CN116777861B - Marking quality detection method and system for laser engraving machine - Google Patents
Marking quality detection method and system for laser engraving machine Download PDFInfo
- Publication number
- CN116777861B CN116777861B CN202310710612.9A CN202310710612A CN116777861B CN 116777861 B CN116777861 B CN 116777861B CN 202310710612 A CN202310710612 A CN 202310710612A CN 116777861 B CN116777861 B CN 116777861B
- Authority
- CN
- China
- Prior art keywords
- image
- marking
- information
- sets
- quality detection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 200
- 238000010147 laser engraving Methods 0.000 title claims abstract description 55
- 238000010276 construction Methods 0.000 claims abstract description 200
- 238000004458 analytical method Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000007689 inspection Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 22
- 238000012360 testing method Methods 0.000 claims description 14
- 238000012795 verification Methods 0.000 claims description 13
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 238000002372 labelling Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 5
- 230000007547 defect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000010205 computational analysis Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000004925 denaturation Methods 0.000 description 1
- 230000036425 denaturation Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010330 laser marking Methods 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a marking quality detection method and a marking quality detection system for a laser engraving machine, which are applied to the technical field of data processing, wherein the method comprises the following steps: marking images by a target laser engraving machine are collected, T image areas are obtained by dividing, marking data in a preset time range in the past are obtained, probability of marking quality problems is obtained, error probability information is obtained, and in a detection model construction parameter analysis model, construction data scale information and accuracy information are obtained. And obtaining a historical marking image and marking depth information set, and constructing and obtaining T marking quality detection models. Dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, and respectively inputting T marking quality detection models to obtain T marking quality detection results. The technical problems of low detection efficiency and low detection accuracy in the marking quality detection of the laser engraving machine in the prior art are solved.
Description
Technical Field
The invention relates to the field of data processing, in particular to a marking quality detection method and system for a laser engraving machine.
Background
The laser engraving machine utilizes numerical control technology to control laser to irradiate on the surface or inside of the processed material so as to gasify or physically denature the processed material, thereby realizing laser engraving. However, in the prior art, marking quality detection of a laser engraving machine is mostly carried out in a macroscopic observation mode, and the problems of low detection efficiency and low detection accuracy exist.
Therefore, in the prior art, the marking quality detection of the laser engraving machine has the technical problems of low detection efficiency and low detection accuracy.
Disclosure of Invention
The application provides a marking quality detection method and a marking quality detection system for a laser engraving machine, which solve the technical problems of low detection efficiency and low detection accuracy in the marking quality detection of the laser engraving machine in the prior art.
The application provides a marking quality detection method of a laser engraving machine, which comprises the following steps: marking a preset image by adopting a target laser engraving machine, and collecting a marked image; dividing the preset image to obtain T image areas, and analyzing the probability of marking quality problems of the T image areas according to marking data of the preset image in a past preset time range to obtain T error probability information; inputting the T pieces of error probability information into a detection model construction parameter analysis model to obtain T pieces of construction data scale information and T pieces of accuracy information; according to the marking data of the preset image within the past preset time range, a historical marking image set and a historical marking depth information set are obtained, and the T image areas are divided according to the T image areas, so that T area image sets and T marking depth information sets are obtained; according to the T construction data scale information, construction data selection is carried out on the T regional image sets and the T marking depth information sets to obtain T image construction data sets and T depth construction data sets, the T image construction data sets and the T depth construction data sets are adopted to construct and obtain T marking quality detection models according to the T accuracy information, and each marking quality detection model comprises an image quality detection unit and a depth quality detection unit; dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, and respectively inputting the T marking quality detection models to obtain T marking quality detection results.
The application also provides a marking quality detection system of the laser engraving machine, which comprises: the image acquisition module is used for acquiring a marking image after marking a preset image by adopting a target laser engraving machine; the error probability information acquisition module is used for dividing the preset image to obtain T image areas, and analyzing the probability of occurrence of marking quality problems of the T image areas according to marking data of the preset image within a past preset time range to obtain T error probability information; the accuracy information acquisition module is used for inputting the T error probability information into a detection model construction parameter analysis model to obtain T construction data scale information and T accuracy information; the image region dividing module is used for obtaining a historical marking image set and a historical marking depth information set according to marking data of the preset image within a past preset time range, dividing the image region according to the T image regions and obtaining T region image sets and T marking depth information sets; the marking quality detection model construction module is used for selecting construction data from the T regional image sets and the T marking depth information sets according to the T construction data scale information to obtain T image construction data sets and T depth construction data sets, constructing and obtaining T marking quality detection models by adopting the T image construction data sets and the T depth construction data sets according to the T accuracy information, wherein each marking quality detection model comprises an image quality detection unit and a depth quality detection unit; and the quality detection result acquisition module is used for dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, and respectively inputting the T marking quality detection models to obtain T marking quality detection results.
The application also provides an electronic device, comprising:
a memory for storing executable instructions;
and the processor is used for realizing the marking quality detection method of the laser engraving machine when executing the executable instructions stored in the memory.
The embodiment of the application provides a computer readable storage medium which stores a computer program, and when the program is executed by a processor, the marking quality detection method of the laser engraving machine is realized.
