CN118446984B - A defect detection system and method for enamel colored porcelain - Google Patents

Abstract

The invention discloses a defect detection system and method for enamel colored porcelain, belonging to the technical field of defect detection of enamel colored porcelain. The method comprises: establishing an embryo model, segmenting the embryo model, and obtaining each unit area; identifying and sorting each unit area to obtain an identification sequence; setting a positioning point; setting corresponding verification information for each unit area; performing image acquisition on the enamel colored porcelain to be inspected to obtain a detection image; obtaining an identification sequence, positioning and identifying the detection image according to the positioning point, and determining the initial positioning area of the correspondingly sorted unit area in the identification sequence in the detection image; obtaining a detection grayscale image of the detection image, marking the initial positioning area in the detection grayscale image, and determining unit detection data of the unit area; verifying the unit detection data according to the verification information corresponding to the unit area to obtain a detection verification result; and so on, until all detections are completed according to the identification sequence.

Description

A defect detection system and method for enamel colored porcelain

Technical Field

The invention belongs to the technical field of enamel colored porcelain defect detection, and in particular relates to an enamel colored porcelain defect detection system and method.

Background Art

As a ceramic handicraft with high artistic value, enamel porcelain has extremely high quality requirements in its production process. However, traditional enamel porcelain defect detection mainly relies on manual visual inspection, which is not only inefficient but also easily affected by human factors, resulting in inaccurate and inconsistent detection results.

With the rapid development of computer vision and machine learning technologies, more and more fields are beginning to try to apply these technologies to automated inspection. However, in the field of enamel porcelain defect detection, the existing automated inspection systems still have some problems. First, due to the complexity and diversity of the enamel porcelain surface, it is difficult to extract effective features. Second, the existing inspection systems can often only identify specific defect types and cannot effectively respond to newly emerging defect types.

Based on this, the present invention provides a defect detection system and method for enamel porcelain.

Summary of the invention

In order to solve the problems existing in the above-mentioned scheme, the present invention provides a defect detection system and method for enamel porcelain.

The purpose of the present invention can be achieved through the following technical solutions:

A method for detecting defects in enamel colored porcelain, the method comprising:

Step 1: Obtain standard enamel colored porcelain, establish an embryo model according to the standard enamel colored porcelain, segment the embryo model according to the standard enamel colored porcelain, and obtain each unit area;

Furthermore, the method for segmenting the embryo model according to the standard enamel porcelain includes:

Obtaining a standard image corresponding to the standard enamel porcelain, segmenting the standard image, and obtaining initial images and transition areas between the initial images;

Performing grayscale processing on the standard image to obtain a standard grayscale image, marking each initial image and transition area in the standard grayscale image; determining the initial image corresponding to each transition area and marking it as a reference image; and setting an initial value of each reference image;

Identify the grayscale value of each pixel in the transition area, analyze each grayscale value through a preset judgment model, and determine the reference image corresponding to each pixel;

Each pixel in the transition area is merged into the corresponding reference image to obtain each unit image; the embryo model is segmented according to the position area of the unit image in the embryo model to obtain each unit area.

Furthermore, the method of setting the initial value of each reference image includes:

Step SA1: identifying the gray value of each pixel in the reference image, merging adjacent pixels with the same gray value to obtain each first region; and determining the first region value of the first region;

Step SA2: Calculate the difference between the first region values of adjacent first regions, marked as the first difference; merge the adjacent first regions whose first difference is not greater than the threshold value X1 to obtain a new first region; determine the new first region value;

Step SA3: looping step SA2 until all first differences are greater than the threshold value X1; identifying the area of each first region;

According to the formula Calculate the corresponding initial value;

In the formula: CZ is the initial value; v represents the corresponding first region, c=1, 2, ..., v, v is a positive integer; Ac represents the corresponding first region area; DZc represents the first region value corresponding to the first region.

Furthermore, the expression of the judgment model is:

Where: s is the gray value of the corresponding pixel in the transition area.

