CN101292263A - Method and system for automatic defect detection of articles in visual inspection machines - Google Patents
Method and system for automatic defect detection of articles in visual inspection machines Download PDFInfo
- Publication number
- CN101292263A CN101292263A CNA2006800384645A CN200680038464A CN101292263A CN 101292263 A CN101292263 A CN 101292263A CN A2006800384645 A CNA2006800384645 A CN A2006800384645A CN 200680038464 A CN200680038464 A CN 200680038464A CN 101292263 A CN101292263 A CN 101292263A
- Authority
- CN
- China
- Prior art keywords
- defect
- parameter
- article
- parameter setting
- defective
- 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.)
- Pending
Links
Images
Abstract
There is provided a method for establishing a parameters setup for inspecting a plurality of articles by an automatic inspection system. The method includes inspecting a first article by the inspection system, applying an automatic defects detection method according to a given set of inspection parameters, receiving an initial map of defects and sorting uncovered defects into defect types according to a predetermined set of defect types. While sorting defects, if new defects not recognized by the inspection system are detected, adding the new defects to the initial map to be sorted and automatically setting the inspection parameters by means of applying computational dedicated algorithms, using a heuristic approach, to form a modified parameters setup. The modified parameters setup is then used for obtaining a modified map of detected defects, and the modified parameters setup for inspecting other of the plurality of articles. A system for establishing a parameters setup for inspecting a plurality of articles is also provided.
Description
Technical field
The present invention relates in the intermediate process steps during the duplication of production article method that in visual inspection system automatically, realizes, more particularly, relate to and carry out the method that inspection parameter is set when detecting defective by the automatic gauging instrument.
Background technology
In comprising the article production run of a plurality of continuous treatment steps, as P.e.c., semiconductor equipment or complex mechanical, managing throughout needs check, verification and quality control step between the step.In order to detect defective article and to avoid carrying out invalid, expensive treatment step, need carry out centre verification for badly damaged article in may some treatment steps formerly.In some cases, can only after finishing whole process of production, just carry out the functional test of article.For this reason, the visual inspection methods in the middle of having developed, it begins to carry out with manual visual inspection machines, and for example U.S. Patent No. 4691426 is described.
Another aspect that relates to automatic processing medium quantity batch processing is the degree of correlation of the physical location (coordinate) on defect type and the article.Some automatic sight check (AVI) technology are by being stored in detected defect coordinate in the database of continuous updating, thereby utilize this correlativity.This database is used to shorten and improve the round of visits of the article that will detect.
In this type systematic, whether serious in order to assess the defective that is detected, need finish several at least article and carry out functional test.As long as on same position, repeat to produce serious defective, this method is exactly acceptable, but if at random or new local defect, just should start the processing procedure of the assessment defective order of severity once more, and generate defective to having formed unacceptable delay between the detection automatically.
The second method that is used for visual inspection system is by the image that obtains the article that are verified and analyzes this image and carry out defects detection.This analysis normally utilizes Flame Image Process, form and pattern recognition device to carry out.Each device all has the intrinsic parameter of himself, and these parameters define the defective that this system can discern.When discerning,, need a group categories rule in order to determine whether that be major defect to suspicious defect report.Subsequently can be by the defective that this system is reported by user or automatically visual and/or reparation.But the accurate differentiation between major defect and the non-major defect is also remarkable, in order accurately to distinguish major defect, uses this system of sample training.For the dissimilar defective that generates in the production run, be difficult to provide accurate major defect sample by manual observation and classification.
In addition, the different user of this system has different examination criterias for their different product.A client thinks that the certain patterns of major defect may be an acceptable for another client.In addition, the defective that should report in the sophisticated product may show as the acceptable quality for the not high product of same client's measure of precision.And the different article of identical product have different demonstrations in the image that is obtained, and therefore need different classifying rules parameter group.
The third method has been used the combination of said method, proposes in U.S. Patent No. 7062081, and a kind of method that is used for the detected defective of production run of analytical electron circuit patterns wherein is provided.Defective on the object that is verified is detected, and the positional information of this detected defective is stored.The specifying information of this defective is gathered at this defective, for it stores the positional information of this defective.The specifying information that is collected is associated with defective locations information and is stored.The object that is verified is tested by the electronics mode, and is stored in the information that occurs wrong position in this Electronic Testing.Defective locations information of having stored and the wrong positional information that generates compare, and the result classifies detected defective based on the comparison.Show the relevant information of defective subsequently with the process classification.