The marking quality detection method and system for the laser engraving machine provided by the application are characterized in that the marking images of the target laser engraving machine are collected and divided to obtain T image areas, marking data in the past preset time range are obtained, the probability of marking quality problems is obtained, error probability information is obtained, and the error probability information is input into a detection model construction parameter analysis model to obtain construction data scale information and accuracy information. And obtaining a historical marking image and marking depth information set, and constructing and obtaining T marking quality detection models. Dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, respectively inputting T marking quality detection models to obtain T marking quality detection results, and realizing the rapid and accurate detection of the marking quality of the laser engraving machine. According to the application, the accuracy of marking quality detection can be improved by partitioning the laser marking image, the possible omission of defects and flaws caused by overall quality inspection is avoided, the data scale and accuracy of model construction are set according to the probability of quality problems in the past time of each region, the calculation force resources required by model construction can be reduced for the region with low probability of quality problems, the model construction speed is improved, the construction data is more for the region with high probability of quality problems, the accuracy requirement is higher, the accuracy of quality detection is ensured, and the technical problems of low detection efficiency and low detection accuracy in marking quality detection of a laser engraving machine in the prior art are solved.
The foregoing description is only an overview of the present application, and is intended to be implemented in accordance with the teachings of the present application in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present application more readily apparent.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly described below. It is apparent that the figures in the following description relate only to some embodiments of the present disclosure and are not limiting of the present disclosure.
Fig. 1 is a schematic flow chart of a marking quality detection method of a laser engraving machine according to an embodiment of the present application;
fig. 2 is a schematic flow chart of obtaining T error probability information by using the marking quality detection method of a laser engraving machine according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of a method for detecting marking quality of a laser engraving machine according to an embodiment of the present application to obtain a detection model and construct a parameter analysis model;
fig. 4 is a schematic structural diagram of a system of a marking quality detection method of a laser engraving machine according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a system electronic device of a marking quality detection method of a laser engraving machine according to an embodiment of the present application.
Reference numerals illustrate: the system comprises an image acquisition module 11, an error probability information acquisition module 12, an accuracy information acquisition module 13, an image region division module 14, a marking quality detection model construction module 15, a quality detection result acquisition module 16, a processor 31, a memory 32, an input device 33 and an output device 34.
Detailed Description
Example 1
The present application will be further described in detail with reference to the accompanying drawings, for the purpose of making the objects, technical solutions and advantages of the present application more apparent, and the described embodiments should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making any inventive effort are within the scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict.
In the following description, the terms "first", "second", "third" and the like are merely used to distinguish similar objects and do not represent a particular ordering of the objects, it being understood that the "first", "second", "third" may be interchanged with a particular order or sequence, as permitted, to enable embodiments of the application described herein to be practiced otherwise than as illustrated or described herein.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the application only.
While the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server, the modules are merely illustrative, and different aspects of the system and method may use different modules.
A flowchart is used in the present application to describe the operations performed by a system according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
As shown in fig. 1, an embodiment of the present application provides a method for detecting marking quality of a laser engraving machine, including:
s10: after marking a preset image by adopting a target laser engraving machine, acquiring a marking image;
S20: dividing the preset image to obtain T image areas, and analyzing the probability of marking quality problems of the T image areas according to marking data of the preset image in a past preset time range to obtain T error probability information;
s30: inputting the T pieces of error probability information into a detection model construction parameter analysis model to obtain T pieces of construction data scale information and T pieces of accuracy information;
specifically, the laser engraving machine utilizes a numerical control technology to control laser irradiation on a processing material to generate physical denaturation of the processing material, so as to realize laser engraving. When the marking quality of the laser engraving machine is improved, marking images after marking of preset images are carried out by collecting the target laser engraving machine, wherein the preset images are images designed before laser engraving marking, and the marking images are images obtained after laser engraving marking is completed. Then, the preset image is subjected to image division to obtain T image areas, wherein the T image areas are preferably divided into T image areas with the same area, and T is a positive integer greater than 1, for example, 10. And analyzing and acquiring the probability of marking quality problems of the T image areas according to marking data of the preset image within the past preset time range, so as to acquire T error probability information. Further, the T pieces of error probability information are input into a detection model construction parameter analysis model, the parameter analysis model is constructed through the detection model, and the T pieces of construction data scale information and the T pieces of accuracy information corresponding to the T pieces of error probability information are respectively obtained.
As shown in fig. 2, the method S20 provided by the embodiment of the present application further includes:
s21: dividing the preset image to obtain T image areas;
s22: marking quality inspection data in the past preset time of the preset image is obtained according to T historical area marking quality inspection data sets, wherein the marking quality inspection data comprise quality inspection data of whether quality problems occur or not;
s23: and calculating and obtaining the T error probability information according to the proportion of quality problems in the T historical region marking quality inspection data sets.
Specifically, image region division is performed on the marked preset image, and the preset image is divided into a plurality of square regions according to the image size. The specific dividing number can be set according to actual conditions, and T image areas are obtained, wherein T is the specific dividing number. And dividing according to the marking quality inspection data of the preset image within the past preset time and the T image areas to obtain T historical area marking quality inspection data sets, namely the historical marking quality inspection data of the T areas. The marking quality inspection data includes quality inspection data of whether quality problems occur in each region, for example, the size of the marking image does not meet the preset image requirement, or flaws occur due to laser spots and the like. And obtaining the T pieces of error probability information according to the proportion of quality problems in the T pieces of historical area marking quality inspection data sets, namely the proportion of the T pieces of error probability information representing the quality problems in the T pieces of historical area marking quality inspection data sets.