Step 2: Identify and sort each unit area to obtain an identification sequence; set corresponding positioning points according to standard enamel porcelain; set corresponding verification information for each unit area;

Furthermore, the method for identifying and sorting each unit area includes:

Obtain unit images corresponding to each unit area, and mark each unit image equally; perform identification sequence combination according to each unit image to obtain each candidate combination;

Based on the standard image of standard enamel porcelain, each candidate combination is recognized and simulated to obtain the corresponding recognition efficiency and recognition accuracy;

Calculate the corresponding simulation value according to the formula PQ = b1 × XL + b2 × DL;

Where: PQ is the simulation value; b1 and b2 are both proportional coefficients, with a value range of 0<b1≤1, 0<b2≤1; XL is the recognition efficiency; DL is the recognition accuracy;

The recognition sequence is set according to the candidate combination with the largest simulation value.

Step 3: Capture images of the enamel porcelain to be inspected to obtain inspection images;

Step 4: Obtain the recognition sequence, perform positioning recognition on the detection image according to the preset positioning points, determine the position of the correspondingly ordered unit area in the recognition sequence in the detection image, and mark it as the initial positioning area;

Step 5: grayscale processing is performed on the detection image to obtain a detection grayscale image, an initial positioning area is marked in the detection grayscale image, image processing is performed according to the initial positioning area, and unit detection data corresponding to the unit area is determined;

Furthermore, the method of performing image processing on the initial positioning area includes:

Step SC1: Identify and locate the grayscale value of each pixel in the initial area, merge adjacent pixels with the same grayscale value to obtain a second area;

Step SC2: calculating the second difference between each adjacent second region; merging the adjacent second regions whose second difference is not greater than the threshold value X2 to obtain a new second region;

Step SC3: looping step SC2 until there is no second difference value not greater than the threshold value X2, and determining the unit initial region according to the remaining second regions;

Step SC4: Identify and locate each pixel outside the initial region and adjacent to the unit initial region, and mark them as out-of-bounds pixels; identify the grayscale value of each out-of-bounds pixel, evaluate whether each out-of-bounds pixel meets the merging requirement, merge each out-of-bounds pixel that meets the merging requirement with the unit initial region, and obtain a new unit initial region;

Step SC5: loop step SC4 until there are no out-of-bounds pixels that meet the merging requirements, mark the unit initial area as the unit control area; and obtain unit detection data according to the unit control area.

Step 6: Position and mark the unit detection data on the corresponding unit area in the embryo model; obtain the verification information corresponding to the unit area, verify the unit detection data according to the verification information, and obtain the corresponding detection verification result;

Furthermore, the method for verifying the unit detection data according to the verification information includes:

Generate a corresponding unit detection graph according to the embryo model, verification information and unit detection data, perform shape evaluation according to the unit detection graph, and obtain a corresponding shape judgment result;

Acquire corresponding calibration grayscale image and unit detection grayscale image according to calibration information and unit detection data;

Compare the verification grayscale image and the unit detection grayscale image to obtain a grayscale difference surface; analyze the grayscale difference surface to determine the abnormal area, identify the area of the abnormal area, make a judgment based on the area of the abnormal area, and obtain the corresponding grayscale verification result;

The shape judgment results and grayscale verification results are integrated into the detection verification results.

Furthermore, the method for analyzing the grayscale difference surface includes:

Set the grayscale limit; establish the corresponding conversion formula according to the grayscale limit Where: r(x, y) is the grayscale difference of the corresponding position in the grayscale difference plane; (x, y) is the coordinate of the corresponding position; R is the grayscale limit; the output data is the conversion value 1 or 0;

The grayscale difference surface is processed by the conversion formula to obtain the conversion value of each position, and the area corresponding to the conversion value of 1 is marked as an abnormal area.

Step 7: Return to step 4 until all detections are completed according to the recognition sequence.

A defect detection system for enamel colored porcelain, comprising a standard analysis module, a collection module and a detection module;

The standard analysis module is used to establish an embryo model according to the standard enamel porcelain, and segment the embryo model to obtain each unit area; identify and sort each unit area to obtain an identification sequence; and set corresponding positioning points according to the standard enamel porcelain; set corresponding verification information for each unit area;

The acquisition module is used to acquire images of the enamel porcelain to be inspected to obtain an inspection image; obtain an identification sequence, perform positioning identification on the inspection image according to a preset positioning point, determine the position of the correspondingly ordered unit area in the identification sequence in the inspection image, and mark it as the initial positioning area; perform grayscale processing on the inspection image to obtain an inspection grayscale image, mark the initial positioning area in the inspection grayscale image, perform image processing according to the initial positioning area, and determine the unit inspection data corresponding to the unit area;

The detection module is used to perform defect detection on enamel porcelain, locate and mark the unit detection data on the corresponding unit area in the embryo model; obtain the verification information corresponding to the unit area, verify the unit detection data according to the verification information, and obtain the corresponding detection verification result; and so on, until all detections are completed according to the recognition sequence.