The shortcoming of above-mentioned the third method comprises: need come verification information according to functional test results, and be difficult to set up the classifying rules of the other products except that semiconductor equipment, and compare with above-mentioned second method, there is wideer image difference in it.Another defective is, is difficult to manually upgrade classifying rules, below with explained in detail.
The importance of these all major defects of System Reports is clearly, but is to use too sensitive classifying rules parameter group also will cause reporting non-major defect.The defective of uncared-for defective or wrong identification can be not intended to the free burial ground for the destitute at last and consume customer resources.
U.S. Patent No. 6674888 is related to the process that classifying rules is provided with parameter, and suggestion repeats regulation rule, until satisfying resulting standard.
As described in above-mentioned patent documentation, in order to satisfy the particular detection standard of each product, obtain the optimum balance between major defect and the non-major defect simultaneously, when setting above-mentioned identification, judgement and Report Parameters, have continuous stages.Initial setting can automatically perform according to the specific characteristic of product, but this setting can not obtain the optimum balance between major defect and the non-major defect all the time.This situation is owing to different reasons causes, and comprising: a) characteristic of the image that is obtained by the article checked can not be predicted all the time in advance, and b) there is unpredictable environmental baseline, as dust granule, lighting condition or material behavior.
Because the testing result behind the initial setting of first article will be applied on all subsequent article in this batch subsequently, a follow-up more completely assignment procedure is employed, to optimize the setting of first article.
At present, at first article of scanning and after receiving initial defect map will, this system itself is carried out follow-up assignment procedure.But by direct change identification or judgement and Report Parameters or by changing the examination criteria that can influence these parameters subsequently, this secondary setup is manually to carry out.
As long as assignment procedure is manually to carry out, just define the quantity of the parameter that can change by common user, its result depends on specific user's skill and the user familiarity to checking system to a great extent.
Therefore, need a kind of automatic intelligent method, be used for setting, the classifying rules of refinement and the automatic sight check process of adjustment.
Summary of the invention
Therefore a kind of method and system has been proposed, correlativity between wherein very big processing, identification, judgement and the Report Parameters group can be in very short setting-up time, and before the checkout procedure, during or under the indicated afterwards restriction by Automatic Optimal concurrently.Described optimizing process is based on arithmetic or cost function minimizing scheme, and it uses the heuristic or learning parameter of logical OR of decision rule.Described optimizing process is also handled the hierarchical structure of image space and color depth resolution, and focuses on different image sources, as imaging sensor, light source, storage source and network source.Described optimizing process can also not carry out user interactions at specific learning process (not finishing automatically), comprises specific visual and judgment device.
The present invention also provides a kind of method, and it utilizes semi-automatic or full-automatic machine study notion, is convenient to the secondary setup in the automatic visual inspection system, thereby has strengthened testing result, makes unfamiliar user also can operate native system.
Described method is, in case (first article preferably of article in this batch, but this is not absolute) be verified, then set up the initial defect map will of being reported, defective is come sorting according to the order of severity, can adjust identification, judgement and Report Parameters automatically afterwards, thereby satisfy by the determined detection criteria of selection process best.
Alternatively,, can carry out learning process once more by carrying out selection from resulting additional defect figure that the subsequent article this batch is tested, thus further refinement parameter adjustment, and further strengthen testing result.In addition, also provide a kind of method of carrying out this assignment procedure from remote location.
According to a preferred embodiment of the present invention, a kind of method of the parameter setting when being used for determining to check a plurality of article by automatic checking system is provided, described method comprises the following step: check first article by described checking system, according to one group of given inspection parameter use a kind of automatic defect method of inspection, obtain an initial defect map will, according to one group of predetermined defect type the defective of being found is carried out sorting according to defect type, when defective is carried out sorting, if detect the new defective that described checking system can not be discerned, then described new defective is added in the described initial defect map will that is classified selection, calculate tailor-made algorithm by using, use heuristic, described inspection parameter is set automatically, to form adjusted parameter setting, use adjusted parameter setting to obtain the adjusted defect map of detected defective, and use described adjusted parameter setting to check all the other article in described a plurality of article.