As shown in fig. 3, the method S30 provided by the embodiment of the present application further includes:
s31: acquiring a sample error probability information set, and analyzing and acquiring a sample construction data scale information set and a sample accuracy information set, wherein the larger the sample error probability information is, the larger the construction data scale information is and the larger the sample accuracy information is;
s32: constructing a construction data scale analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample construction data scale information set;
s33: and constructing an accuracy analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample accuracy information set to obtain the detection model construction parameter analysis model.
Specifically, a sample error probability information set is obtained, wherein the sample error probability information set is a numerical value set for recording error probability information of all history areas. In one embodiment. The method comprises the steps of analyzing a sample error probability information set by a professional evaluator to obtain a sample construction data scale information set and a sample accuracy information set, wherein sample construction data scale information in the sample construction data scale information set is the data quantity of sample construction data for constructing a marking quality detection model, the sample accuracy information set is the accuracy of model output during training of the marking quality detection model, and different error probability information in the sample error probability information set corresponds to one element in the sample construction data scale information set and one element in the sample accuracy information set.
In one embodiment, the sample construction data size information set and the sample accuracy information set are obtained from the sample error probability information set by computational analysis. For example, according to the error probability average value of the sample error probability information set, the corresponding data scale is set to 5000 groups of data, and the accuracy is 90%. Setting a data scale interval to 2000-10000 groups, setting an accuracy interval to 80-95%, and then calculating and setting corresponding sample construction data scale and sample accuracy in the data scale interval and the accuracy interval according to the difference value of the error probability of each sample and the mean value and the percentage of the mean value to obtain a sample construction data scale information set and a sample accuracy information set.
The larger the sample error probability information of a certain area is, the more quality problems of the area are, when the construction model performs marking quality detection, the more accurate the output results of the constructed T marking quality detection models are by increasing the data scale of the construction model and improving the accuracy information, the smaller the sample error probability information is, the less the quality problems of the area are, and by reducing the data scale of the construction model and properly reducing the accuracy information, the workload of model construction is reduced, the training efficiency of the T marking quality detection models is improved, and the accuracy of marking quality detection is met.
And constructing a construction data scale analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample construction data scale information set. The data scale analysis unit is used for outputting the data quantity for constructing the marking quality detection model according to the specific error probability information. And constructing an accuracy analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample accuracy information set to obtain the detection model construction parameter analysis model. The accuracy analysis unit is used for outputting the model output accuracy of the marking quality detection model according to the specific error probability information. Because the training time of different data volumes and different output accuracy of different marking quality detection models is different, a parameter analysis model is constructed by acquiring the detection models, and the targeted setting of the training parameters of the marking quality detection models in different areas is realized, so that the training efficiency and the model output accuracy of the marking quality detection models are improved.
The method S32 provided by the embodiment of the application further comprises the following steps:
s321: based on the error probability information as a decision feature, acquiring a plurality of division thresholds according to the sample error probability information set;
S322: adopting the multiple division thresholds to construct multi-layer decision division nodes, wherein each layer of decision division nodes carries out classification division decision on the input error probability information according to the division thresholds;
s323: obtaining a plurality of final division results of the multi-layer decision division node;
s324: and taking a plurality of sample construction data scale information in the sample construction data scale information set as a plurality of decision results and taking the plurality of sample construction data scale information as the decision results of the plurality of final division results to obtain the data scale analysis unit.
Specifically, based on error probability information as a decision feature, a plurality of division thresholds are acquired according to the sample error probability information set, and a decision tree diagram is constructed. Then, a plurality of division thresholds are adopted to construct multi-layer decision division nodes, wherein each planning division node corresponds to one division threshold. And the decision dividing nodes of each layer divide the input error probability information into a class which is larger than the dividing threshold value and a class which is smaller than or equal to the dividing threshold value according to the dividing threshold value. And then, obtaining a plurality of final division results of the multi-layer division decision by the multi-layer decision division node, taking a plurality of pieces of sample construction data scale information in the sample construction data scale information set as a plurality of decision results, correspondingly setting the decision results as the plurality of final division results according to the size of error probability information in the plurality of final division results, and obtaining the data scale analysis unit. And the error probability information is input into a data scale analysis unit, and the corresponding decision result is obtained, namely the data scale information is constructed. Furthermore, the construction mode of the accuracy analysis unit in the detection model construction parameter analysis model is the same as that of the data scale analysis unit, but specific decision results are obtained through a sample accuracy information set, and are not described in detail herein.
S40: according to the marking data of the preset image within the past preset time range, a historical marking image set and a historical marking depth information set are obtained, and the T image areas are divided according to the T image areas, so that T area image sets and T marking depth information sets are obtained;
s50: according to the T construction data scale information, construction data selection is carried out on the T regional image sets and the T marking depth information sets to obtain T image construction data sets and T depth construction data sets, the T image construction data sets and the T depth construction data sets are adopted to construct and obtain T marking quality detection models according to the T accuracy information, and each marking quality detection model comprises an image quality detection unit and a depth quality detection unit; and
s60: dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, and respectively inputting the T marking quality detection models to obtain T marking quality detection results.