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

The present invention can realize intelligent detection of enamel colored porcelain. By segmenting and distributing the identification and verification of enamel colored porcelain, the adverse effects caused by the complexity and diversity of the surface of enamel colored porcelain can be reduced, and the detection accuracy can be improved. Newly emerged production defects can also be identified, thereby improving the comprehensiveness and applicability of detection. Human participation can be greatly reduced, labor costs can be reduced, and the stability of detection can be improved. These high-risk operations can also be replaced by humans, thereby reducing safety risks in the production process.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

FIG1 is a flow chart of the method of the present invention.

DETAILED DESCRIPTION

The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

As shown in FIG1 , a method for detecting defects of enamel colored porcelain includes:

Step 1: Obtain standard enamel colored porcelain, establish an embryo model according to the standard enamel colored porcelain, wherein the embryo model is a three-dimensional data model established according to the standard enamel colored porcelain, and segment the embryo model according to the standard enamel colored porcelain to obtain each unit area;

Methods for segmenting embryo models according to standard enamel porcelain include:

Obtain a standard image corresponding to the standard enamel porcelain, identify the boundaries of each pattern in the standard image, which refers to the boundaries of different colors. The intersection of different colors will form a boundary; use the identified boundary as a reference to perform a certain shift to both sides and then perform segmentation to obtain each initial image; avoid inaccurate direct segmentation, and perform segmentation after shifting, which will ensure the accuracy of the initial image. Adjacent initial images will form a transition area due to the shifting segmentation, and the transition area will be analyzed later, and the transition area will be assigned a corresponding initial image; this step can be quickly segmented based on existing technology; such as establishing an initial segmentation model based on a deep neural network; create a sample data set, the sample data set includes material samples and manually annotated sample sets, the material samples are the original picture sets, and the manually annotated sample sets are the picture sets after the original pictures are format converted and the manually annotated initial images are processed, and the sample data set contains the material samples The ratio of the sample set to the manually labeled sample set is 2:1; the pictures in the manually labeled sample set are binarized and saved in a single-channel form, and the photo sample set and the manually labeled sample set are divided according to the ratio to form a second photo sample set and a second manually labeled sample set; a Linknet network model is built based on the Linknet network structure under the Pytorch deep learning framework, the parameters of the Linknet network model are set, the second photo sample set and the second manually labeled sample set are input into the Linknet network model, and the Linknet network model is trained based on the Pytorch deep learning framework. Multiple models are saved during the training process, and the model with the smallest error is selected as the initial segmentation model using the validation set data; the standard image is processed by the initial segmentation model to obtain each initial image and the corresponding transition area;

Performing grayscale processing on the standard image to obtain a standard grayscale image, marking each initial image and the corresponding transition area in the standard grayscale image; determining the initial image corresponding to each transition area and marking it as a reference image; setting the initial value of each reference image;

Identify the grayscale value of each pixel in the transition area, evaluate the grayscale value according to the initial value of each reference image, and determine which reference image the pixel of the grayscale value belongs to. Because the grayscale values of different colors have large differences, they can be compared intuitively. In order to facilitate processing, a judgment model is established according to the above steps. The expression of the judgment model is: Wherein: s is the input data, i.e., the gray value of the corresponding pixel in the transition area; the proximity standard is set according to the initial value of the reference image, such as setting the interval range of each reference image according to each initial value, and taking the corresponding interval range as the proximity standard, or directly taking whether its gray value is closest to the initial value of the reference image as the proximity standard; the gray value of each pixel in the transition area is analyzed by the judgment model to obtain the reference image to which each pixel belongs, and the corresponding pixels are merged into the corresponding reference image; the current reference images or initial images are marked as unit images; the embryo model is segmented according to the position area corresponding to the unit image to obtain the unit area corresponding to each unit image.