The present invention also provides a kind of system of the parameter setting when being used for determining a plurality of article of check, described system comprises a checking system, be used for checking first article of one batch, to form an initial defect map will, also comprise a controller, it is operated and is used for receiving described initial defect map will from described checking system, each the described defective that shows described initial graph, make the operator to carry out sorting to each defective according to defect type, and a selection result is input in this system, use tailor-made algorithm to determine adjusted parameter setting, to obtain adjusted defect map, described adjusted defect map has the desirable ratio between true defect and the false defect, and provides described parameter setting to check other article in described batch.
The present invention also provides a kind of system of the parameter setting when automatically or semi-automatically being identified for checking a plurality of article, and described system comprises a sensor, is used for imaging is carried out in the zone of the article that are verified; One testing agency is used to select to be used on the described article position of handling or showing; One storer can be preserved the image of the Examination region that is obtained by described sensor; One decision unit is used to obtain best defect map; One searches mechanism, is used to seek the best output result's that can produce the decision unit defined parameter value; And provide described parameter value to check the device of all the other article in described batch.
Therefore, different with method and system of the prior art, according to the present invention, parameter setting not only can control chart as processing parameter, and all parameters that can control system are as the illumination intensity of article.
Description of drawings
It specifically describes the present invention with reference to schematic figures and in conjunction with specific preferred embodiment below, so that can be understood more fully.
Specify in conjunction with the accompanying drawings now, it is emphasized that, shown in content only be to describe by example, its objective is and schematically explain the preferred embodiments of the present invention that it provides thinks the most useful at present and be convenient to understand specifying of principle of the present invention and notion.Thus, the explanation of CONSTRUCTED SPECIFICATION of the present invention can be more specifically more not required for the present invention than basic comprehension, to those skilled in the art, can very clearly obtain several form that can implement in practice of the present invention according to instructions and accompanying drawing.
Among the figure:
Fig. 1 is the process flow diagram that the detected parameters of automatic visual inspection system is carried out the method for semi-automatic adjustment;
Fig. 2 obtains and is stored in the example that image in the storer carries out sorting in the rudimentary check process;
Fig. 3 shows the example that utilizes the sorting that live video obtains;
Fig. 4 shows best identified/report of optional method select a to(for) parameter, and
Fig. 5 shows best identified/report of another kind of optional method select a to(for) parameter.
Embodiment
With reference to the accompanying drawings, Fig. 1 shows the process flow diagram that the detected parameters in the automatic visual inspection system is carried out the method for semi-automatic adjustment.Whether this method is considered to automanual, seriously manually carried out by the user because judge the defective in the initial defect map will, preferably by veteran user, carry out as the deviser of article, is perhaps automatically performed by system.The flow process of treatment step is sorted automatically by controller.
In the step (a) of module 11, first article in batch or other article use initial parameter scanning to test by automated optical inspection (AOI) system.These initial parameters can obtain automatically by initial setting in the system or default value, perhaps manually select from parameter database.Preferably, select responsive parameter group, this will cause being positioned at all major defects on the article, comprising that some non-major defects detect.
In the step (b) of module 12, use described initial parameter group to generate a selected user's of reporting to defect map.This initial defect map will comprises detected major defect and non-major defect.
In the step (c) of module 13, during first step or next procedure of check, the image of defect area is stored in the memory device, to be used for analysis subsequently.
In the next procedure (d) of module 14, the image of expression defective is displayed to the user.These images can be the images that is stored in the memory device, also can be the images that obtains from another data source (comprise live intercepting, but this being not unique situation).
Referring now to Fig. 2 and 3, can see the synoptic diagram of this sorting.By watching these images, the user judges whether each defective is serious.Alternatively, the user can determine that detected defective needs meticulousr check.In addition, the user also can add not by the detected manual detection defective of system.Advantageously, not only defective is presented to the user, and display position, the latter will help the automatic adjustment of parameter.Above-mentioned image selection process can be carried out at remote location.
This process can use one of the following option to continue:
A) step (e1), module 15: with different parameter group the image of storage is handled again, thereby obtained new defect map.Handling does not need article are checked once more again, perhaps
B) step (e2), module 16: receive the output of self-identifying/annunciator, to be used for analysis subsequently.