Specifically, according to marking data of a preset image within a past preset time range, a historical marking image set and a historical marking depth information set are obtained, and the T image areas are divided according to the T image areas, so that T area image sets and T marking depth information sets are obtained. And selecting construction data from the T regional image sets and the T marking depth information sets according to the T construction data scale information to obtain T image construction data sets and T depth construction data sets, wherein the data scales of the T image construction data sets and the T depth construction data sets are consistent with the construction data scale information acquired by the corresponding regions. And constructing a data set by adopting the T images according to the T pieces of accuracy information, and constructing and obtaining T marking quality detection models by adopting the T pieces of image construction data sets and the T pieces of depth construction data sets, wherein the model output accuracy of the T marking quality detection models is consistent with the corresponding T pieces of accuracy information, and each marking quality detection model comprises an image quality detection unit and a depth quality detection unit. By acquiring the detection model construction parameter analysis model, construction data scale information and model output accuracy information corresponding to different detection areas are obtained, calculation force resources required by model construction are further reduced, model construction speed is improved, and quality detection requirements are guaranteed. The image quality detection unit is used for carrying out quality detection and evaluation on the image quality marked by the laser engraving machine, and the depth quality detection unit is used for carrying out quality detection and evaluation on the depth marked by the laser engraving machine. Finally, according to the T image areas, carrying out area division and depth detection on the marking image of the laser engraving machine, further obtaining T area images and T area depths marked by the laser engraving machine, respectively inputting the T marking quality detection models, obtaining T marking quality detection results, and realizing rapid and accurate detection on the marking quality of the laser engraving machine.
The method S40 provided by the embodiment of the application further comprises the following steps:
s41: according to the T construction data scale information, randomly selecting data from the T regional image sets respectively, and obtaining T selected regional image sets;
s42: marking quality ratings are carried out according to the region images in the T selected region image sets and the preset images, T image quality information sets are obtained, and the T image construction data sets are obtained by combining the T selected region image sets;
s43: according to the T construction data scale information, randomly selecting data from the T marking depth information sets respectively, and obtaining T selected marking depth information sets;
s44: and carrying out marking quality rating according to marking depth information in the T selected marking depth information sets to obtain T depth quality information sets, and combining the T selected marking depth information sets to obtain the T depth construction data sets.
Specifically, according to the T pieces of construction data scale information, the T pieces of construction data scale information of the regional image data are randomly selected from the T regional image sets, and the T selected regional image sets are obtained. And then, according to the obtained region images in the T selected region image sets, carrying out manual marking quality grading according to the preset image, namely the image designed before laser engraving marking, comparing the marked images of all regions with the preset image, judging whether defects exist and the defect degree, such as the size and the number of the defects, and further obtaining the quality grading of the images of all the regions, wherein the quality grading is lower as the defect degree is larger, and obtaining the image quality information sets of the T regions, namely the T image quality information sets according to the grading result. And then, combining the corresponding T selected area image sets to obtain T image construction data sets, wherein each image construction data set comprises the selected area image set of one image area and the corresponding image quality information set.
And according to the T pieces of construction data scale information, randomly selecting data from the T pieces of marking depth information sets respectively, namely randomly selecting each piece of region mark depth information corresponding to each region to construct data scale information from the T pieces of marking depth information sets according to the obtained T pieces of construction data scale information, and obtaining T pieces of selected marking depth information sets. And then, marking quality rating is carried out on each region in a manual identification mode according to the T selected marking depth information sets, whether the marking depth in each region meets the depth requirement is evaluated, further marking quality rating information of each region is obtained, the larger the deviation from the depth requirement is, the lower the quality rating is, the depth quality information sets of the T regions, namely T depth quality information sets, are obtained, the T selected marking depth information sets are combined, the T depth construction data sets are obtained, and each depth construction data set comprises one selected marking depth information set and the corresponding depth quality information set.
The method S40 provided by the embodiment of the application further comprises the following steps:
s45: constructing a first image quality detection unit in a first marking quality detection model according to first accuracy information in the T pieces of accuracy information by adopting a first image construction data set in the T pieces of image construction data sets;
S46: constructing a first depth quality detection unit in the first marking quality detection model according to first accuracy information in the T accuracy information by adopting a first depth construction data set in the T depth construction data sets;
s47: and (5) continuing to adopt other T-1 image construction data sets and depth construction data sets, and constructing other T-1 marking quality detection models according to other T-1 accuracy information.
Specifically, a first image construction dataset in T image construction datasets is adopted, the first image construction dataset is used as training data, a convolutional neural network model is trained, and a first image quality detection unit in a first marking quality detection model is constructed according to first accuracy information in the T accuracy information acquired by a detection model construction parameter analysis model. The first marking quality detection model corresponds to a random one of the T image regions. And training until the accuracy of the output result of the model meets the corresponding first accuracy information, and obtaining the trained model to obtain a first image quality detection unit. The first depth quality detection unit is constructed through a BP neural network or a feedforward neural network model, and the acquired first depth construction data set is identical to the acquired first accuracy information area.
And adopting a first depth construction data set in the T depth construction data sets, taking the first depth construction data set as training data, wherein the input data is a selected marking depth information set, and the output data is a corresponding depth quality information set. And performing supervision training on the BP neural network model, and constructing a first depth quality detection unit in the first marking quality detection model according to first accuracy information in the T accuracy information. And obtaining the trained model to obtain a first depth quality detection unit until the output result of the model meets the corresponding first accuracy information. Wherein the acquiring of the first depth build dataset is the same as the acquiring of the region of the first accuracy information. Further, other T-1 image construction data sets and depth construction data sets are continuously adopted, other T-1 marking quality detection models are constructed according to other T-1 accuracy information, the construction process is the same as that of the first marking quality detection model, but the construction data are the image construction data sets and depth construction data sets of other T-1 image areas, and the description is omitted here.