The method of setting the initial value of each reference image includes:

Step SA1: Identify the grayscale value of each pixel in the reference image, merge adjacent pixels with the same grayscale value to obtain each first region; calculate the average value of the grayscale value corresponding to each pixel in the first region, and mark it as the first region value;

Step SA2: Calculate the difference between the first region values of adjacent first regions, marked as the first difference; merge the adjacent first regions whose first difference is not greater than the threshold value X1 to obtain a new first region; determine the new first region value;

Step SA3: looping step SA2 until all first differences are greater than the threshold value X1; identifying the area of each first region;

According to the formula Calculate the corresponding initial value;

In the formula: CZ is the initial value; v represents the corresponding first region, c=1, 2, ..., v, v is a positive integer; Ac represents the corresponding first region area; DZc represents the first region value corresponding to the first region.

Step 2: Identify and sort each unit area to obtain an identification sequence; and set corresponding positioning points according to standard enamel porcelain; the positioning points are used to determine the position of each image during subsequent image recognition; according to the prior art, positioning points can be set, and multiple positioning points can be set; corresponding verification information is set for each unit area, and the verification information is set for the standard unit image corresponding to the unit area; including corresponding standard verification data; such as shape, size, gray value and other related data;

The method for identifying and sorting each unit area includes:

Obtain the unit images corresponding to each unit area, and mark each unit image identically, that is, mark the same unit image with an identical marking symbol, and judge whether each unit image is identical, because there may be many identical images in enamel porcelain;

According to the identification sequence combination of each unit image, each candidate combination is obtained, and each unit image with the same mark is continuously in the same order; that is, various possible combinations are determined according to each unit image and marked as candidate combinations;

Based on the standard image of the standard enamel porcelain, each candidate combination is recognized and simulated, that is, the corresponding unit images in the standard image are recognized in the order of the candidate combinations, and the recognized unit images are extracted. After the extraction, the standard image will not have the unit images, and then the recognition is performed in the recognition order after the candidate combination, and so on, until the recognition is completed; the corresponding recognition efficiency and recognition accuracy are determined according to the recognition results;

The obtained recognition efficiency and recognition accuracy are marked as XL and DL respectively;

Calculate the corresponding simulation value according to the formula PQ = b1 × XL + b2 × DL;

Where: PQ is the simulation value; b1 and b2 are both proportional coefficients, with a value range of 0<b1≤1, 0<b2≤1; XL is the recognition efficiency; DL is the recognition accuracy;

The recognition sequence is set according to the candidate combination with the largest simulation value.

Step 3: Capture images of the enamel porcelain to be inspected to obtain inspection images;

Step 4: Get the recognition sequence, locate and identify the detection image according to the preset positioning points, determine the position of the corresponding ordered unit area in the recognition sequence in the detection image, and mark it as the initial positioning area; the initial positioning area is an approximate position, and the corresponding position in the detection image can be determined based on the coordinate area of the positioning point and the unit area, but there will generally be a certain deviation, but it does not affect the rapid positioning of the corresponding position area; the corresponding ordered unit area in the recognition sequence is determined based on the image that has been recognized.

Step 5: grayscale processing is performed on the detection image to obtain a detection grayscale image, an initial positioning area is marked in the detection grayscale image, image processing is performed on the initial positioning area, and unit detection data corresponding to the unit area is determined;

The method of performing image processing on the initial positioning area includes:

Step SC1: Identify and locate the grayscale value of each pixel in the initial area, merge adjacent pixels with the same grayscale value to obtain a second area;

Step SC2: Calculate the second difference between each adjacent second area, that is, the difference of the corresponding gray value; merge the adjacent second areas whose second difference is not greater than the threshold value X2 to obtain a new second area, the threshold value X2 is set according to the maximum gray value difference of the color that may exist in the corresponding verification information; use the average gray value as the basis for subsequent calculations;

Step SC3: looping step SC2 until there is no second difference value not greater than the threshold value X2, determining the unit initial region according to the remaining second regions, that is, according to the real gray value of the unit region, the second region corresponding to the unit region can be determined from the second regions and marked as the unit initial region;