In the step (f) of module 17, system uses heuristic by using specific calculating tailor-made algorithm, and selection can provide optimum detection result's parameter combinations, to form a new parameter setting.When carrying out heuristic, test various parameter settings, generate a new defect map at every turn.Select best defect map, thereby select best parameter setting, as new parameter setting.Heuritic approach can be used in combination with deterministic approach, wherein in case receive through the defect map behind the sorting and some or each through the detected parameters of the defective behind the sorting, each parameter then is set, to obtain best new defect map.Dedicated rules is used to desirable ratio between definite defect type, and these rules are set up, and to obtain to detect the new parameter setting of new defect map, all these defectives are included in the database of default defect type.These rules realize that with mathematical function or logical function perhaps the combination with them realizes.Wherein a kind of available mathematical function is a cost function.Best combination can be stipulated neatly.Alternatively, for cost function of all applied in any combination (as described in following example) of parameter setting and find its extreme value.The method of determining optimal parameter can be respectively to each parameter enforcement, perhaps to one group of parameter enforcement.Figure 4 and 5 show the mode of selecting parameter in greater detail below.
According to the next procedure in the module 18 (g), the initial identification/Report Parameters of system is self-adjusting according to the parameter of above-mentioned selection.This process can use down one of column selection to continue:
A) step (h1), module 19: check same article again with a new parameter group, obtain to have the new defect map of better testing result, and
B) step (h2), module 20: continue the next article of check with this new parameter group.
Advantageously, in step (i), module 21, repeated execution of steps (a) is to (h), with the adjustment of refinement parameter.
Full-automatic method of adjustment is the same with semi-automatic adjustment, except automatically performing the sorting in the step (d) of module 14, it uses more high-resolution image, higher computational resource or longer execution time than all the other steps in this workflow.More high-resolution image can be to have higher color resolution, spatial resolution or the two image.In this case, the stage of only manually carrying out in aforementioned workflow is automatically performed.
With reference to Fig. 2, the example that it shows sorting uses an image 22 among the regional 22a, and this image is acquired during rudimentary check and is stored in the storer.Image deflects zone 23 with those suspected defects 23a is displayed near the correct image 22.The image 22 of reference article is added in the database alternatively, to strengthen the further detection of detected defective.Mathematical filter is applied on the image, to strengthen the visual of defective.Alternatively, sorting is carried out at remote location.
Fig. 3 shows and uses live video acquisition to carry out an example of sorting.The image in display defect zone 24 is to illustrate defective 24a.Alternatively, the image 25 with reference article of correct form 25a is added in the database, to strengthen the further detection of detected defective.Mathematical filter can be applied on the image, to strengthen the visual of defective.Alternatively, sorting is carried out at remote location.
Fig. 4 shows a kind of method for optimizing of selection best identified/Report Parameters.Set up a chart 26 for each parameter of needs adjustment.The value of the adjustable parameter of the X-axis of these charts 26 (26a) expression, and the number of Y-axis (26b) expression detected major defect and non-major defect when changing this parameter.Other correlation parameters also can add in these charts.
Fig. 5 shows a kind of method for optimizing of selection best identified/Report Parameters.By using specific cost function, for each parameter value or parameter combinations are determined a value at cost to each independent parameter or to one group of parameter.By seeking the cost extreme value, extract optimal parameter and be inserted in the checking system.This accompanying drawing is described and has been shown the function of cost function 27 as selected parameter value 28, and has obtained maximum gross error and minimum nonfatal error at optimal selection 29 places.
To utilize cost function to select parameter in order describing, can to use following Example:
The assumed cost function can be expressed as-A* (major defect)+B* (non-major defect)+C* (from the variation of initial parameter value)+D* (defective of the non-classified selection of adding owing to parameter change).
Suppose A=1000, B=10, C=5, D=20.
The initial value of supposing parameter is 60.
Suppose the result (simultaneously referring to Fig. 5) of following table:
| |
40 | 50 | 60 | 70 | 80 | 85 |
| Seriously | 6 | 7 | 7 | 7 | 5 | 6 |
| Non-serious | 20 | 15 | 25 | 10 | 2 | 10 |
| Add | 15 | 2 | - | 1 | 5 | 1 |
The application cost function can obtain:
| |
40 | 50 | 60 | 70 | 80 | 85 |
| Cost | -5400 | -6750 | -6750 | -6930 | -4980 | -6005 |
Therefore, in order to obtain optimum, numerical value 70 (numerical value with least cost) will be selected for this parameter by this system.