The method S45 provided by the embodiment of the application further comprises the following steps:
S451: performing data labeling and dividing on the first image construction data set to obtain a first training set, a first verification set and a first test set;
s452: based on a convolutional neural network, constructing the first image quality detection unit, wherein input data are first area images of first image areas in T image areas, and output data are first image quality information;
s453: and performing supervision training, verification and testing on the first image quality detection unit by adopting the first training set, the first verification set and the first testing set, and obtaining the first image quality detection unit when the accuracy of the first image quality detection unit meets the first accuracy information.
Specifically, the first image construction data set is subjected to data labeling and division, namely the first image construction data set is subjected to data division, the first image construction data set is divided into a first training set, a first verification set and a first test set according to a preset proportion, and the divided data is labeled. The preset proportion can be set according to actual conditions, and the proportion of the first training set is larger than that of the first verification set and the first test set, for example, the proportion can be 7:2:1. And constructing the first image quality detection unit based on a convolutional neural network, wherein input data is a first area image of a first image area in the T image areas, and output data is first image quality information reflecting the marking image quality of the first area. And performing supervision training, verification and testing on the first image quality detection unit by adopting the first training set, the first verification set and the first testing set, and obtaining the first image quality detection unit when the accuracy of the first image quality detection unit meets the first accuracy information.
According to the technical scheme provided by the embodiment of the invention, the target laser engraving machine is used for marking images, T image areas are obtained by dividing, marking data in a preset time range are obtained, the probability of marking quality problems is obtained, error probability information is obtained, and the error probability information is input into a detection model construction parameter analysis model to obtain construction data scale information and accuracy information. And obtaining a historical marking image and marking depth information set, and constructing and obtaining T marking quality detection models. And by acquiring the detection model construction parameter analysis model, construction data scale information and model output accuracy information corresponding to different detection areas are obtained, so that calculation force resources required by model construction are further reduced, model construction speed is improved, and quality detection requirements are ensured. Dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, respectively inputting T marking quality detection models to obtain T marking quality detection results, and realizing the rapid and accurate detection of the marking quality of the laser engraving machine. The technical problems of low detection efficiency and low detection accuracy in the marking quality detection of the laser engraving machine in the prior art are solved.
Example two
Based on the same inventive concept as the marking quality detection method of the laser engraving machine in the foregoing embodiment, the present invention also provides a system of the marking quality detection method of the laser engraving machine, which can be implemented by hardware and/or software, and can be generally integrated in an electronic device, for executing the method provided by any embodiment of the present invention. As shown in fig. 4, the system includes:
the image acquisition module 11 is used for acquiring a marking image after marking a preset image by adopting a target laser engraving machine;
the error probability information obtaining module 12 is configured to divide the preset image to obtain T image areas, and analyze the probability that the T image areas have a problem of marking quality according to marking data in a preset time range in the past of the preset image, so as to obtain T error probability information;
the accuracy information obtaining module 13 is configured to input the T pieces of error probability information into a detection model construction parameter analysis model, and obtain T pieces of construction data scale information and T pieces of accuracy information;
the image region dividing module 14 is configured to obtain a historical marking image set and a historical marking depth information set according to marking data in a past preset time range of the preset image, and divide the image region according to the T image regions to obtain T regional image sets and T marking depth information sets;
The marking quality detection model construction module 15 is configured to select construction data from the T area image sets and the T marking depth information sets according to the T construction data scale information, obtain T image construction data sets and T depth construction data sets, and construct and obtain T marking quality detection models by using the T image construction data sets and the T depth construction data sets according to the T accuracy information, where each marking quality detection model includes an image quality detection unit and a depth quality detection unit; and
the quality detection result obtaining module 16 is configured to divide and detect the depth of the marking image according to the T image areas, obtain T area images and T area depths, and input the T marking quality detection models respectively to obtain T marking quality detection results.
Further, the error probability information obtaining module 12 is further configured to:
dividing the preset image to obtain T image areas;
marking quality inspection data in the past preset time of the preset image is obtained according to T historical area marking quality inspection data sets, wherein the marking quality inspection data comprise quality inspection data of whether quality problems occur or not;
And calculating and obtaining the T error probability information according to the proportion of quality problems in the T historical region marking quality inspection data sets.
Further, the accuracy information obtaining module 13 is further configured to:
acquiring a sample error probability information set, and analyzing and acquiring a sample construction data scale information set and a sample accuracy information set, wherein the larger the sample error probability information is, the larger the construction data scale information is and the larger the sample accuracy information is;
constructing a construction data scale analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample construction data scale information set;
and constructing an accuracy analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample accuracy information set to obtain the detection model construction parameter analysis model.
Further, the accuracy information obtaining module 13 is further configured to:
based on the error probability information as a decision feature, acquiring a plurality of division thresholds according to the sample error probability information set;
adopting the multiple division thresholds to construct multi-layer decision division nodes, wherein each layer of decision division nodes carries out classification division decision on the input error probability information according to the division thresholds;
Obtaining a plurality of final division results of the multi-layer decision division node;
and taking a plurality of sample construction data scale information in the sample construction data scale information set as a plurality of decision results and taking the plurality of sample construction data scale information as the decision results of the plurality of final division results to obtain the data scale analysis unit.