Step SC4: Identify the grayscale value of each pixel outside the initial positioning area and adjacent to the unit initial area, and mark the corresponding pixel as an out-of-bounds pixel; determine whether each out-of-bounds pixel meets the merging requirement, that is, whether the difference is not greater than the threshold value X2; merge each out-of-bounds pixel that meets the merging requirement with the unit initial area to obtain a new unit initial area;

Step SC5: looping step SC4 until there are no out-of-bounds pixels that meet the merging requirements, marking the unit initial area as the unit control area; obtaining the position data of the unit control area, the gray value of each pixel, the shape data, the gray image and other related data, and integrating them into the unit detection data;

Step 6: Mark the unit detection data on the corresponding unit area in the embryo model, that is, mark the unit area according to the corresponding positioning position in the unit detection data to understand whether it is misplaced; obtain the verification information corresponding to the unit area, verify the unit detection data according to the verification information, and obtain the corresponding detection verification result;

The method for verifying unit detection data according to verification information includes:

According to the unit detection shape and position corresponding to the unit detection data of the embryo model recognition unit, a corresponding unit detection map is formed, and the unit area shape and position corresponding to the unit area are marked in the unit detection map; shape evaluation is performed according to the unit detection map to determine whether the shape is qualified, mainly from the two directions of position deviation and shape similarity, and corresponding judgment standards are preset; corresponding shape judgment results are obtained, including qualified shape judgment and unqualified shape judgment;

According to the calibration information and the unit detection data, a corresponding calibration grayscale image and a unit detection grayscale image are obtained. The calibration grayscale image is the grayscale image corresponding to the unit area in the standard grayscale image; the unit detection grayscale image is the grayscale image of the unit control area;

The calibration grayscale image and the unit detection grayscale image are compared to obtain a grayscale difference surface; that is, the position is first adjusted according to whether the position is deviated, and then the grayscale values of the corresponding pixels are subtracted to obtain the absolute value of the grayscale difference corresponding to each pixel to form a grayscale difference surface;

The grayscale difference surface is analyzed to determine the abnormal area, identify the area of the abnormal area, make a judgment based on the area of the abnormal area, and obtain the corresponding grayscale verification result; that is, when the area of the abnormal area exceeds the preset value, the grayscale verification is judged to be unqualified, otherwise, the grayscale verification is judged to be qualified.

The shape judgment results and grayscale verification results are integrated into the detection verification results.

Methods for analyzing the grayscale difference surface include:

According to the inspection requirements of enamel porcelain, set the corresponding maximum grayscale difference, marked as the grayscale limit; establish the corresponding conversion formula according to the grayscale limit Where: r(x, y) is the grayscale difference of the corresponding position in the grayscale difference plane; (x, y) is the coordinate of the corresponding position; R is the grayscale limit; the output data is the conversion value 1 or 0;

The grayscale difference surface is processed by the conversion formula to obtain the conversion value of each position, and the area corresponding to the conversion value of 1 is marked as an abnormal area.

Step 7: Return to step 4 until all detections are completed according to the recognition sequence.

The present invention can realize intelligent detection of enamel colored porcelain. By segmenting and distributing the identification and verification of enamel colored porcelain, the adverse effects caused by the complexity and diversity of the surface of enamel colored porcelain can be reduced, and the detection accuracy can be improved. Newly emerged production defects can also be identified, thereby improving the comprehensiveness and applicability of detection. Human participation can be greatly reduced, labor costs can be reduced, and the stability of detection can be improved. These high-risk operations can also be replaced by humans, thereby reducing safety risks in the production process.

A defect detection system for enamel colored porcelain, comprising a standard analysis module, a collection module and a detection module;

The standard analysis module is used to establish an embryo model according to the standard enamel porcelain, and segment the embryo model to obtain each unit area; identify and sort each unit area to obtain an identification sequence; and set corresponding positioning points according to the standard enamel porcelain; and set corresponding verification information for each unit area.

The acquisition module is used to acquire images of the enamel porcelain to be inspected to obtain a detection image; obtain a recognition sequence, locate and identify the detection image according to a preset positioning point, determine the position of the correspondingly ordered unit area in the recognition sequence in the detection image, and mark it as the initial positioning area; perform grayscale processing on the detection image to obtain a detection grayscale image, mark the initial positioning area in the detection grayscale image, perform image processing according to the initial positioning area, and determine the unit detection data corresponding to the unit area.

