CN115373345A

CN115373345A - Method and system for intelligent processing

Info

Publication number: CN115373345A
Application number: CN202110556050.8A
Authority: CN
Inventors: 陈文生; 李彦志
Original assignee: SYNPOWER CO Ltd
Current assignee: SYNPOWER CO Ltd
Priority date: 2021-05-21
Filing date: 2021-05-21
Publication date: 2022-11-22

Abstract

The present invention provides an intelligent processing method and system for integrating the on-line processing, detection and calibration of processing equipment. The system comprises a marking device, a rough judgment device, a fine judgment device, a processing information acquisition device and a measurement/parameter correction device. The marking device is used for marking the processing information on the processed object. The rough judgment device is an automatic vision detection device realized by an optical detection instrument, and can primarily screen the products with flaws. The precise determination device identifies the type of the defect through artificial intelligence. The processing information acquiring device is used for identifying the processing parameters of the processing information. The measuring/parameter correcting device can measure the size of the flaw and calculate the best result of adjusting the processing parameter according to the flaw size and the processing parameter.

Description

Method and system for intelligent processing

Technical Field

The present invention relates to an optimization of automatic processing flow, and more particularly to a method and system for intelligent processing.

Background

The production process is a loop-locked process from the processes of processing, detecting, and calibrating the processing machine, and any step error will eventually lead to the problems of insufficient product yield and reduced productivity. The above-mentioned manufacturing processes are often cut into different departments of responsibility, which is also prone to waste of time and cost in manual communication or coordination. For example: although the automatic optical inspection of the product can replace manual general inspection, when a defective product is found, manual re-inspection is still required to measure and identify the defect, and the defect data is then sent to the production unit, which then corrects the parameters of the production equipment. The above-mentioned process is not continuous and is useless in the day, and most cases can not obtain the ideal yield immediately by one-time calibration of the parameters of the machine, actually requiring many times of calibration.

With the development of technology, the term "industrial 4.0" was proposed in 2011, which is mainly intended to be intelligent and automatic in the production process and actively eliminate the problems during production.

Therefore, in order to solve the above problems, the art should develop a processing method and system conforming to the spirit of "industrial 4.0" to increase the efficiency of the production process and continuously increase the yield of the product.

Disclosure of Invention

In order to solve the above problems, embodiments of the present invention develop an intelligent processing method and system, which integrate product traceability, optical instrument initial inspection, artificial intelligence based review, virtual measurement, automatic parameter optimization, and other technologies. The applicant of the present invention refers to the patent cases filed in the past, including: TW109132726, TW109124011, TW110107077, TW109212512 and TW109136592 are incorporated by reference in the present invention.

Specifically, an embodiment of the present invention provides an intelligent processing system, which is applied to at least one processing device, and the processing device processes an object to be processed according to processing parameters and generates the object to be processed. The intelligent processing system comprises at least one marking device, a rough judgment device, a fine judgment device, a processing information acquisition device and a measurement/parameter correction device.

According to another embodiment, the marking device marks processing information on the object to be processed, wherein the processing information includes at least one of the processing parameters. The recording method of the processing information of the marking device includes characters, numbers, symbols and graphic codes, wherein the graphic codes include one-dimensional bar codes and two-dimensional bar codes. The processing information of the marking device further includes a manufacturing lot, a product model, and a manufacturing date.

According to another embodiment, the rough-decision device is an automatic visual inspection device for initially checking whether the workpiece has at least one defect and generating at least one workpiece image. The initial detection method of the automatic visual detection device comprises the steps of profile comparison, position coordinate comparison, 3D profile comparison, color comparison and brightness comparison. The automatic vision detection device comprises at least one image capturing unit, at least one distance measuring unit, a calculating unit and a database. The distance measuring unit is used for generating an image-capturing distance information for capturing the processed object. The database stores a plurality of standard dimension information, each of which includes a resolution per unit area, an image capture distance, a corresponding pixel matrix, and a corresponding actual dimension and a film thickness height. The calculating unit compares the standard size information in the database according to the resolution information of the processed object image and the image-capturing distance information, and calculates and generates the processed object image with actual size information. The standard size information in the database may be, for example, a two-dimensional size or a three-dimensional size, and the calculation unit may further calculate the actual size having the two-dimensional size or the three-dimensional size. The processed object image may be 2D or 3D.