The present invention also provides a kind of system for carrying out said process that is used to carry out, and it comprises a checking system, is used for detecting one batch article, determine initial defect map will, and a controller, its operation is used for obtaining initial defect map will from checking system, and shows each defective to the operator.This system makes the operator to carry out sorting to each defective according to type.Controller is used above-mentioned tailor-made algorithm to the selection result who is gathered subsequently, determines that a new parameter setting is used for follow-up check.By using new parameter, obtain an improved defect map, it has the desired ratio between true defect and the false defect.This new parameter setting is used for checking all the other article of this batch.
Checking system also comprises a sensor, be used for imaging is carried out in the zone of article to be checked, a testing agency is used to select position or the display-memory element carried out, and a judgement mechanism, it comprises criterion or the regular device that is used for the optimal result that regulation will search.Also comprise a searched structure, it is used to seek the parameter value of generation by the optimum output result of judgement mechanism defined, and provides parameter value to be used to check the required device of all the other article in this batch.
This system can utilize any optical sensor, and described sensor is to visible light, colorama or gray-level light or to other part sensitivities of electromagnetic spectrum, the optionally capable or array of tdi sensor.
Testing agency uses the data that receive from sensor to detect those suspected defects or zone, and execution parameter is set so better.Testing agency can compare its result alternatively with the reference value that is stored in memory component or the database.Memory component is preserved the image in the zone of being detected that is obtained by sensor, can only preserve the position in the zone of being detected, and can also preserve the data relevant with the reason that causes the detected defective of testing agency in addition.
This system comprises an indication mechanism, and it shows the image in the zone that at least one detects to the user, and described image can be coloured image, gray level image or binary picture, perhaps is stored in the user images in the memory component.Alternatively, display can show and uses mathematical filter or optical filter processed images, perhaps shows other data relevant with the reason that causes the detected defective of testing agency, perhaps about other data of the feature of shown image.
This system also comprises a known user interface own, and it can choose into major defect and non-major defect to shown classification of defects.Judgement mechanism is used to use mathematics and/or logical function to come desirable ratio between the regulation defect type, is configured to obtain to be used to detect the new parameter setting of new defect map, and these defectives all are included in the database of predetermined defect type.Alternatively, described mathematical function can be a cost function of refinement in the parameter setting process, to obtain optimal result.
Data after search mechanism uses heuristic to the classification of defects selection are analyzed, and to obtain new parameter setting, during this period, test different parameter settings at every turn, generate a new defect map.Best defect map, and the optimal parameter that obtains is thus set the parameter setting that selected conduct is new.Only determine parameter setting according to heuristic analysis or in conjunction with deterministic approach, wherein in case receive through the defect map behind the sorting and some or each through the detected parameters in the defective behind the sorting, each parameter then is set to reach best new defect map.Optionally, from several spaces or color resolution, determine at least a parameter setting with layered mode.
It will be apparent to one skilled in the art that the detail that the invention is not restricted to the foregoing description, the present invention can realize in other specific forms, and can not depart from its purport and base attribute.Therefore only to be considered to be restrictive schematically and not in all aspects of present embodiment, scope of the present invention is represented by appending claims, rather than above-mentioned instructions, therefore all changes of carrying out in the implication of the equivalents of claim and scope are all thought acceptable.
Claims (45)
1, a kind of method that is used for determining the parameter setting when automatic checking system is checked a plurality of article, described method comprises the following steps:
Check first article by described checking system;
According to one group of given inspection parameter use a kind of automatic defect method of inspection;
Obtain an initial defect map will;
According to one group of predetermined defect type, the defective of being found is carried out sorting according to defect type;
When defective is carried out sorting,, then described new defective is added in the described initial defect map will that is classified selection if detect the new defective that described checking system can not be discerned;
Calculate tailor-made algorithm by using, use heuristic, described inspection parameter is set automatically, to form adjusted parameter setting;
Use adjusted parameter setting to obtain the adjusted defect map of detected defective, and
Use described adjusted parameter setting to check all the other article in described a plurality of article.