The detection module is used to perform defect detection on enamel porcelain, locate and mark the unit detection data on the corresponding unit area in the embryo model; obtain the verification information corresponding to the unit area, verify the unit detection data according to the verification information, and obtain the corresponding detection verification result; and so on, until all detections are completed according to the recognition sequence.

The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, is also included in the patent protection scope of the present invention.

It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

For the convenience of description, the above devices are described in terms of functions and are divided into various units and described separately. Of course, when implementing the present application, the functions of each unit can be implemented in the same one or more software and/or hardware. It should be understood by those skilled in the art that the embodiments of the present invention can be provided as methods, systems or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

The present invention is described with reference to flowcharts and/or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and/or box in the flowchart and/or block diagram, as well as the combination of the processes and/or boxes in the flowchart and/or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one process or multiple boxes in the flowchart.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

Claims (9)

1. A method for detecting defects in enamel colored porcelain, characterized in that the method comprises: Step 1: Obtain standard enamel colored porcelain, establish an embryo model according to the standard enamel colored porcelain, segment the embryo model according to the standard enamel colored porcelain, and obtain each unit area; Step 2: Identify and sort each unit area to obtain an identification sequence; set corresponding positioning points according to standard enamel porcelain; set corresponding verification information for each unit area; Step 3: Capture images of the enamel porcelain to be inspected to obtain inspection images; Step 4: Obtain the recognition sequence, perform positioning recognition on the detection image according to the preset positioning points, determine the position of the correspondingly ordered unit area in the recognition sequence in the detection image, and mark it as the initial positioning area; Step 5: grayscale processing is performed on the detection image to obtain a detection grayscale image, an initial positioning area is marked in the detection grayscale image, image processing is performed according to the initial positioning area, and unit detection data corresponding to the unit area is determined; Step 6: Position and mark the unit detection data on the corresponding unit area in the embryo model; obtain the verification information corresponding to the unit area, verify the unit detection data according to the verification information, and obtain the corresponding detection verification result; Step 7: Return to step 4 until all detections are completed according to the recognition sequence. 2. The method for detecting defects of enamel colored porcelain according to claim 1, characterized in that the method for segmenting the embryo model according to the standard enamel colored porcelain comprises: Obtaining a standard image corresponding to the standard enamel porcelain, segmenting the standard image, and obtaining initial images and transition areas between the initial images; Performing grayscale processing on the standard image to obtain a standard grayscale image, marking each initial image and transition area in the standard grayscale image; determining the initial image corresponding to each transition area and marking it as a reference image; and setting an initial value of each reference image; Identify the grayscale value of each pixel in the transition area, analyze each grayscale value through a preset judgment model, and determine the reference image corresponding to each pixel; Each pixel in the transition area is merged into the corresponding reference image to obtain each unit image; the embryo model is segmented according to the position area of the unit image in the embryo model to obtain each unit area. 3. The method for detecting defects in enamel porcelain according to claim 2, wherein the method for setting the initial value of each reference image comprises: Step SA1: identifying the gray value of each pixel in the reference image, merging adjacent pixels with the same gray value to obtain each first region; and determining the first region value of the first region; Step SA2: Calculate the difference between the first region values of adjacent first regions, marked as the first difference; merge the adjacent first regions whose first difference is not greater than the threshold value X1 to obtain a new first region; determine the new first region value; Step SA3: looping step SA2 until all first differences are greater than the threshold value X1; identifying the area of each first region; According to the formula Calculate the corresponding initial value; In the formula: CZ is the initial value; v represents the corresponding first region, c=1, 2, ..., v, v is a positive integer; Ac represents the corresponding first region area; DZc represents the first region value corresponding to the first region. 