According to another embodiment, the fine determination device is in signal connection with the rough determination device for rechecking whether the image of the workpiece has at least one defect, and identifying the type of the defect contained in the image of the workpiece, and generating the identification result. The precise judging device further comprises a classification module, wherein the classification module is used for identifying the defect type by applying an artificial neural network. The training data of the classification module includes the recognition result and the manual re-judgment result.

According to another embodiment, the processing information acquiring device is connected to the precise judging device via signals, and is used to identify the processing information of the processed object image and generate the processing parameters.

According to another embodiment, the measuring/parameter calibrating device is connected to the fine-judging device by signals, and is used for measuring the defect position of the image of the workpiece to generate a defect size, calculating the best result of the processing parameter according to the defect size and the processing parameter, and further generating a calibrated processing parameter, and transmitting the calibrated processing parameter to the processing device. The measurement/parameter calibration apparatus further includes a parameter optimization module, wherein the parameter optimization module calculates the optimal result of the processing parameter by applying an algorithm. The flaw size may be, for example, a two-dimensional size or a three-dimensional size.

According to another embodiment, the measurement/parameter calibration apparatus further superimposes the design drawing on the image of the workpiece to generate a superimposed image for measuring the defect size when the workpiece has the design drawing.

According to another embodiment, the measuring/parameter calibrating apparatus further includes a step of providing at least one monitoring measuring area preset in the design drawing, and measuring the monitoring measuring area of the superimposed image to generate monitoring dimension information. The monitoring measurement Region can be regarded as a Region of interest (ROI) or an important Region (ROI for short) in image processing.

According to another embodiment, the measurement/parameter calibration apparatus may further generate monitoring statistical information by performing statistics on the monitoring dimension information, and the monitoring statistical information is used as calculation data for calculating the processing parameter. The monitor size information may be information of a two-dimensional size or a three-dimensional size, for example.

According to another embodiment, when the machining is performed according to the machining parameters, the number of defects found by the precise judgment device can be used to further calculate the production yield corresponding to the machining parameters.

According to another embodiment, the qualification criteria of the production yield can be adjusted according to user requirements, and if the yield of the process parameters meets the qualification criteria, the process parameters can be further used as reference parameters for performing Virtual Metrology (VM). The virtual measurement refers to estimating the production result (or quality) through the processing parameters to achieve the purpose of full inspection.

According to another embodiment, the measurement/parameter calibration device stops automatically calibrating the processing parameters and generates an alarm notification when the defect size exceeds a predetermined threshold.

According to another embodiment, the precise determination device receives the warning notification and does not use the recognition result corresponding to the warning notification as the training data for training the classification module.

According to another embodiment, the alert notification can notify the system administrator in real time through a short message or a mail.

The embodiment of the invention further provides an intelligent processing method, which comprises the following steps. Defaulting processing parameters in a processing device; processing the object to be processed and generating a processed object; marking processing information on the processed object, wherein the processing information comprises the processing parameters; using a rough judgment device to perform a preliminary inspection on the processed object, and capturing at least one processed object image and image-capturing distance information, wherein the rough judgment device can be an automatic vision detection device; comparing the image-capturing distance information with a plurality of standard size information loaded by the rough-judging device to generate the actual size information of the processed object image; rechecking by a precise judging device, identifying the defect type contained in the processed object image by using a classification module, and generating an identification result, wherein the classification module executes the identification of the defect type by using an artificial neural network; superposing the design drawing of the processed object with the processed object image to generate a superposed image, and measuring the flaw position of the superposed image to generate flaw size; identifying the processing information to generate the processing parameters; the measuring/parameter correcting device calculates the best result of the processing parameter according to the flaw size and the processing parameter through a parameter optimizing module, and generates a corrected processing parameter to be transmitted to the processing device, wherein the parameter optimizing module executes the processing parameter optimization by applying an artificial neural network.

According to another embodiment, the method further comprises a step of quality monitoring, wherein at least one monitoring measurement area in the design drawing of the workpiece is set in advance through the measurement/parameter calibration device, and the monitoring measurement area of the superimposed image is measured to generate monitoring dimension information.

According to another embodiment, the measurement/parameter calibration apparatus according to the method further generates monitor statistical information by counting the monitor dimension information, and the monitor statistical information is used as calculation data for calculating the processing parameters.

According to another embodiment, the method further calculates the yield corresponding to the processing parameters according to the number of defects found by the precise determination device during processing according to the processing parameters.

