference between the substrate and the mold is calculated, and the relative position difference includes the lateral displacement difference , longitudinal displacement difference and the rotation angle difference ; Rotation angle difference Estimated by affine transformation between matching points; ; .

7. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 6, characterized in that: Rotation angle difference When the estimation is performed by affine transformation between matching points, it includes: defining a feature point coordinate set, the feature point coordinate set including a feature point coordinate set in the substrate image in the matching feature point pair , the corresponding feature point coordinate set in the mold image ;in, , is the number of matched feature point pairs; an affine transformation model is established, and the affine transformation model is: ； in, Represents the parameters of rotation and scaling, Represents the parameters of translation, rotation and scaling And the parameters of translation Denote as affine transformation parameters; substitute the feature point coordinate set into the affine transformation model to construct a linear equation system: ； The affine transformation parameters are solved using the least squares method to minimize the error of the linear equations for the matching point pairs; the rotation angle difference is calculated based on the solved affine transformation parameters. ; 。 8. The method for transferring texture inside a mobile phone case based on visual recognition as claimed in claim 7, characterized in that: The preset reference point pair includes predetermined corner points or center points in the substrate image Image base and the mold image Image template.

9. A texture transfer system for a mobile phone case based on visual recognition, characterized in that: The system uses the method for transferring texture inside a mobile phone case based on visual recognition as described in any one of claims 1-8.

10. A mobile phone case, characterized in that: The mobile phone case uses the mobile phone case internal texture transfer method based on visual recognition as described in any one of claims 1 to 8 to perform internal texture transfer in the transfer process.

Images