Further, the image area dividing module 14 is further configured to:
according to the T construction data scale information, randomly selecting data from the T regional image sets respectively, and obtaining T selected regional image sets;
marking quality ratings are carried out according to the region images in the T selected region image sets and the preset images, T image quality information sets are obtained, and the T image construction data sets are obtained by combining the T selected region image sets;
according to the T construction data scale information, randomly selecting data from the T marking depth information sets respectively, and obtaining T selected marking depth information sets;
and carrying out marking quality rating according to marking depth information in the T selected marking depth information sets to obtain T depth quality information sets, and combining the T selected marking depth information sets to obtain the T depth construction data sets.
Further, the image area dividing module 14 is further configured to:
constructing a first image quality detection unit in a first marking quality detection model according to first accuracy information in the T pieces of accuracy information by adopting a first image construction data set in the T pieces of image construction data sets;
constructing a first depth quality detection unit in the first marking quality detection model according to first accuracy information in the T accuracy information by adopting a first depth construction data set in the T depth construction data sets;
and (5) continuing to adopt other T-1 image construction data sets and depth construction data sets, and constructing other T-1 marking quality detection models according to other T-1 accuracy information.
Further, the image area dividing module 14 is further configured to:
performing data labeling and dividing on the first image construction data set to obtain a first training set, a first verification set and a first test set;
based on a convolutional neural network, constructing the first image quality detection unit, wherein input data are first area images of first image areas in T image areas, and output data are first image quality information;
And performing supervision training, verification and testing on the first image quality detection unit by adopting the first training set, the first verification set and the first testing set, and obtaining the first image quality detection unit when the accuracy of the first image quality detection unit meets the first accuracy information.
The included units and modules are only divided according to the functional logic, but are not limited to the above-mentioned division, so long as the corresponding functions can be realized; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present invention.
Example III
Fig. 5 is a schematic structural diagram of an electronic device provided in a third embodiment of the present invention, and shows a block diagram of an exemplary electronic device suitable for implementing an embodiment of the present invention. The electronic device shown in fig. 5 is only an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention. As shown in fig. 5, the electronic device includes a processor 31, a memory 32, an input device 33, and an output device 34; the number of processors 31 in the electronic device may be one or more, in fig. 5, one processor 31 is taken as an example, and the processors 31, the memory 32, the input device 33 and the output device 34 in the electronic device may be connected by a bus or other means, in fig. 5, by bus connection is taken as an example.
The memory 32 is used as a computer readable storage medium for storing a software program, a computer executable program, and modules, such as program instructions/modules corresponding to a marking quality detection method of a laser engraving machine in an embodiment of the present invention. The processor 31 executes various functional applications of the computer device and data processing by running software programs, instructions and modules stored in the memory 32, i.e. implements a laser engraving machine marking quality detection method as described above.
Note that the above is only a preferred embodiment of the present invention and the technical principle applied. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.
Claims (10)
1. A method for detecting marking quality of a laser engraving machine, which is characterized by comprising the following steps:
After marking a preset image by adopting a target laser engraving machine, acquiring a marking image;
dividing the preset image to obtain T image areas, and analyzing the probability of marking quality problems of the T image areas according to marking data of the preset image in a past preset time range to obtain T error probability information;
inputting the T pieces of error probability information into a detection model construction parameter analysis model to obtain T pieces of construction data scale information and T pieces of accuracy information;
according to the marking data of the preset image within the past preset time range, a historical marking image set and a historical marking depth information set are obtained, and the T image areas are divided according to the T image areas, so that T area image sets and T marking depth information sets are obtained;
according to the T construction data scale information, construction data selection is carried out on the T regional image sets and the T marking depth information sets to obtain T image construction data sets and T depth construction data sets, the T image construction data sets and the T depth construction data sets are adopted to construct and obtain T marking quality detection models according to the T accuracy information, and each marking quality detection model comprises an image quality detection unit and a depth quality detection unit; and
Dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, and respectively inputting the T marking quality detection models to obtain T marking quality detection results.
2. The method of claim 1, wherein dividing the preset image to obtain T image areas, and analyzing the probability of occurrence of a marking quality problem for the T image areas according to marking data of the preset image within a preset time range in the past, comprises:
dividing the preset image to obtain T image areas;
marking quality inspection data in the past preset time of the preset image is obtained according to T historical area marking quality inspection data sets, wherein the marking quality inspection data comprise quality inspection data of whether quality problems occur or not;
and calculating and obtaining the T error probability information according to the proportion of quality problems in the T historical region marking quality inspection data sets.
3. The method of claim 1, wherein inputting the T error probability information into a test model build parameter analysis model to obtain T build data scale information and T accuracy information, comprises:
Acquiring a sample error probability information set, and analyzing and acquiring a sample construction data scale information set and a sample accuracy information set, wherein the larger the sample error probability information is, the larger the construction data scale information is and the larger the sample accuracy information is;
constructing a construction data scale analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample construction data scale information set;
and constructing an accuracy analysis unit in the detection model construction parameter analysis model by adopting the sample error probability information set and the sample accuracy information set to obtain the detection model construction parameter analysis model.