According to another embodiment, the qualification criteria of the production yield according to the method may be adjusted according to user requirements, and if the yield generated by the processing parameters meets the qualification criteria, the processing parameters may be further used as reference parameters for performing Virtual Metrology (VM). The virtual measurement refers to estimating the production result (or quality) through the processing parameters to achieve the purpose of full inspection.

According to another embodiment, the training data of the classification module according to the method includes the recognition result and the manual re-judgment result.

According to another embodiment, when the defect size exceeds the threshold value according to the method, the measurement/parameter calibration apparatus stops automatically calibrating the processing parameter and generates an alarm notification.

According to another embodiment, the precise determination device according to the method receives the warning notification, and does not use the recognition result corresponding to the warning notification as the training data for training the classification module.

According to another embodiment, the preliminary detection method of the rough judgment device according to the above method includes 3D contour comparison, position coordinate comparison, color comparison and brightness comparison.

According to another embodiment, the standard dimension information includes a resolution per unit area, a capturing distance, a corresponding pixel matrix, and a corresponding actual dimension and a film thickness height.

According to another embodiment, the recording method of the processing information according to the method includes characters, numbers, symbols and graphic codes, wherein the graphic codes include one-dimensional bar codes and two-dimensional bar codes.

In combination with the above-mentioned technical features of the embodiments, the following effects can be specifically claimed.

(1) The system is applied to the processing procedure to achieve the effects of intellectualization and full automation, and can be connected in series with the processing equipment of a plurality of processing stations simultaneously to automatically execute the initial judgment and the repeated judgment detection of the processed object of the processing equipment of each station.

(2) The products processed by the system can trace the processing parameters of each station in real time by the processing information on the products, and can quickly clear which stations cause the product yield deficiency.

(3) The system can automatically classify the flaw types and measure the flaws through the artificial neural network according to the initial judgment and the repeated judgment detection results, and acquire flaw size information.

(4) After the system collects the processing parameters and the flaw sizes, the optimal result of the processing parameters can be further calculated through an algorithm, corrected processing parameters are generated, and the corrected processing parameters are transmitted back to the processing equipment, so that the effect of optimizing the processing parameters in real time is achieved.

Drawings

In order to make the aforementioned and other objects, features, advantages and embodiments of the invention more comprehensible, the following description is given:

FIG. 1 is a schematic diagram of an apparatus of an intelligent processing system according to an embodiment of the invention.

Fig. 2 is a schematic diagram of an apparatus for applying an intelligent processing system to a plurality of processing tools according to an embodiment of the present invention.

FIG. 3 is a flow chart of an intelligent processing system according to an embodiment of the invention.

FIG. 4 is a flow chart illustrating quality monitoring of an intelligent processing system according to an embodiment of the present invention.

Detailed Description

To more particularly describe the embodiments of the present invention, reference is made to the following description taken in conjunction with the accompanying drawings.

Referring to fig. 1, fig. 1 is a schematic diagram illustrating an apparatus of an intelligent processing system according to an embodiment of the invention. In fig. 1, an intelligent processing system 100 is provided, according to one embodiment. The intelligent processing system 100 is applied to at least one processing device 110, the processing device 110 processes the object to be processed 200 according to the processing parameters and generates the object to be processed 220, the system 100 includes: a marking device 120, a rough judgment device 130, a fine judgment device 140, a processing information acquisition device 160, and a measurement/parameter calibration device 180. Each of the components of the intelligent processing system 100 may be implemented through software plus hardware or through pure hardware, and the invention is not limited thereto.

The marking device 120 marks processing information on the object 220, wherein the processing information includes at least one of the processing parameters. The recording method of the processing information of the marking device 120 includes characters, numbers, symbols and graphic codes, wherein the graphic codes include one-dimensional bar codes and two-dimensional bar codes. The processing information of the marking device 120 further includes a manufacturing lot, a product model, and a manufacturing date. For example: the two-dimensional graph code is printed on the printed circuit board finished product or semi-finished product, and the historical processing parameter record can be generated by scanning or identifying the two-dimensional graph code at a later date.