4. The method for detecting defects in enamel colored porcelain according to claim 2, wherein the expression of the judgment model is: Where: s is the gray value of the corresponding pixel in the transition area. 5. The method for detecting defects in enamel porcelain according to claim 2, characterized in that the method for identifying and sorting each unit area comprises: Obtain unit images corresponding to each unit area, and mark each unit image equally; perform identification sequence combination according to each unit image to obtain each candidate combination; Based on the standard image of standard enamel porcelain, each candidate combination is recognized and simulated to obtain the corresponding recognition efficiency and recognition accuracy; Calculate the corresponding simulation value according to the formula PQ = b1 × XL + b2 × DL; Where: PQ is the simulation value; b1 and b2 are both proportional coefficients, with a value range of 0<b1≤1, 0<b2≤1; XL is the recognition efficiency; DL is the recognition accuracy; The recognition sequence is set according to the candidate combination with the largest simulation value. 6. The method for defect detection of enamel porcelain according to claim 1, characterized in that the method of performing image processing on the initial positioning area comprises: Step SC1: Identify and locate the grayscale value of each pixel in the initial area, merge adjacent pixels with the same grayscale value to obtain a second area; Step SC2: calculating the second difference between each adjacent second region; merging the adjacent second regions whose second difference is not greater than the threshold value X2 to obtain a new second region; Step SC3: looping step SC2 until there is no second difference value not greater than the threshold value X2, and determining the unit initial region according to the remaining second regions; Step SC4: Identify and locate each pixel outside the initial region and adjacent to the unit initial region, and mark them as out-of-bounds pixels; identify the grayscale value of each out-of-bounds pixel, evaluate whether each out-of-bounds pixel meets the merging requirement, merge each out-of-bounds pixel that meets the merging requirement with the unit initial region, and obtain a new unit initial region; Step SC5: loop step SC4 until there are no out-of-bounds pixels that meet the merging requirements, mark the unit initial area as the unit control area; and obtain unit detection data according to the unit control area. 7. The method for defect detection of enamel porcelain according to claim 1, characterized in that the method for verifying the unit detection data according to the verification information comprises: Generate a corresponding unit detection graph according to the embryo model, verification information and unit detection data, perform shape evaluation according to the unit detection graph, and obtain a corresponding shape judgment result; Acquire corresponding calibration grayscale image and unit detection grayscale image according to calibration information and unit detection data; Compare the verification grayscale image and the unit detection grayscale image to obtain a grayscale difference surface; analyze the grayscale difference surface to determine the abnormal area, identify the area of the abnormal area, make a judgment based on the area of the abnormal area, and obtain the corresponding grayscale verification result; The shape judgment results and grayscale verification results are integrated into the detection verification results. 8. The method for detecting defects of enamel porcelain according to claim 7, characterized in that the method for analyzing the grayscale difference surface comprises: Set the grayscale limit; establish the corresponding conversion formula according to the grayscale limit Where: r(x, y) is the grayscale difference of the corresponding position in the grayscale difference plane; (x, y) is the coordinate of the corresponding position; R is the grayscale limit; the output data is the conversion value 1 or 0; The grayscale difference surface is processed by the conversion formula to obtain the conversion value of each position, and the area corresponding to the conversion value of 1 is marked as an abnormal area. 9. A defect detection system for enamel colored porcelain, characterized in that it implements a defect detection method for enamel colored porcelain as claimed in any one of claims 1 to 8, comprising a standard analysis module, a collection module and a detection module; The standard analysis module is used to establish an embryo model according to the standard enamel porcelain, and segment the embryo model to obtain each unit area; identify and sort each unit area to obtain an identification sequence; and set corresponding positioning points according to the standard enamel porcelain; set corresponding verification information for each unit area; The acquisition module is used to acquire images of the enamel porcelain to be inspected to obtain a detection image; obtain a recognition sequence, perform positioning recognition on the detection image according to a preset positioning point, determine the position of the correspondingly ordered unit area in the recognition sequence in the detection image, and mark it as the initial positioning area; perform grayscale processing on the detection image to obtain a detection grayscale image, mark the initial positioning area in the detection grayscale image, perform image processing according to the initial positioning area, and determine the unit detection data corresponding to the unit area; The detection module is used to perform defect detection on enamel porcelain, locate and mark the unit detection data on the corresponding unit area in the embryo model; obtain the verification information corresponding to the unit area, verify the unit detection data according to the verification information, and obtain the corresponding detection verification result; and so on, until all detections are completed according to the recognition sequence.