4. A method according to claim 3, wherein constructing a construction data scale analysis unit within the detection model construction parameter analysis model using the set of sample error probability information and the set of sample construction data scale information comprises:
based on the error probability information as a decision feature, acquiring a plurality of division thresholds according to the sample error probability information set;
adopting the multiple division thresholds to construct multi-layer decision division nodes, wherein each layer of decision division nodes carries out classification division decision on the input error probability information according to the division thresholds;
Obtaining a plurality of final division results of the multi-layer decision division node;
and taking a plurality of sample construction data scale information in the sample construction data scale information set as a plurality of decision results and taking the plurality of sample construction data scale information as the decision results of the plurality of final division results to obtain the data scale analysis unit.
5. The method of claim 1, wherein selecting the construction data from the T regional image sets and the T marking depth information sets according to the T construction data scale information to obtain T image construction data sets and T depth construction data sets, comprises:
according to the T construction data scale information, randomly selecting data from the T regional image sets respectively, and obtaining T selected regional image sets;
marking quality ratings are carried out according to the region images in the T selected region image sets and the preset images, T image quality information sets are obtained, and the T image construction data sets are obtained by combining the T selected region image sets;
according to the T construction data scale information, randomly selecting data from the T marking depth information sets respectively, and obtaining T selected marking depth information sets;
And carrying out marking quality rating according to marking depth information in the T selected marking depth information sets to obtain T depth quality information sets, and combining the T selected marking depth information sets to obtain the T depth construction data sets.
6. The method of claim 5, wherein constructing to obtain T marking quality detection models using the T image construction datasets and T depth construction datasets according to the T accuracy information, comprises:
constructing a first image quality detection unit in a first marking quality detection model according to first accuracy information in the T pieces of accuracy information by adopting a first image construction data set in the T pieces of image construction data sets;
constructing a first depth quality detection unit in the first marking quality detection model according to first accuracy information in the T accuracy information by adopting a first depth construction data set in the T depth construction data sets;
and (5) continuing to adopt other T-1 image construction data sets and depth construction data sets, and constructing other T-1 marking quality detection models according to other T-1 accuracy information.
7. The method of claim 6, wherein constructing a first image quality detection unit in a first marking quality detection model from a first one of the T accuracy information using a first one of the T image construction data sets, comprises:
Performing data labeling and dividing on the first image construction data set to obtain a first training set, a first verification set and a first test set;
based on a convolutional neural network, constructing the first image quality detection unit, wherein input data are first area images of first image areas in T image areas, and output data are first image quality information;
and performing supervision training, verification and testing on the first image quality detection unit by adopting the first training set, the first verification set and the first testing set, and obtaining the first image quality detection unit when the accuracy of the first image quality detection unit meets the first accuracy information.
8. A laser engraving machine marking quality detection system, the system comprising:
the image acquisition module is used for acquiring a marking image after marking a preset image by adopting a target laser engraving machine;
the error probability information acquisition module is used for dividing the preset image to obtain T image areas, and analyzing the probability of occurrence of marking quality problems of the T image areas according to marking data of the preset image within a past preset time range to obtain T error probability information;
The accuracy information acquisition module is used for inputting the T error probability information into a detection model construction parameter analysis model to obtain T construction data scale information and T accuracy information;
the image region dividing module is used for obtaining a historical marking image set and a historical marking depth information set according to marking data of the preset image within a past preset time range, dividing the image region according to the T image regions and obtaining T region image sets and T marking depth information sets;
the marking quality detection model construction module is used for selecting construction data from the T regional image sets and the T marking depth information sets according to the T construction data scale information to obtain T image construction data sets and T depth construction data sets, constructing and obtaining T marking quality detection models by adopting the T image construction data sets and the T depth construction data sets according to the T accuracy information, wherein each marking quality detection model comprises an image quality detection unit and a depth quality detection unit; and
the quality detection result acquisition module is used for dividing and detecting the depth of the marking image according to the T image areas to obtain T area images and T area depths, and inputting the T marking quality detection models respectively to obtain T marking quality detection results.
9. An electronic device, the electronic device comprising:
a memory for storing executable instructions;
a processor for implementing a laser engraving machine marking quality detection method according to any one of claims 1 to 7 when executing the executable instructions stored in the memory.