The rough-judging device 130 is used to initially detect whether the workpiece 220 has at least one defect and generate at least one workpiece image, wherein the rough-judging device is an automatic visual inspection device. The initial inspection method of the automatic visual inspection device comprises 3D contour comparison, position coordinate comparison, color comparison and brightness comparison. The rough judgment device 130 of the present system is usually strict in the initial judgment standard to avoid neglecting the product with the real defect, but there is a case of misjudgment, for example, a fine dust appears at a certain position in the printed circuit board, which is judged as a defect, so that the yield of the product can be ensured, and the false defect can be relatively easily judged. In the past, the above-mentioned false defects can be eliminated only by manual re-judgment, but in the present invention, the re-judgment and defect classification of defects can be performed by the above-mentioned precise judgment device 140, so that the number of false defects and the number of manual re-judgment needed can be greatly reduced.

According to another embodiment of the present invention, the automatic vision inspection apparatus comprises at least one image capturing unit, at least one distance measuring unit, a calculating unit and a database. The distance measuring unit is used for generating an image-capturing distance information for capturing the processed object. The database stores a plurality of standard dimension information, each of which includes a resolution per unit area, an image capture distance, a corresponding pixel matrix, and a corresponding actual dimension and a film thickness height. The calculating unit compares the standard size information in the database according to the resolution information of the image of the workpiece 220 and the image-capturing distance information, and calculates and generates the workpiece image with actual size information. The rough judgment device 130 can be, for example, a calculator (including a CPU and a GPU) in cooperation with an image recognition device.

The fine determination device 140, which is connected to the rough determination device 130 via signals, is used to re-check whether the processed object image has at least one defect, and identify the defect type contained in the processed object image, and generate an identification result. The precise determination device further includes a classification module 142, and the classification module 142 is implemented by applying an artificial neural network to identify the defect type. The training data of the classification module 142 includes the recognition result and the manual re-judgment result. The artificial Neural Network applied by the classification module 142 includes a Convolutional Neural Network (CNN) model. The CNN model can be selected from one of R-CNN (Region-based connected Neural Network), fast R-CNN, RPN (Region pro-potential Network), mask R-CNN and FCN (full connected Network) for defect identification and classification. In addition, besides the implementation of the convolution neural network model, other precise algorithms for classifying the components in the image can be used to implement the precise determination device 140, and the implementation manner of the precise determination device 140 is not limited in the present invention. The precise determination device 140 can effectively reduce the number of false defects and the number of manual re-determination required, and whether the manual re-determination is required can be determined by the default threshold of the measurement/parameter calibration device 180. The above false defect may be, for example, a printed circuit board, and may include: false points on the board edge, fine board scraps, dust and the daughter board to be tested are different from the mother board. The precise judgment device 140 can be, for example, a calculator (with CPU) having computing capability.

The processing information obtaining device 160 is connected to the precise determination device 140 by signal connection, and is used to identify the processing information of the processed object image and generate the processing parameters. The identification method can be character identification, image code identification, RFID and other induction tags. The processing information acquiring device 160 may be, for example, a calculator device (including a CPU and a GPU) in cooperation with an image recognition device.

The measurement/parameter calibration device 180, which is connected to the fine determination device 140 via signals, measures the defect position of the image of the workpiece to generate a defect size, calculates the optimal result of the processing parameter according to the defect size and the processing parameter to generate a calibrated processing parameter, and transmits the calibrated processing parameter to the processing device 110. The metrology/parameter calibration apparatus 180 further includes a parameter optimization module 182, wherein the parameter optimization module 182 calculates the optimal result of the processing parameters by applying an algorithm. The algorithm applied by the parameter optimization module 182 may be any algorithm that can calculate or generalize the optimal result of the processing parameters, such as: gradient descent method, newton method, conjugate gradient method, linear search, confidence domain method, neural network, particle swarm algorithm, simulated annealing, support vector machine, ant swarm algorithm, differential evolution algorithm, K-nearest neighbor algorithm (K-nearest neighbor). The measurement/parameter calibration device 180 can be, for example, a calculator (with a CPU) with computing capability.

According to another embodiment of the present invention, when the object 220 has a design drawing, the measurement/parameter calibration device 180 can further superimpose the design drawing on the image of the object 220 to be processed, which can be used as a basis for measuring the size of the defect.

According to another embodiment of the present invention, the measurement/parameter calibration device 180 further comprises a quality monitoring process, which sets at least one monitoring measurement area in the design drawing of the workpiece in advance through the measurement/parameter calibration device, and measures the monitoring measurement area of the superimposed image to generate monitoring dimension information. Wherein the quality monitoring process can be performed by general inspection or spot inspection. Taking the detection of the printed circuit board as an example, the monitoring and measuring area can be selected by frames according to a specific welding position of the printed circuit board.

According to another embodiment of the present invention, the monitoring dimension information can be further statistically generated by the measurement/parameter calibration device 180 to generate monitoring statistical information. And the monitoring statistic information is used as the data for calculating the processing parameters.

According to another embodiment of the present invention, when the defect size exceeds a threshold, the measurement/parameter calibration device 180 stops automatically calibrating the processing parameters and generates an alarm notification.

According to another embodiment, when the precise judgment device 140 receives the alert notification, the recognition result corresponding to the alert notification will not be used as the training data for training the classification module.

According to another embodiment, the warning notification can notify the system administrator in real time through short message or mail, and remind the system administrator that the identification result corresponding to the warning notification will need to be manually determined again. For example: it was found that the area of exposed nickel on the pcb was greater than 1.0cm2, and if the magnitude of the above-mentioned processing parameters could be adjusted too much according to the size of the defect, the subsequent processing flow would be seriously affected. Therefore, the size of the defect exceeding the threshold value is not used as the data source for adjusting the processing parameters and the training data.

According to another embodiment, the present invention can be applied to various stages of a manufacturing process, and referring to fig. 2, fig. 2 is a schematic diagram of an apparatus for applying an intelligent processing system to a plurality of processing equipments according to an embodiment of the present invention.

Referring to fig. 3, fig. 3 is a flow chart illustrating an intelligent processing system according to an embodiment of the invention.

In fig. 3, step 300 is to begin processing.

In step 302, the processing parameters of the processing device 110 are set.

In step 304, the machining device 110 performs a machining operation to generate the workpiece 220.

In step 306, the processed object 220 is marked with processing information. The recording method of the processing information includes characters, numbers, symbols and graphic codes, wherein the graphic codes include one-dimensional bar codes and two-dimensional bar codes.

In step 308, the object 220 is determined by the rough determination device 130, and if the object 220 has at least one defect, the image of the object 220 is captured to generate at least one object image. The rough judgment device 130 can further calculate the actual size information of the processed object image, and the calculation method is as described above and will not be described herein again.

In step 310, the rough judgment device 130 makes an initial judgment that the workpiece 220 has at least one defective portion and judges the workpiece as a defective object, and then the process proceeds to step 312. If the rough determination device 130 initially determines that no defect is found, go to step 322.

In step 312, the processed object image is determined again by the precise determination device 140, and the defect type is identified.

In step 314, after the fine determination device 140 determines that the processed object image has at least one defect, the type of the defect is further identified, and the method for identifying the defect is as described above and will not be described herein again. If the fine judgment means 140 judges that the processed object image has no defect, the process goes to step 322 to indicate that the defect part judged by the rough judgment means 130 is a false defect.

In step 316a, the defect position is measured by the measurement/parameter calibration device 180 to generate the defect size, and the measurement method is as described above and will not be described herein again.

In step 316b, the processing information is identified and the processing parameters are generated by the processing information obtaining device 160, and the method for identifying the processing information is as described above and will not be described herein again.

In step 318, the algorithm of the measurement/parameter calibration device 180 calculates the best result of calibrating the processing parameters according to the defect size and the processing parameters, generates calibration processing parameters, and transmits the calibration processing parameters to the processing device.

In step 320, the processing device 110 adjusts the original processing parameters according to the corrected processing parameters, and jumps back to step 304 to continue processing.

In step 322, the workpiece 220 is continuously processed because there is at least one defect in the initial judgment of the rough judgment device 130 or the precise judgment of the precise judgment device 140.

Referring to fig. 4, in accordance with another embodiment of the present invention, fig. 4 is a flow chart illustrating a quality monitoring process of an intelligent processing system according to an embodiment of the present invention.

The process flow of the embodiment of fig. 3 of the present invention can be combined with the quality monitoring process flow of the embodiment of fig. 4 to increase the accuracy of calculating the above-mentioned correction processing parameters, wherein the process flows of the embodiments of fig. 3 and fig. 4 can be executed synchronously or separately.

In step 402, a measurement area to be monitored in the design drawing of the workpiece 220 is set (or selected) in the measurement/parameter calibration device 180 in advance.

In step 404, an image of the workpiece 220 is captured by the rough determination device 130.

In step 406, the design drawing of the workpiece 220 is superimposed with the processed image to generate a superimposed image.

In step 408, the size of the monitoring measurement area set in advance is measured on the superimposed image.

In step 410, monitor size information is generated according to the measurement result, and monitor statistical information is further calculated.

In step 412, the monitored statistical information may be used as data for the metrology/parameter calibration device 180 to calculate calibration process parameters.

According to another embodiment of the present invention, in step 414, the processing parameters adjusted in steps 402-412 are used to further calculate the production yield corresponding to the processing parameters according to the number of defects found in steps 300-314. And can inspect whether the production yield meets the qualified standard (for example, the yield needs to reach 95.0%)

In step 416, the process parameters may be further measured by Virtual Metrology (VM) to increase the production yield. For example: through the virtual measurement calculation, in the process of laying a metal copper layer on a substrate when the printed circuit board is manufactured, the electroplating time of the copper sulfate bath is reduced by 0.5 second, so that the occurrence probability of short circuit of the circuit can be reduced by 1%. Therefore, the production yield can be further increased by adjusting the production parameters through the virtual measurement. By the virtual measurement, the monitored measurement area is measured without performing the steps 402-412, and instead, the yield of the product is calculated by the virtual measurement, so as to achieve the purpose of full inspection.

In summary, the embodiment of the invention provides an intelligent processing method and system, which can process and detect the production line in real time. And further, tracing, flaw judgment and correction of the processing equipment are automatically performed.

While the present invention has been described in terms of its preferred embodiments, it will be understood by those skilled in the art that the foregoing embodiments are provided for illustrative purposes only, and are not intended to limit the scope of the invention. All changes and substitutions that are equivalent or equivalent to the above-described embodiments should be understood to be included within the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be defined by the following claims.

Description of the symbols

100: intelligent processing system, 110: machining device, 120: marking device, 130: rough judgment device, 140: precise judgment device, 142: classification module, 160: processing information acquisition device, 180: measurement/parameter calibration device, 182: parameter optimization module, 200: object to be processed, 220: processed object, 300-322: step 402-416: and (5) carrying out the following steps.

Claims

1. A method of intelligent processing, the method comprising the steps of:

defaulting a processing parameter in a processing device;

processing an object to be processed and generating a processed object;

marking a piece of processing information on the processed object, wherein the processing information comprises the processing parameter;

using a rough-judging device to perform a preliminary inspection on the processed object and capturing at least one processed object image and an image-capturing distance information, wherein the rough-judging device is an automatic vision inspection device;

comparing the image-capturing distance information with a plurality of standard size information loaded by the rough-judging device to generate the actual size information of the processed object image;

rechecking by a precise judging device, identifying the defect type contained in the processed object image by using a classification module, and generating an identification result, wherein the classification module executes the identification of the defect type by using an artificial neural network;

superposing the design drawing of the processed object with the processed object image to generate a superposed image, measuring the flaw position of the superposed image and generating a flaw size;

identifying the processing information to generate the processing parameter; and

a measurement/parameter correction device calculates the best result of the processing parameter according to the flaw size and the processing parameter through a parameter optimization module, and generates a corrected processing parameter to be transmitted to the processing device, wherein the parameter optimization module executes processing parameter optimization by applying an algorithm.

2. An intelligent processing method as claimed in claim 1, wherein the method further comprises a quality monitoring step, which sets at least one monitoring measurement region in the design drawing of the processed object through the measurement/parameter calibration device, and then measures the monitoring measurement region of the superimposed image to generate at least one monitoring dimension information.

3. The intelligent processing method as claimed in claim 1 or claim 2, wherein the measurement/parameter calibration device further comprises generating a monitoring statistic by counting the monitoring dimension information, and the monitoring statistic is used as the calculation data for calculating the processing parameter.

4. An intelligent processing method as claimed in claim 3, wherein the processing parameter calculated by the monitoring statistic information is used as a reference parameter for virtual measurement to execute the virtual measurement.

5. The intelligent processing method as claimed in claim 1, wherein the training data of the classification module includes the recognition result and the manual review result.

6. The intelligent processing method as claimed in claim 1, wherein when the defect size exceeds a threshold, the measurement/parameter calibration device stops automatically calibrating the processing parameters and generates an alarm notification.

7. The intelligent processing method as claimed in any one of claim 1 or claim 6, wherein the precise determination device receives the warning notification and does not use the recognition result corresponding to the warning notification as the training data for training the classification module.

8. The intelligent processing method as claimed in claim 1, wherein the preliminary detection method of the rough judgment device comprises 3D contour comparison, position coordinate comparison, color comparison and brightness comparison.

9. The intelligent processing method as claimed in claim 1, wherein the standard dimension information respectively includes resolution per unit area, image capture distance, corresponding pixel matrix, and corresponding actual dimension and film thickness height.

10. The intelligent processing method as claimed in claim 1, wherein the recording manner of the processing information comprises characters, numbers, symbols and graphic codes, wherein the graphic codes comprise one-dimensional bar codes and two-dimensional bar codes.

11. An intelligent processing system, the system is applied to at least one processing device, the processing device processes an object to be processed according to a processing parameter, and generates an object to be processed, the system includes:

at least one marking device for marking a processing information on the processed object, wherein the processing information comprises at least one processing parameter;

a rough-judging device for initially checking whether the processed object has at least one defect and generating at least one processed object image, wherein the rough-judging device is an automatic visual inspection device;

a fine judgment device, which is connected with the rough judgment device by signals and is used for rechecking whether the processed object image has at least one defect and identifying the defect type contained in the processed object image to generate an identification result;

a processing information acquisition device, signal connected to the precise judgment device, for identifying the processing information of the processed object image and generating the processing parameters; and

a measuring/parameter correcting device, signal connected to the fine judging device, for measuring the defect position of the processed object image and generating a defect size, then calculating the best result of the processing parameter according to the defect size and the processing parameter, generating a corrected processing parameter, and transmitting the corrected processing parameter to the processing device.

12. The intelligent processing system as claimed in claim 11, wherein the automatic vision inspection device comprises an outline matching, a position coordinate matching, a color matching and a brightness matching.

13. The intelligent processing system as claimed in claim 11, wherein the automatic vision inspection device comprises at least one image capturing unit, at least one distance measuring unit, a computing unit and a database.

14. The intelligent processing system as claimed in claim 11, wherein the distance measuring unit is used to generate an image-capturing distance information for capturing the processed object.

15. The intelligent processing system as claimed in claim 11, wherein the database stores a plurality of standard dimension information, each of the standard dimension information comprises a resolution per unit area, a capture distance, a corresponding pixel matrix, and a corresponding actual dimension and a film thickness.

16. The intelligent processing system as claimed in any one of claims 11 to 15, wherein the calculating unit compares the standard dimension information in the database with the resolution information of the processed object image and the image-capturing distance information to calculate and generate the processed object image with actual dimension information.

17. The intelligent processing system of claim 11 wherein said metrology/parameter calibration device further comprises a parameter optimization module, said parameter optimization module calculating optimal results for said processing parameters by applying an algorithm.

18. The intelligent processing system as claimed in claim 11, wherein the measurement/parameter calibration device, when the object has a design drawing, further superimposes the design drawing with the image of the object to be processed to generate a superimposed image for use as a basis for measuring the defect size.

19. The intelligent processing system as claimed in claim 11 or claim 18, wherein the measurement/parameter calibration device further comprises at least one monitoring measurement region preset in the design drawing, and measures the monitoring measurement region of the overlay image to generate at least one monitoring dimension information.

20. The intelligent processing system as recited in claim 19, wherein said metrology/parameter calibration device further comprises statistical generation of a monitoring statistic from said monitored dimension information, and said monitoring statistic is used as calculation data for calculating said processing parameter.

21. The intelligent processing system as claimed in claim 20, wherein the processing parameters calculated by the monitoring statistics are used as reference parameters for virtual metrology to perform the virtual metrology.

22. The intelligent processing system as claimed in claim 11, wherein the processing information of the marking device is recorded by text, numbers, symbols and graphic codes, wherein the graphic codes include one-dimensional bar codes and two-dimensional bar codes.

23. The intelligent processing system as claimed in claim 11, wherein the processing information of the indicating device further comprises a manufacturing lot, a product model, and a manufacturing date.

24. The intelligent processing system of claim 11, wherein the refining device further comprises a classification module for identifying the defect type by applying an artificial neural network.

25. The intelligent processing system as claimed in claim 24, wherein the training data of the classification module comprises the recognition result and the manual review result.

26. The intelligent processing system of claim 11, wherein the metrology/parameter calibration device stops automatically calibrating the processing parameters and generates an alert when the defect size exceeds a threshold.

27. The intelligent processing system as claimed in claim 11 or claim 26, wherein the refining device receives the alert notification and does not use the recognition result corresponding to the alert notification as the training data for training the classification module.