2, the method for claim 1 also comprises and utilizes identical method to check additional article, with the described adjusted parameter setting of further refinement.
3, the method for claim 1 is wherein by realizing check with the described article that are verified of automated optical inspection system scanning.
4, the method for claim 1, wherein said method is automatically performed by controller.
5, the method for claim 1, wherein said sorting is manually realized by professional operating personnel.
6, the method for claim 1, wherein said sorting utilize live video image to realize.
7, the method for claim 1, wherein said sorting are to utilize the image that is stored in the memory component to realize.
8, the method for claim 1, wherein said sorting is realized at remote location.
9, the method for claim 1, wherein said initial defect map will are to check described article to obtain when being set to represent the numerical value of high detection sensitivity when inspection parameter.
10, the method for claim 1, wherein said defect type are classified as major defect and non-major defect.
11, the method for claim 1, wherein said heuristic may further comprise the steps:
The various parameter settings that form adjusted defect map are tested, and
Selection can provide optimum detection result's parameter combinations.
12, method as claimed in claim 11, wherein at least one described parameter setting is determined according to tailor-made algorithm.
13, method as claimed in claim 12, wherein said algorithm uses heuristic to analyzing through the defective behind the sorting, to obtain adjusted parameter group, wherein in case receive described through the defect map behind the sorting and some or each through the detected parameters in the defective behind the sorting, each parameter then is set, to obtain best adjusted defect map.
14, the method for claim 1 wherein uses dedicated rules to come desirable ratio between the regulation defect type.
15, method as claimed in claim 14, wherein said rule are set to obtain to be used for detecting the adjusted parameter setting of adjusted defect map, and these defectives are included in the database of predetermined defect type.
16, method as claimed in claim 15 wherein uses at least a mathematics and/or logical function to realize described rule.
17, method as claimed in claim 16, wherein said mathematical function is a cost function.
18, method as claimed in claim 14, wherein said rule during the parameter setting up procedure by refinement, to obtain optimal results.
19, the method for claim 1 is wherein determined at least one described parameter setting with layered mode from several spaces or color resolution.
20, the method for claim 1, wherein said parameter setting are operated all parameters with control system.
21, a kind of system of the parameter setting when being identified for checking a plurality of article comprises:
One checking system is used for checking first article of one batch, forms an initial defect map will, and a controller, and it is operated and is used for:
Receive described initial defect map will from described checking system;
Show each the described defective in the described initial graph, make operating personnel to carry out sorting to each defective, and the selection result is input in this system according to defect type;
Use tailor-made algorithm to determine adjusted parameter setting, to obtain adjusted defect map, described adjusted defect map has the desirable ratio between true defect and the false defect, and
Provide described parameter setting to check all the other article in described batch.
22, system as claimed in claim 20, wherein said parameter setting is operated to control all parameters of this system.
23, a kind of system of the parameter setting when automatically or semi-automatically being identified for checking a plurality of article comprises:
One sensor is used for imaging is carried out in the zone of the article that are verified;
One testing agency is used to select to be used on the described article position of handling or showing;
One storer can be preserved the image by the Examination region of described sensor acquisition;
One decision unit is used to obtain best defect map;
One searches mechanism, is used to seek the best output result's that can produce the decision unit defined parameter value, and
Provide described parameter value to check the device of all the other article in described batch.
24, system as claimed in claim 23, wherein said sensor is selected from some the part sensor sensitive group to electromagnetic spectrum, comprises visible light sensor; The row of tdi sensor or array, perhaps color sensor or gray-scale sensor.
25, system as claimed in claim 23, wherein said testing agency is selected from following group: the testing agency that uses the data that obtain from described sensor; Defects detection mechanism; Detect those suspected defects and zone and can better carry out the mechanism of described parameter group, and the testing agency of detecting the doubtful zone on the detected article and comparing or do not compare with the institute stored reference value.
26, system as claimed in claim 23, wherein said storer can be preserved the position in the zone of being detected.
27, system as claimed in claim 23, wherein said storer can be preserved the data that expression causes the reason of the detected defective of described testing agency.
28, system as claimed in claim 23 also comprises display.
29, system as claimed in claim 28, wherein said display shows the live image of at least one surveyed area, the form of described image is selected from following image sets: coloured image, gray level image or binary picture.
30, system as claimed in claim 28, wherein said display shows the image that obtains from described storer.
31, system as claimed in claim 28, wherein said display shows by using the image after mathematical filter or optical filter are handled.
32, system as claimed in claim 28, wherein said display shows that expression causes the data of the reason of the detected defective of described testing agency.
33, system as claimed in claim 28, wherein said display shows the additional data relevant with the feature of described image.
34, system as claimed in claim 23 also comprises user interface, can be major defect and non-major defect to the classification of defects selection.
35, system as claimed in claim 23, wherein said decision unit can be stipulated the desirable ratio between the defect type.
36, system as claimed in claim 35, wherein said decision unit is configured to obtain described adjusted parameter setting, is used for detecting the adjusted defect map that is included in predetermined defect type database.
37, system as claimed in claim 35, wherein said decision unit uses at least a mathematics and/or logical function to realize.
38, system as claimed in claim 37, wherein said mathematical function is a cost function.
39, system as claimed in claim 35, wherein said decision unit in the parameter setting up procedure by refinement, to obtain optimal results.
40, system as claimed in claim 23, wherein said search mechanism is used to obtain new parameter setting.
41, system as claimed in claim 23, wherein said search mechanism use heuritic approach to analyzing through the defective behind the sorting, to obtain adjusted parameter setting.
42, system as claimed in claim 41, wherein said heuristic comprises:
The every kind of parameter setting that forms adjusted defect map is tested, and
Selection can provide optimum detection result's parameter combinations.
43, system as claimed in claim 23, wherein said search mechanism use deterministic approach to analyzing through the data after the classification of defects selection, to obtain adjusted parameter setting.
44, system as claimed in claim 43, wherein said deterministic approach comprises:
In case receive described through sorting defect map and some or each through the detected parameters in the defective behind the sorting, each parameter then is set, to obtain best adjusted defect map.
45, system as claimed in claim 23, wherein said parameter setting is operated to control all parameters of this system.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US71142505P | 2005-08-26 | 2005-08-26 | |
| US60/711,425 | 2005-08-26 | ||
| IL177689 | 2006-08-24 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101292263A true CN101292263A (en) | 2008-10-22 |
Family
ID=40035654
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2006800384645A Pending CN101292263A (en) | 2005-08-26 | 2006-08-27 | Method and system for automatic defect detection of articles in visual inspection machines |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN101292263A (en) |
| IL (2) | IL177689A0 (en) |
| TW (1) | TW200716967A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102426440A (en) * | 2010-06-15 | 2012-04-25 | 卡姆特有限公司 | Method and system for monitoring operator of printed circuit board inspection station |
| CN105334216A (en) * | 2014-06-10 | 2016-02-17 | 通用电气公司 | Method and system used for automatic parts inspection |
| US9645097B2 (en) | 2014-06-20 | 2017-05-09 | Kla-Tencor Corporation | In-line wafer edge inspection, wafer pre-alignment, and wafer cleaning |
| US9885671B2 (en) | 2014-06-09 | 2018-02-06 | Kla-Tencor Corporation | Miniaturized imaging apparatus for wafer edge |
| CN109643425A (en) * | 2016-06-28 | 2019-04-16 | 罗伯特·博世有限公司 | System and method for identifying product |
| CN111060520A (en) * | 2019-12-30 | 2020-04-24 | 歌尔股份有限公司 | Product defect detection method, device and system |
| CN111712769A (en) * | 2018-03-06 | 2020-09-25 | 欧姆龙株式会社 | Method, apparatus, system, and program for setting lighting condition, and storage medium |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7937233B2 (en) * | 2008-04-17 | 2011-05-03 | 3M Innovative Properties Company | Preferential defect marking on a web |
-
2006
- 2006-08-24 IL IL177689A patent/IL177689A0/en unknown
- 2006-08-27 CN CNA2006800384645A patent/CN101292263A/en active Pending
- 2006-08-28 TW TW095131607A patent/TW200716967A/en unknown
-
2008
- 2008-02-24 IL IL189710A patent/IL189710A/en not_active IP Right Cessation
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102426440A (en) * | 2010-06-15 | 2012-04-25 | 卡姆特有限公司 | Method and system for monitoring operator of printed circuit board inspection station |
| CN102426440B (en) * | 2010-06-15 | 2015-07-22 | 卡姆特有限公司 | Method and system for monitoring operator of printed circuit board inspection station |
| US9885671B2 (en) | 2014-06-09 | 2018-02-06 | Kla-Tencor Corporation | Miniaturized imaging apparatus for wafer edge |
| CN105334216A (en) * | 2014-06-10 | 2016-02-17 | 通用电气公司 | Method and system used for automatic parts inspection |
| US9645097B2 (en) | 2014-06-20 | 2017-05-09 | Kla-Tencor Corporation | In-line wafer edge inspection, wafer pre-alignment, and wafer cleaning |
| CN109643425A (en) * | 2016-06-28 | 2019-04-16 | 罗伯特·博世有限公司 | System and method for identifying product |
| CN109643425B (en) * | 2016-06-28 | 2024-01-02 | 罗伯特·博世有限公司 | System and method for identifying products |
| CN111712769A (en) * | 2018-03-06 | 2020-09-25 | 欧姆龙株式会社 | Method, apparatus, system, and program for setting lighting condition, and storage medium |
| US11631230B2 (en) | 2018-03-06 | 2023-04-18 | Omron Corporation | Method, device, system and computer-program product for setting lighting condition and storage medium |
| CN111060520A (en) * | 2019-12-30 | 2020-04-24 | 歌尔股份有限公司 | Product defect detection method, device and system |
| CN111060520B (en) * | 2019-12-30 | 2021-10-29 | 歌尔股份有限公司 | Product defect detection method, device and system |
| US11836907B2 (en) | 2019-12-30 | 2023-12-05 | Goertek, Inc. | Product defect detection method, device and system |
Also Published As
| Publication number | Publication date |
|---|---|
| IL189710A0 (en) | 2008-06-05 |
| IL189710A (en) | 2013-07-31 |
| IL177689A0 (en) | 2006-12-31 |
| TW200716967A (en) | 2007-05-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20080281548A1 (en) | Method and System for Automatic Defect Detection of Articles in Visual Inspection Machines | |
| US10964013B2 (en) | System, method for training and applying defect classifiers in wafers having deeply stacked layers | |
| CN101292263A (en) | Method and system for automatic defect detection of articles in visual inspection machines | |
| JP5619776B2 (en) | Method for selecting one or more parameters for wafer inspection | |
| US8660336B2 (en) | Defect inspection system | |
| US8660340B2 (en) | Defect classification method and apparatus, and defect inspection apparatus | |
| CN107004123B (en) | Iterative defect filtering process | |
| CN104303264B (en) | For method, computer system and equipment that the formula of the Automated inspection of semiconductor device is produced | |
| US6941009B2 (en) | Method for evaluating pattern defects on a water surface | |
| CN105184765A (en) | Inspection Apparatus, Inspection Method, And Program | |
| JP6544716B2 (en) | DATA GENERATION APPARATUS, DATA GENERATION METHOD, AND DATA GENERATION PROGRAM | |
| KR20080028278A (en) | Optical inspection system | |
| JP2007198968A (en) | Image-classifying method and image-classifying apparatus | |
| CN109884078A (en) | It is layered wafer inspection | |
| KR102000938B1 (en) | Method for Setting Inspection Criteria Automatically by Indicating Regions and Smart Learning Method for X-ray Inspection Using the Same | |
| KR20200044172A (en) | Deep learning method for analyzing defectiveness | |
| JP5374225B2 (en) | Wafer inspection condition determination method, wafer inspection condition determination system, and wafer inspection system | |
| WO2022190518A1 (en) | Integrated model generation method, image inspection system, image inspection model generation device, image inspection model generation program, and image inspection device | |
| CN107290378A (en) | Material adulterated method and its system are quickly judged based on X-fluorescence | |
| TW577995B (en) | An image searching defect detector | |
| Warner | Using" Just Noticeable Difference" to Automate Visual Inspection of Displays According to Human Visual Perception | |
| KR20150093959A (en) | Method for Setting Inspection Criteria Automatically by Indicating Regions and Smart Learning Method for X-ray Inspection Using the Same |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20081022 |