10. A computer readable medium having stored thereon a computer program, which when executed by a processor, implements a laser engraving machine marking quality detection method as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310710612.9A CN116777861B (en) | 2023-06-15 | 2023-06-15 | Marking quality detection method and system for laser engraving machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310710612.9A CN116777861B (en) | 2023-06-15 | 2023-06-15 | Marking quality detection method and system for laser engraving machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116777861A CN116777861A (en) | 2023-09-19 |
| CN116777861B true CN116777861B (en) | 2023-11-14 |
Family
ID=87988942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310710612.9A Active CN116777861B (en) | 2023-06-15 | 2023-06-15 | Marking quality detection method and system for laser engraving machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116777861B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117476509B (en) * | 2023-12-27 | 2024-03-19 | 联合富士半导体有限公司 | Laser engraving device for semiconductor chip product and control method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20110009564A (en) * | 2009-07-22 | 2011-01-28 | (주) 엔지온 | Method for manufacturing semiconductor chip with two sided sawing process of semiconductor wafer and an apparatus for manufacturing semiconductor chip by the method |
| CN110188843A (en) * | 2019-05-09 | 2019-08-30 | 上海赫丁格热处理有限公司 | Workpiece quality detection label record system |
| CN111652879A (en) * | 2020-07-01 | 2020-09-11 | 济南中正金码科技有限公司 | Marking character quality detection system and method |
| CN112733920A (en) * | 2020-12-31 | 2021-04-30 | 中国地质调查局成都地质调查中心 | Image identification method and system based on deep learning |
| CN114332536A (en) * | 2021-12-30 | 2022-04-12 | 广州大学 | Forged image detection method, system and storage medium based on posterior probability |
| CN114419397A (en) * | 2022-01-20 | 2022-04-29 | 中山大学·深圳 | Data set construction method and device based on data cleaning and data generation |
| CN114798495A (en) * | 2022-07-01 | 2022-07-29 | 深圳市信润富联数字科技有限公司 | Sheet metal stamping part sorting method, device, equipment and medium based on quality detection |
| CN115993344A (en) * | 2023-03-23 | 2023-04-21 | 苏州斌智科技有限公司 | Quality monitoring and analyzing system and method for near infrared spectrum analyzer |
-
2023
- 2023-06-15 CN CN202310710612.9A patent/CN116777861B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20110009564A (en) * | 2009-07-22 | 2011-01-28 | (주) 엔지온 | Method for manufacturing semiconductor chip with two sided sawing process of semiconductor wafer and an apparatus for manufacturing semiconductor chip by the method |
| CN110188843A (en) * | 2019-05-09 | 2019-08-30 | 上海赫丁格热处理有限公司 | Workpiece quality detection label record system |
| CN111652879A (en) * | 2020-07-01 | 2020-09-11 | 济南中正金码科技有限公司 | Marking character quality detection system and method |
| CN112733920A (en) * | 2020-12-31 | 2021-04-30 | 中国地质调查局成都地质调查中心 | Image identification method and system based on deep learning |
| CN114332536A (en) * | 2021-12-30 | 2022-04-12 | 广州大学 | Forged image detection method, system and storage medium based on posterior probability |
| CN114419397A (en) * | 2022-01-20 | 2022-04-29 | 中山大学·深圳 | Data set construction method and device based on data cleaning and data generation |
| CN114798495A (en) * | 2022-07-01 | 2022-07-29 | 深圳市信润富联数字科技有限公司 | Sheet metal stamping part sorting method, device, equipment and medium based on quality detection |
| CN115993344A (en) * | 2023-03-23 | 2023-04-21 | 苏州斌智科技有限公司 | Quality monitoring and analyzing system and method for near infrared spectrum analyzer |
Non-Patent Citations (4)
| Title |
|---|
| A digital blind watermarking for depth-image-based rendering 3D images;Lin Y H等;《IEEE transactions on Broadcasting》;第57卷(第2期);602-611 * |
| 一种低复杂度的激光打标拐点检测算法;梁文斌等;《激光技术》;第39卷(第2期);202-205 * |
| 激光打标机的标刻质量及评判标准的研究;唐海缤;《中国优秀硕士学位论文全文数据库 (工程科技Ⅰ辑)》(第5期);B022-247 * |
| 遥感图像区域变化检测技术的研究;张晖;《中国优秀硕士学位论文全文数据库 (信息科技辑)》(第4期);I140-677 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116777861A (en) | 2023-09-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7138753B2 (en) | Image defect detection method, device, electronic device, storage medium and product | |
| JP6573226B2 (en) | DATA GENERATION DEVICE, DATA GENERATION METHOD, AND DATA GENERATION PROGRAM | |
| KR102258942B1 (en) | System and method for the automatic determination of critical parametric electrical test parameters for inline yield monitoring | |
| CN108491302B (en) | Method for detecting spark cluster node state | |
| CN116777861B (en) | Marking quality detection method and system for laser engraving machine | |
| CN111353983A (en) | Defect detection and identification method and device, computer readable medium and electronic equipment | |
| CN110969600A (en) | Product defect detection method and device, electronic equipment and storage medium | |
| CN112926642B (en) | Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system | |
| CN109815855B (en) | Electronic equipment automatic test method and system based on machine learning | |
| CN109543693A (en) | Weak labeling data noise reduction method based on regularization label propagation | |
| CN113591948A (en) | Defect pattern recognition method and device, electronic equipment and storage medium | |
| CN113110961B (en) | Equipment abnormality detection method and device, computer equipment and readable storage medium | |
| CN117372424B (en) | Defect detection method, device, equipment and storage medium | |
| CN114169460A (en) | Sample screening method, sample screening device, computer equipment and storage medium | |
| JP7477608B2 (en) | Improving the accuracy of classification models | |
| CN109102486B (en) | Surface defect detection method and device based on machine learning | |
| CN111949530A (en) | Test result prediction method and device, computer equipment and storage medium | |
| Fakhri et al. | Road crack detection using gaussian/prewitt filter | |
| CN116843605A (en) | Fruit and vegetable defect detection method and system based on AI algorithm | |
| CN113139332A (en) | Automatic model construction method, device and equipment | |
| CN113537407B (en) | Image data evaluation processing method and device based on machine learning | |
| CN114530163A (en) | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering | |
| CN114359670A (en) | Unstructured data labeling method and device, computer equipment and storage medium | |
| JP7206892B2 (en) | Image inspection device, learning method for image inspection, and image inspection program | |
| CN114138328A (en) | Software reconstruction prediction method based on code peculiar smell |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |