CN102421279A - Shape basic matching parameter adjusting device, adjusting method, and part installing device - Google Patents

Shape basic matching parameter adjusting device, adjusting method, and part installing device Download PDF

Info

Publication number
CN102421279A
CN102421279A CN2011102886329A CN201110288632A CN102421279A CN 102421279 A CN102421279 A CN 102421279A CN 2011102886329 A CN2011102886329 A CN 2011102886329A CN 201110288632 A CN201110288632 A CN 201110288632A CN 102421279 A CN102421279 A CN 102421279A
Authority
CN
China
Prior art keywords
parameter
image
unit
precision
template
Prior art date
Application number
CN2011102886329A
Other languages
Chinese (zh)
Other versions
CN102421279B (en
Inventor
山田和范
Original Assignee
Juki株式会社
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Priority to JP2010214489A priority Critical patent/JP5759133B2/en
Priority to JP2010-214489 priority
Application filed by Juki株式会社 filed Critical Juki株式会社
Publication of CN102421279A publication Critical patent/CN102421279A/en
Application granted granted Critical
Publication of CN102421279B publication Critical patent/CN102421279B/en

Links

Abstract

The invention provides a shape basic matching parameter adjusting device, a shape basic matching parameter adjusting method, and a part installing device which are corresponding to the characteristics of precision, processing time, reliability and the like in the shape basic matching processing so as to generate the optimal template data. For the part installing device which uses the shape basic matching processing to carry out the carrying positioning of an electronic part (2), the part installing device obtains a searching evaluation image and adjusts the above parameter based on the searching evaluation image. Firstly, the parameter is set, so the positioning processing precision can satisfy the required precision of a user, and then, the parameter changes gradually to reduce the production beat of the positioning processing, and at the same time, the precision evaluation is carried out, and the precision is acquired by taking the required precision marginal value as the optimal parameter. The optimal parameter is processed, and when the part batch changes, the optimal parameter changes correspondingly.

Description

Adjusting device, method of adjustment and the apparatus for mounting component of shape basis match parameter

Technical field

The present invention relates to a kind of shape informations such as edge taper are being compared, from the searching object image, find out the parameter of using in the shape basis coupling of object adjusting device, this parameter method of adjustment and utilize shape basis coupling to carry out the apparatus for mounting component of the location of boarded parts.

Background technology

As utilizing the template data that generates in advance, find out the method that is present in the object in the searching object image, there is a kind of shape basis matching process that shape informations such as edge taper are compared.In this shape basis matching process, the edge judges that the parameter setting of usefulness is extremely important.According to the difference of this parameter setting, the shape of detected edge line can fine change, to matching treatment and utilize the position probing of this processing characteristics such as position detection accuracy, processing time and reliability in handling to exert an influence.

These characteristics require characteristic all different with the adjustment degree to each object, and are current for above-mentioned trickle parameter adjustment, and the user must be corresponding with environment for use and be adjusted.In existing apparatus; To the related parameter of rim detection; For example Cmin value, minimum edge strength, filtering size etc. are provided with the numerical value input function, perhaps be provided with the edge effectively/invalidly utilize window to surround and structure that users such as appointment can utilize dialogic operation to adjust.

In addition, also have following method, that is, carry out and carry out the handling property relevant evaluation of position probing when handling, the template candidate that evaluation of estimate is the highest is confirmed as the template of reality etc., generates template data (for example, with reference to patent documentation 1) automatically.

Patent documentation 1: No. 4470503 communique of Japan's patent

Summary of the invention

But, carry out the user under the situation of aforesaid trickle parameter adjustment, if algorithm and the characteristic handled for position probing do not possess experience, then can't suitably set, existence can't be brought into play the problem of performance to greatest extent.In addition, also have following requirement, that is, be desirably in corresponding expected accuracy of purposes and processing speed under, be extremely carefully adjust, but in existing apparatus, be difficult to realize above-mentioned requirements shape basis matching treatment.

In addition; In apparatus for mounting component, handle and carry out under the situation of location of boarded parts in the position probing of having used shape basis matching treatment, variations such as the material of the parts that cause owing to the change of parts batch and color; The trickle size variation that perhaps causes etc. owing to foozle; Template data during with teaching produces difference, therefore, positioning accuracy fine takes place sometimes worsen problems such as perhaps productive temp delay.For the deterioration that does not become above-mentioned error level,, be difficult to the timing of judging that parameter is adjusted once more if do not examine then can't detect.

Therefore; Problem of the present invention is; The adjusting device of a kind of shape basis match parameter, the method for adjustment and the apparatus for mounting component of shape basis match parameter are provided; It is used for corresponding with the characteristics such as precision, processing time and reliability of shape basis matching treatment, generates optimized template data.

In order to solve above-mentioned problem; The adjusting device of the shape basis match parameter that technical scheme 1 is related; It is adjusted the parameter of in the matching treatment of shape basis, using; It is characterized in that having: designating unit, the precision prescribed of its designated shape basis matching treatment and require productive temp; Evaluation obtains the unit with image, and it will apply the image of the object of the posture after any variation for the benchmark posture, obtains with image as the evaluation of said parameter; And parameter adjustment unit; It uses image based on obtaining the said evaluation that obtains the unit by said evaluation with image; Said parameter is adjusted; Said parameter adjustment unit has: the parameter setting unit, and the said parameter of its automatic setting is so that the precision of shape basis matching treatment satisfies the said precision prescribed by said designating unit appointment; And parameter update unit; It is that productive temp does not satisfy when requiring productive temp by said designating unit appointment said utilizing the result who said evaluation is carried out shape basis matching treatment with image by the said parameter of said parameter setting unit automatic setting; The direction gradual change that the said parameter set by said parameter setting unit is shortened to said productive temp; On one side the precision of shape basis matching treatment being estimated, is the critical parameters value that said precision can be guaranteed said precision prescribed with said parameter update.

Thus, can automatically set the parameter of the best that the precision and requiring of meeting the demands uses in the shape basis matching treatment of productive temp.In addition, when not satisfying precision prescribed and requiring productive temp, can automatically set the parameter of productive temp of the best of the precision that only meets the demands.Therefore,, also can suitably carry out the adjustment of template parameter, can carry out generation and the retrieval of best template data setting with parameter even the user does not possess experience for the algorithm of shape basis matching treatment etc.In addition, the user can specify precision prescribed and require productive temp, therefore, can carry out and the corresponding extremely trickle parameter adjustment of purposes.

In addition, the adjusting device of the shape basis match parameter that technical scheme 2 is related is characterised in that, in the related invention of technical scheme 1, have: take the unit, it is taken said object; And template image is obtained the unit; It utilizes said shooting unit that the said object of benchmark posture is taken; And the image of the said object that will photograph is obtained as template image; Said evaluation obtains the unit based on obtained the said template image that the unit is obtained by said template image with image, generates and has applied the correction image of variable quantity arbitrarily for this template image, and said correction image is obtained with image as said evaluation.

As stated, internally apply change and generate to estimate and use image, utilize the evaluation of generation to carry out parameter adjustment, therefore, can carry out parameter adjustment through only taking 1 template image with image for template image.Therefore, can tackle the less system of the capacity of memory of storage photographic images, can't prepare to estimate the system of the real image of usefulness.

And; The adjusting device of the shape basis match parameter that technical scheme 3 is related is characterised in that; In the related invention of technical scheme 2; Said parameter setting unit has: the fringe region extraction unit, and it extracts the fringe region of being obtained the said template image of obtaining the unit by said template image; And rim detection parameter setting unit; It is based on the distribution of the local edge value in the said fringe region that is extracted by said fringe region extraction unit; With the parameter relevant in the said parameter, be set at the value that obtains best rim detection precision in predefined a plurality of parameter candidate value with the rim detection precision.

As stated, through extracting the fringe region of template image, the distribution of measuring the local edge value of contrast, edge strength, edge ratio etc., thereby the best rim detection parameter of setting.Thus, so that the rim detection precision that produces considerable influence for the precision of shape basis matching treatment becomes good mode setup parameter, the parameter setting of the precision that can meet the demands reliably.

In addition; The adjusting device of the shape basis match parameter that technical scheme 4 is related is characterised in that; In the related invention of technical scheme 2; The precision of shape basis matching treatment is utilized absolute precision evaluation, and this absolute precision is to obtain said variable quantity that the unit applies to said template image as true value by said evaluation with image.

And; The adjusting device of the shape basis match parameter that technical scheme 5 is related is characterised in that; In the related invention of technical scheme 3; The precision of shape basis matching treatment is utilized absolute precision evaluation, and this absolute precision is to obtain said variable quantity that the unit applies to said template image as true value by said evaluation with image.

As stated, owing to can carry out the absolute precision evaluation, the precision evaluation of shape basis matching treatment becomes easy.

In addition, the adjusting device of the shape basis match parameter that technical scheme 6 is related is characterised in that, in the related invention of technical scheme 1, have: take the unit, it is taken said object; And template image is obtained the unit; It utilizes said shooting unit that the said object of benchmark posture is taken; And the image of the said object that will photograph is obtained as template image; Said evaluation obtains the said shooting of unit by using unit with image, repeatedly takes being directed to the said object that the benchmark posture applied the posture that changes arbitrarily, and a plurality of images that photograph are obtained with image as said evaluation.

As stated, image is used in the evaluation that is utilized under a plurality of shooting conditions of mobility scale of the actual searching object image of expression, carries out parameter adjustment, therefore, can obtain the higher parameter of reliability.

In addition; The adjusting device of the shape basis match parameter that technical scheme 7 is related is characterised in that; In the related invention of technical scheme 6; Said parameter setting unit has: extraction unit, and it extracts 2 maximum evaluations of opposed edges variable quantity and uses image from being obtained a plurality of evaluations under the same shooting condition of obtaining the unit with image by said evaluation with the image; The mobile rim extraction unit, it extracts regulation is compared in 2 evaluations that extracted by said extraction unit with edge of image unanimity degree the low mobile rim of decision threshold; And parameter gradual change unit; It makes said parameter to the direction gradual change that will from the object of shape basis matching treatment, remove with the corresponding edge of mobile rim that is extracted by said mobile rim extraction unit on one side; On one side the precision of shape basis matching treatment is estimated, this precision is satisfied the parameter value of said precision prescribed and confirm as said parameter.

As stated, for the more weak edge of contrast etc. since shooting condition and edge detection results becomes unsettled position can from the object of shape basis matching treatment, remove.Therefore, can the meet the demands parameter setting of the best of precision.

In addition; The adjusting device of the shape basis match parameter that technical scheme 8 is related is characterised in that; In the related invention of technical scheme 7; Said parameter comprises: template parameter, and it is utilizing said template image, uses when generating the template data that is made up of the coordinate of the marginal point of said object and edge taper vector; And parameter is used in retrieval; It is based on said template data; Use when from the searching object image, detecting said object; Said parameter gradual change unit, the direction to will from said template data, removing with the corresponding edge of mobile rim that is extracted by said mobile rim extraction unit makes said template parameter gradual change.

As stated, increase, the edge ratio value is reduced, thereby when utilizing template image to generate template data, can mobile rim not extracted through making the minimum edge strength in the template parameter.Therefore, can reliably mobile rim be removed from the object of shape basis matching treatment.

In addition; The adjusting device of the shape basis match parameter that technical scheme 9 is related is characterised in that; In the related invention of technical scheme 7; Said parameter comprises: template parameter, and it is utilizing said template image, uses when generating the template data that is made up of the coordinate of the marginal point of said object and edge taper vector; And parameter is used in retrieval; It is based on said template data; Use when from the searching object image, detecting said object; Said parameter gradual change unit, to be difficult to with the corresponding edge of mobile rim that extracts by said mobile rim extraction unit detected direction from said searching object image, make said retrieval use the parameter gradual change.

And; The adjusting device of the shape basis match parameter that technical scheme 10 is related is characterised in that; In the related invention of technical scheme 8; Said parameter comprises: template parameter, and it is utilizing said template image, uses when generating the template data that is made up of the coordinate of the marginal point of said object and edge taper vector; And parameter is used in retrieval; It is based on said template data; Use when from the searching object image, detecting said object; Said parameter gradual change unit, to be difficult to with the corresponding edge of mobile rim that extracts by said mobile rim extraction unit detected direction from said searching object image, make said retrieval use the parameter gradual change.

As stated, reduce with the minimum edge strength in the parameter, the edge ratio value is increased, thereby when utilizing the searching object image to extract the edge, can mobile rim not extracted through making retrieval.Thus, can reliably mobile rim be removed from the object of shape basis matching treatment.

And; The adjusting device of the shape basis match parameter that technical scheme 11 is related is characterised in that; In technical scheme 6 to 10 in each related invention; The precision of shape basis matching treatment is utilized repeatable accuracy evaluation, and this repeatable accuracy is carried out statistical disposition to the result who has carried out shape basis matching treatment with image to said a plurality of evaluations and obtained.

As stated, carry out the repeatable accuracy evaluation, therefore can carry out the higher evaluation of reliability.

In addition; The adjusting device of the shape basis match parameter that technical scheme 12 is related is characterised in that, in technical scheme 1 to 11 in each related invention, and said shape basis matching treatment; Carrying out coarse search for the result who carries out edge detection process to the searching object image handles; Based on this coarse search process result, carry out smart retrieval process, said parameter update unit; On one side to said edge detection process, said coarse search are handled and the productive temp of said smart retrieval process shortens respectively direction; Make the said parameter gradual change of in this each processing, using, on one side the precision of shape basis matching treatment is estimated, said parameter update can be guaranteed the critical parameters value of said precision prescribed for this precision.

As stated, the relevant parameter of the productive temp with each processing in the matching treatment of shape basis is adjusted respectively, therefore, can further be carried out best parameter adjustment.

And; Method of adjustment in the basis of the shape described in the technical scheme 13 match parameter; It is adjusted the parameter of in the matching treatment of shape basis, using, and it is characterized in that having following step: the precision prescribed of designated shape basis matching treatment and require productive temp; To apply the image of the object of the posture after any variation for the benchmark posture, obtain with image as the evaluation of said parameter; Set said parameter, so that the precision of shape basis matching treatment satisfies said precision prescribed; Use the said parameter of setting, carry out shape basis matching treatment to said evaluation with image, and measure productive temp; And do not satisfy said when requiring productive temp in the productive temp that measures; The direction that shortens to this productive temp on one side makes said parameter gradual change; On one side the precision of shape basis matching treatment is estimated, said parameter update can be guaranteed the critical parameters value of said precision prescribed for this precision.

Thus, can automatically be set in the parameter of the best of using in the matching treatment of shape basis.Therefore, do not possess experience for the algorithm of shape basis matching treatment etc., can carry out method of adjustment yet with the shape basis matching template of the corresponding extremely trickle parameter adjustment of purposes even can become the user.

Apparatus for mounting component described in the technical scheme 14, it utilizes adsorption mouth attract electrons parts, and the assigned position on substrate is installed this electronic unit, it is characterized in that having:

The adjusting device of each described shape basis match parameter in said claim 1 to 12; Memory cell, it is stored in the parameter of using in the matching treatment of shape basis; The searching object image is obtained the unit, and it is obtained the said image that photographs being taken by the image of the electronic unit of said adsorption mouth absorption as the searching object image; Positioning unit, it is utilized in stored parameters in the said memory cell to obtaining the searching object image of obtaining the unit by said searching object image, carries out shape basis matching treatment, carries out the lift-launch location of said electronic unit; Detecting unit, its change to parts batch detects; And judging unit; It is when the change that is gone out said parts batch by said detection; Judge whether to carry out the adjustment of stored parameters in said memory cell; Be in the time of to carry out the adjustment of said parameter by said judgment unit judges, utilizing the adjusting device of said shape basis match parameter to carry out the adjustment of said parameter.

Thus, the information to parts batch change detects aborning, judges whether to carry out parameter adjustment, therefore, can carry out the adjustment once more of parameter in appropriate timing.Thus, can utilize best parameter to carry out shape basis matching treatment (localization process) all the time, can carry out suitable parts and install.

The effect of invention

According to the present invention, can be corresponding with characteristics such as precision in the matching treatment of shape basis, productive temp, reliabilities, set optimized shape basis match parameter.At this moment, with the precision prescribed of user's appointment and require productive temp corresponding, automatically adjust parameter,, also can carry out extremely trickle parameter adjustment accordingly with purposes even therefore the user does not possess experience for the algorithm of shape basis matching treatment etc.

In addition, according to apparatus for mounting component involved in the present invention, the change of detection part batch is adjusted parameter once more, therefore, can carry out the adjustment once more of parameter in appropriate timing.Therefore; Can prevent to cause the variations such as material and color of parts in change owing to parts batch, or because foozle causes that size fine changes etc. down cause precision fine to worsen the perhaps situation of productive temp delay because the template data during with teaching produces difference.

Description of drawings

Fig. 1 is the block diagram of the structure of the apparatus for mounting component among expression the present invention.

Fig. 2 is the figure of the task structure in the presentation video processing unit 12.

Fig. 3 is the flow chart that the optimal parameter of expression the 1st execution mode obtains handling process.

Fig. 4 is the flow chart that the minimum edge strength of expression is set handling process.

Fig. 5 is that the template of expression the 2nd execution mode generates the flow chart with the image taking handling process.

Fig. 6 is the flow chart that the optimal parameter of expression the 2nd execution mode obtains handling process.

Fig. 7 is the flow chart of expression precision parameter adjustment handling process.

Fig. 8 is the figure of the method for distilling of explanation mobile rim.

Fig. 9 is the flow chart of expression productive temp parameter adjustment process flow.

Figure 10 is that the figure of the relation of vector and rotating vector is dwindled in parallel mobile vector, the amplification on the representation feature point.

Embodiment

Below, based on accompanying drawing, execution mode of the present invention is described.

(the 1st execution mode)

(structure)

Fig. 1 is with the block diagram under the situation of adjusting device use in apparatus for mounting component of shape basis matching template involved in the present invention.

Among the figure, label 1 is an apparatus for mounting component.This apparatus for mounting component 1 has: adsorption mouth 3, its attract electrons parts 2; Lighting device 4, electronic unit 2 irradiates lights that it moves to the camera site of regulation to utilizing moving of adsorption mouth 3; And standard camera 5a and high-resolution camera 5b, they are taken the electronic unit 2 that is configured on the above-mentioned camera site.In addition, apparatus for mounting component 1 has: plant control unit 11, the action of its control adsorption mouth 3 and lighting device 4; And image processing apparatus 12, its control camera 5a, 5b, and the image that camera 5a, 5b photograph handles, and the object (electronic unit 2) that is present in the photographic images is carried out position probing (retrieval process), carries out the location of electronic unit 2.In this execution mode, in localization process, use shape basis matching treatment.

Usually according to the electrode size of electronic unit 2, camera 5a or 5b that selection is taken, electronic unit 2 move adsorption mouth 3, so that electronic unit 2 is positioned at the camera site of selected camera after being adsorbed by adsorption mouth 3 to plant control unit 11.And, at this moment, to take in order to utilize selected camera, plant control unit 11 moves lighting device 4 and makes its bright lamp.And plant control unit 11 sends the order of the execution of indication localization process with selected camera channel information to image processing apparatus 12.If receive the response (localization process result) with respect to this order transmission information from image processing apparatus 12, then plant control unit 11 makes adsorption mouth 3 move to the component mounting position of regulation, and electronic unit 2 is carried on substrate.At this moment, based on the absorption position bias and the absorption index angular misalignment value of the electronic unit that from the localization process result, obtains 2, decision parts loading position.

In addition, plant control unit 11 has when parts batch change function from this information to image processing apparatus 12 that notify.At this moment, plant control unit 11 sends the various command parameter relevant with the adjustment of the template data that in localization process, uses.As command parameter, comprise the precision prescribed of localization process and require productive temp.Above-mentioned parameter can be specified (designating unit) by the operator via the man-machine interface 6 that is connected with plant control unit 11.

Image processing apparatus 12 is controlled appointed camera 5a or 5b according to the order that slave unit control device 11 sends, and the image of electronic unit 2 is taken, and carried out various processing to the image that photographs.

Specifically; Image processing apparatus 12 is when slave unit control device 11 receives the order of carrying out localization process, based on the camera channel information; Camera 5a or 5b control; Electronic unit 2 to being positioned on the camera site is taken (the searching object image is obtained the unit), positions processing to photographic images, and its result is fed back (positioning unit) to plant control unit 11.In addition; When image processing apparatus 12 receives the order of producing beginning at slave unit control device 11, when beginning to stipulated time (for example, carrying the time of the electronic unit 2 of specified quantity) and each localization process from production; Carry out following processing; That is, measure the precision and the productive temp of this localization process, and keep in the job memory 24 that this repeatable accuracy and productive temp are stated backward.And; When image processing apparatus 12 receives the order of change of notice parts batch at slave unit control device 11; Carry out following processing (optimal parameter obtains processing); That is, template parameter that is used for being created on the template data that localization process uses and the retrieval of in retrieval process, using are adjusted with parameter.

Image processing apparatus 12 has A/D transformation component 21, video memory 22, operational part 23, operation with memory 24, parameter storage part 25, control part 26, concurrent operation portion 27, interface 28 and D/A transformation component 29.

21 pairs of view data that photographed by camera 5a, 5b of A/D transformation component are carried out the A/D conversion, and store in video memory 22 as multivalue image data.Operational part 23 is based on image stored data in the video memory 22, positions processing and optimal parameter and obtains processing.Operation is stored with 24 pairs of deal with data that in processing, generated by operational part 23 of memory.Parameter is used in parameter storage part 25 storing template parameters, template data and retrieval.Control part 26 control camera 5a, 5b.The processing that filtering operation etc. requires processing speed is carried out in concurrent operation portion 27 and the processing in the operational part 23 concurrently.Interface 28 carries out the reception of signal and sends between itself and plant control unit 11.D/A transformation component 29 will the image stored data carry out the D/A conversion in video memory 22, will in display 7, be shown by the image that camera 5a, 5b photograph.

Fig. 2 is the figure of the task structure in the presentation video processing unit 12.

As this is shown in Figure 2, image processing apparatus 12 is made up of execute the task 12c and template adjustment task 12d of command analysis task 12a, image incoming task 12b, identification.

Command analysis task 12a carries out from the reception of the order of plant control unit 11 and localization process result's etc. the response transmission to plant control unit 11.In addition, do not need the execution of the order of parallel processing to be undertaken by command analysis task 12a.

Receive localization process etc. at slave unit control device 11 and require under the exectorial situation of processing of productive temps at a high speed, command analysis task 12a exports the execution requirement respectively to image incoming task 12b, the identification 12c that executes the task.Therefore, during the 12a that executed the task by identification positions processing, can take next image of component, can position processing concurrently with the image input by image incoming task 12b.

And, when command analysis task 12a receives the change notice of parts batch at slave unit control device 11, carry out requirement to template adjustment task 12d output, carry out optimal parameter by background process and obtain processing.

Below, explain that the optimal parameter of being carried out by image processing apparatus 12 obtains the overall flow of processing.

Image processing apparatus 12 receives at slave unit control device 11 under the situation of order of change of notice parts batch; Command analysis task 12a adjusts task 12d starting (detecting unit) with template; Precision and productive temp that measurement and positioning is handled, and with operation with memory 24 in the value of maintenance compare.Here; For precision and productive temp to localization process are estimated; Obtain electronic unit 2 is taken and the retrieval evaluation map picture that obtains,, be utilized in the template data that this timing stores and retrieve and use parameter in parameter storage part 25 to retrieval evaluation map picture; Position processing, thereby precision and productive temp are measured.

And; Surpass certain permissible range with respect to the value that keeps in memory 24 in operation and under the situation about worsening in the precision of the localization process that measures or productive temp; Be judged as the adjustment (optimal parameter obtains processing) that must carry out template data, command analysis task 12a will be made as ON (judging unit) at the marks for treatment that operation is provided with in memory 24.At this moment, template adjustment task 12d temporarily becomes dormant state.

If image incoming task 12b detects above-mentioned marks for treatment and is made as ON, the photographing process of production usefulness is interrupted, the image (template image) of template data adjustment usefulness is taken, and storage in video memory 22.

If through image incoming task 12b the image of template data adjustment usefulness is taken, then command analysis task 12a adjusts task 12d starting with template once more, and the beginning optimal parameter obtains processing.At this moment, the relative importance value of template adjustment task 12d is corresponding and definite with the deterioration degree of precision and productive temp, under the lower situation of urgency (deterioration degree), handles down between at one's leisure etc., does not make production efficiency decline as far as possible.On the other hand, under urgency (deterioration degree) condition with higher, carry out the adjustment once more of template etc. at once, reduce can not make production stability.

If optimal parameter obtains processing and finishes, command analysis task 12a precision and the productive temp handled of measurement and positioning once more then, the value renewal that will keep in memory 24 in operation, and marks for treatment is made as OFF.At this moment, template adjustment task 12d becomes dormant state.

Below, specify the optimal parameter of carrying out by image processing apparatus 12 and obtain processing.

Fig. 3 is the flow chart that the expression optimal parameter obtains handling process.

This optimal parameter obtains processing, is to detect under the situation that above-mentioned marks for treatment is ON at image incoming task 12b to begin to carry out.In this execution mode, be (adjustment of template data) that obtain of from 1 template image, carrying out optimal parameter.

At first, in step S1,12 couples of camera 5a or 5b by plant control unit 11 appointments of image processing apparatus control, and the image of the object (electronic unit) 2 of benchmark posture is taken.The image that photographs carries out digitlization by A/D transformation component 21, and stores in video memory 22 as template image data.

In step S2, image processing apparatus 12 carries out the initial setting of template parameter then, and it is stored in parameter storage part 25, jumps to step S3.In this execution mode; As template parameter, use Cmin value (being regarded as the minimum value of the concentration value of object), minimum edge strength (being regarded as the edge strength of the minimum at edge), edge ratio (width of edge taper=filtering size), maximal margin count (counting of the edge of logining as template data)., the initial value of each parameter for example being set at Cmin value=80 here, minimum edge strength=32, edge ratio=2, edge count=and 1024.

In step S3, image processing apparatus 12 utilizes known techniques of discriminant analysis to the template image data of storage in video memory 22, obtains the threshold value that the background area is separated with the object zone, and it is obtained as the Cmin value.

Then, in step S4, image processing apparatus 12 is utilized in the Cmin value that obtains among the said step S3, with the template image data binaryzation.And, corresponding with rim detection filtering size, carry out expansion process, adjustment Cmin value is so that the edge taper interval becomes the object area side.And, upgrade with the initial value of adjusted Cmin value to storage in parameter storage part 25.

Then, in step S5, image processing apparatus 12 carries out Filtering Processing, and jumps to step S6 to the template image data of storage in video memory 22.,, in predefined edge ratio candidate value, utilize maximum edge ratio value (2.5) here, carry out Filtering Processing in order accurately to obtain best result.

In step S6; Image processing apparatus 12 is utilized in the binaryzation data of obtaining among the said step S4; To the filtering output value of in said step S5, obtaining; Carry out statistical disposition respectively for background area, object zone, obtain mean value, dispersion value, minimum value, the maximum of each regional CONCENTRATION DISTRIBUTION.And, utilize above-mentioned value, carry out processing shown in Figure 4, obtain minimum edge strength MinPwr.

As shown in Figure 4, at first in step S6a, minimum edge strength MinPwr is set at background maximal margin intensity a, and jumps to step S6b, whether background maximal margin intensity a is judged than the minimum edge strength b of object for a short time.And, under the situation of a<b, jump to step S6c, final minimum edge strength MinPwr is set at the mean value (a+b)/2 of the minimum edge strength b of background maximal margin intensity a and object, end process.

In addition, in step S6b, be judged as under the situation of a >=b, jump to step S6d, whether background maximal margin intensity a is judged than the average edge strength c of object for a short time.And, under the situation of a<c, jump to step S6e, under the situation of a >=c, minimum edge strength MinPwr is remained the background maximal margin intensity a that in step S6a, sets, directly end process.

In step S6e, based on following formula, the minimum edge strength b of calculating object thing, and jump to step S6f.

b=c-d·k……(1)

Here, d is the object dispersion value, and k is a correction coefficient.

And, in step S6f, whether background maximal margin intensity a is judged than the minimum edge strength b of object for a short time.And, under the situation of a<b, jump to step S6g, final minimum edge strength MinPwr is set at the mean value (a+b)/2 of the minimum edge strength b of background maximal margin intensity a and object, end process.

On the other hand, in step S6f, be judged as under the situation of a >=b, minimum edge strength MinPwr is remained the background maximal margin intensity a that in step S6a, sets, directly end process.

That is,,, then can realize best rim detection here,, carry out minimum edge strength and obtain processing based on such thinking if between background maximal margin intensity a and the minimum edge strength b of object, threshold value (minimum edge strength MinPwr) is set.As stated, in this execution mode, extract the fringe region of template image,, set best rim detection parameter (minimum edge strength) based on the distribution of the local edge value of contrast, edge strength, edge ratio etc.

Be back to Fig. 3, in step S7, image processing apparatus 12 is utilized in the minimum edge strength MinPwr that is provided with among the said step S6, to the filtering output value of in said step S5, obtaining, carries out threshold process, detected edge points.

Then, in step S8, whether the adjustment of 12 pairs of edge ratio values of image processing apparatus finishes to judge.,, utilize predefined whole edge ratio candidate value here, carry out filtering operation respectively, judge and whether carried out estimating (edge evaluation) to each filtering operation result to detected marginal point in said step S7.Edge ratio candidate value for example is made as 1,2,2.5.And, being judged as under the situation that whole edge evaluations do not have to finish, jump to step S9, be judged as under the situation that whole edge evaluations finishes the step S12 that states after jumping to.

In step S9, image processing apparatus 12 switches to the little 1 grade value of not carrying out filtering operation with the edge ratio value, jumps to step S10.In addition, when first, the edge ratio value is made as maximum (2.5), with it as edge ratio fiducial value.

In step S10, image processing apparatus 12 carries out filtering operation to the detected marginal point of said step S7 with the edge ratio value of being set by said step S9.

Then, in step S11, image processing apparatus 12 carries out the edge evaluation, and jumps to said step S8 to the filtering operation result (filtering output value) who in said step S10, carries out.Specifically; At each marginal point place; To the filtering output value in said step S10 and on the edge of the filtering output value under the ratio fiducial value compare, with big (reaction of filtering the is stronger) person's of filtering output value edge ratio value, select as the edge ratio value of the best of this marginal point.And the edge ratio value of the best that will select at maximum marginal point places as template parameter, is made as best edge ratio value, and with it as edge ratio fiducial value, and the edge ratio value that will in parameter storage part 25, store is updated to this value.In addition, big if filtering size (edge ratio) becomes, then filtering output value also becomes greatly, therefore, even,, carry out the edge evaluation with the output valve normalization of each filtering size under the different filtering size, also can compare the mode of magnitude relationship.

According to above processing, accomplish the obtaining of template parameter value of the best relevant with the precision of localization process.Promptly; At this regularly; The parameter relevant in the template parameter (Cmin value, minimum edge strength, edge ratio value) with the rim detection precision; Become the value that can in the scope of preset parameters candidate value, obtain best rim detection precision, become the state of the precision prescribed that satisfies operator's appointment.

Therefore, in following processing, carry out the obtaining of template parameter value of the best relevant with productive temp.Here, to the direction change template parameter (maximal margin is counted) that productive temp is shortened, carry out retrieval process at every turn and precision and productive temp are estimated.And, the method for using the parameter value under the critical productive temp can keep precision prescribed to obtain as the parameter value of the best.

In step S12, image processing apparatus 12 is utilized in the template parameter value that this timing is stored in parameter storage part 25, generates template data, jumps to step S13., utilize the template parameter value here, detect the edge of template image, generate the list that coordinate and edge taper vector by marginal point constitute, and with it as template data.At this moment, count and elongate at interval and handle, confirm the marginal point of in template data, logining in order to be no more than maximal margin.The template data that generates is stored in parameter storage part 25.

In step S13, image processing apparatus 12 generates the image that usefulness is estimated in retrieval, and storage in video memory 22.,, generate the view data applied any variation (parallelly move, rotation, amplify and dwindle) here to the template image data of storage in video memory 22, and with it as retrieval evaluation map picture.In addition, the variable quantity that here applies, as localization process result's true value, and as the absolute precision evaluation of estimate of localization process and use.

Then, in step S14, image processing apparatus 12 is utilized in the template data that generates among the said step S12, looks like to carry out retrieval process to the retrieval evaluation map that in said step S13, generates.

The retrieval process flow process is described simply here.

Retrieval process in this execution mode is to utilize the pattern match of pyramid structure retrieval to handle, and parameter is used in the retrieval that is based on storage in the parameter storage part 25, and the position that is present in the object in the photographic images is detected.Use parameter as retrieval, use Cmin value, minimum edge strength, edge ratio value, pyramid to begin layer.

In the pyramid structure retrieval, at first begin from the searching object image, utilize above-mentioned retrieval to retrieve the edge with parameter, this image is compressed to begin the corresponding compression ratio of layer with pyramid.Carry out bulk processing (coarse search) for compressed image then, the coordinate position roughly that the detected object thing exists.Then, near detected coordinate, utilize template data to carry out detailed process (smart retrieval), further detect correct position.This essence retrieval is along with level is advanced, and precision is improved gradually carry out.As stated, through after carrying out coarse search and having confirmed approximate location, based on this result; In specific scope, carry out detailed retrieval; Thereby needn't carry out detailed retrieval for integral image, can only accurately check successively total process for producing beat is shortened necessary part.

That is, in this step S14, the retrieval evaluation map picture that will in said step S13, generate carries out retrieval process as above-mentioned searching object image.The retrieval of using in the retrieval process here is based on the template data that this is regularly obtained with parameter, utilizes following rule to set.

(1) uses under the situation of parameter value as default value in template, become default value.

(2) under the situation that is not default value, set with parameter than template loose (for example, based on the candidate value of each parameter shown in the table 1, downward 1 grade).

(3) pyramid begin the layer set accordingly with the profile size of object.

[table 1]

Project Default value Beyond the default value Cmin 80 From 50,80,100, select Minimum edge strength 32 From 16,32,64, select The edge ratio value 2 From 1,2,2.5, select Pyramid begins layer 3 It is 0 that template size is less than or equal under 64 the situation

In step S15; Image processing apparatus 12 is according to the retrieval process result of said step S14; Obtain productive temp and precision (bias of localization process result and true value (variable quantity that in said step S13, applies)), they are kept in job memory 24 as fiducial value.

Below, in step S16, carry out the productive temp adjustment and handle the judgement that whether finishes.Here; The productive temp that in job memory 24, keeps reaches under the situation of requirement productive temp; Productive temp is judged as productive temp adjustment processing and finishes not than surpassing under the situation of predefined stipulated number under the situation of last sub-value shortening and at the adjustment number of times.And, being judged as under the situation that continues productive temp adjustment processing, jump to step S17, under the situation that is judged as productive temp adjustment processing end, the step S23 that states after jumping to.

In step S17, image processing apparatus 12 carries out the optimization that maximal margin is counted.The productive temp of localization process depends on the quantity of the marginal point of in template data, logining.Therefore, if make the minimizing of counting of this edge, then can shorten productive temp.But precision can worsen.

In this execution mode, count through the edge that will in template data, login and to reduce interimly, the precision of localization process is confirmed at every turn, can keep the edge under the critical productive temp of precision prescribed to count thereby obtain.And, this edge counted counts as the edge of the best.Here, the minimizing of counting in the edge is with the marginal point mode of equivalent arrangements on the line on the edge of, utilizes to elongate to handle at interval and carries out.That is, count with 1,2,3 through elongating at interval ... Order increase, count thereby reduce the edge interimly.

When utilizing template parameter to generate template data, automatically elongate at interval and handle to be no more than mode that maximal margin counts, therefore,, thereby can carry out the reduction (elongating the increase of counting in the interval) of counting at the edge through minimizings that maximal margin is counted.Therefore, in this step S17, maximal margin counted is set at the value of the little setting of value than storage in parameter storage part 25.

Then, in step S18, image processing apparatus 12 is utilized in the template parameter of setting among the said step S17, regenerates template data, jumps to step S19.

In step S19, image processing apparatus 12 is utilized in the template data that regenerates among the said step S18, and the retrieval evaluation map picture in said step S13, generating carries out retrieval process, jumps to step S20.

In step S20; Image processing apparatus 12 utilizes the retrieval process result of said step S19; Obtain productive temp and precision, jump to step S21, the precision of the localization process in said step S20, obtained and the fiducial value that in job memory 24, keeps are compared.And the precision of localization process at this moment and fiducial value relatively do not have to worsen under the situation of (deterioration of precision drops in the allowed band), jump to step S22, and the fiducial value of productive temp and precision is upgraded, and jump to said step S16.

On the other hand, the precision of localization process at this moment and fiducial value comparison and worsen under the situation of (deterioration of precision drop on allowed band outer), previous maximal margin is counted is judged as the critical value that can keep precision prescribed.Therefore, in this case, jump to step S23, carry out the initial setting of retrieval process with parameter.Here, the template of the best that obtains based on the processing that utilizes hereto use parameter value, utilizes the initial value of setting retrieval process in above-mentioned rule (1)~(3) with parameter.

In addition, standard camera 5a and high-res camera 5b are corresponding with the shooting unit, and parameter storage part 25 is corresponding with memory cell.In addition, in Fig. 3, it is corresponding that step S1 and template image are obtained the unit, and step S3~S5 is corresponding with the fringe region extraction unit, and step S6 is corresponding with rim detection parameter setting unit, step S13 with estimate that to obtain the unit with image corresponding.In addition, step S3~S11 is corresponding with the parameter setting unit, and step S12~S22 is corresponding with the parameter update unit.

(action)

Below, the action of the 1st execution mode is described.

Current, in the parameter storage part 25 of image processing apparatus 12, parameter is used in the retrieval that stores best template data and the best.Under this state lower component erecting device 1 carried out situation that the lift-launch of electronic unit 2 handles, plant control unit 11 made adsorption mouth 3 supply with the position to the parts of electronic part feeder (not shown) and moves, and carries out the absorption of electronic unit 2.If attract electrons parts 2, then plant control unit 11 makes adsorption mouth 3 move to the camera site of regulation.At this moment, select the camera 5a or the 5b that take accordingly, adsorption mouth 3 is moved so that electronic unit 2 is positioned at the camera site of selected camera with adsorption element.If adsorption mouth 3 moves to the camera site, then plant control unit 11 control lighting devices 4 to electronic unit 2 irradiating illumination light, with selected camera channel information, send the order of the execution of indication localization process to image processing apparatus 12.

Image processing apparatus 12 receives via interface 28 should order.Through command analysis task 21a image incoming task 12b is started then, control part 26 control cameras, thus the electronic unit 2 that is positioned on the camera site is taken.The image of the electronic unit 2 that photographs is by 21 digitlizations of A/D transformation component, and storage in video memory 22.This view data becomes the searching object image.In addition, this moment, image stored data in video memory 22 were simulated by D/A transformation component 29, and in display 7, showed.

If accomplish the preparation of searching object image, then command analysis task 12a utilizes operational part 23 and concurrent operation portion 27 to carry out the localization process of electronic units 2 the identification 12c starting of executing the task.In this localization process, at first, to the searching object image be applied in the parameter storage part 25 store with parameter as retrieval with the corresponding filtering of edge ratio, utilize minimum edge strength, carry out threshold process, extract the interior marginal point of searching object image.

Then, with this edge image with begin the corresponding compression ratio of layer as retrieval with the pyramid of parameter and compress, be utilized in the template data of storing in the parameter storage part 25, carry out coarse search.In this coarse search, utilize edge retrieval, regular coordinate indexing, vague generalization Hough conversion, how much known retrieval techniques such as hashing, from the searching object image, obtain the approximate location and the posture of object.

Then, carry out the essence retrieval, detect the detailed position of object for the search domain of obtaining by coarse search.In this essence retrieval; The position of the object that will in coarse search, obtain and posture be position and starting position to start with; The template (pattern) and searching object picture registration that will generate by template data; Through parallel move or rotation is moved, ratio changes of carrying out template repeatedly, thereby detect the highest position of consistent degree with the interior object of searching object image.And,, obtain the detail location of the object in the searching object image based on the amount of movement of this template.Therefore; Can obtain localization process result (to parallel amount of movement, rotation amount of movement, the ratio variable quantity of the object in the searching object image of template); Absorption position bias, absorption index angular misalignment value can be known, thereby the location can be carried by the electronic unit 2 of adsorption mouth 3 absorption.

If localization process finishes, then discerning the 12c that executes the task becomes dormant state, and command analysis task 12a as the response of sending to plant control unit 11, sends via interface 28 the localization process result to plant control unit 11.

If receive the localization process result from image processing apparatus 12, then plant control unit 11 is corresponding with this localization process result, and the adjustment component loading position carries electronic unit 2 on substrate.Therefore, can electronic unit 2 suitably be carried on the position of expectation.

Then, if parts batch change, then plant control unit 11 is notified this information to image processing apparatus 12.At this moment, plant control unit 11 also to image processing apparatus 12 send by the precision prescribed of the localization process of operator's appointment with require productive temp.

If image processing apparatus 12 slave unit control device 11 receive the change notice of parts batch via interface 28; Then image incoming task 12b is started through command analysis task 12a; Control part 26 camera are controlled, thereby the image of the electronic unit 2 of benchmark posture is taken.The image of the electronic unit 2 that photographs is by 21 digitlizations of A/D transformation component, and storage in video memory 22.Then, command analysis task 12a adjusts task 12d starting with template, generates the view data that has applied any variation (parallelly move, rotation, amplify and dwindle) for image stored data in video memory 22, and it is stored in job memory 24.This view data becomes precision and productive temp that is used for the measurement and positioning processing and the retrieval evaluation map picture that need to judge whether parameter adjustment.And, for the retrieval evaluation map picture that generates, be utilized in the template data of storage in the parameter storage part 25, carry out retrieval process, obtain the localization process result.

At this moment, through the localization process result that obtains and true value (variable quantity that uses when the generation of retrieval evaluation map picture) are compared, thereby obtain the precision of localization process.In addition, the also productive temp handled of measurement and positioning simultaneously.And; With they with operation with memory 24 in the maintenance value compare; Under the situation about worsening surpassing permissible range, be judged as template data and the retrieval that reply store and adjust with parameter in parameter storage part 25, the execution optimal parameter obtains processing.

Obtain in the processing in optimal parameter, the parameter that will be used to generate template data is that Cmin value, minimum edge strength, edge ratio, maximal margin are counted and be adjusted to best value.In this execution mode; The image that template is generated usefulness is that template image is taken (the step S1 of Fig. 3); This template image is handled; Template parameter that will be relevant with the precision of localization process (Cmin value, minimum edge strength, edge ratio) is set at the best optimal value of precision (step S3~S11).

And, utilize the template data that generates based on above-mentioned template parameter, to the retrieval evaluation map picture that generates by template image, carry out retrieval process (retrieval test), obtain the precision of localization process, it is kept as fiducial value (step S12~S15).Then; Template parameter (maximal margin is counted) that will be relevant with the productive temp of localization process; Change to the direction that productive temp improves gradually from initial value; And in each adjustment, precision is being estimated, obtain (step S16~S22) as the parameter value of the best with maintaining the meet the demands critical value of said reference value of precision of this precision.

In the adjustment of existing templates parameter, in fact the operator carries out the numerical value input of parameter value, and perhaps effective/invalid the window that utilizes with the edge surrounds and appointment etc., and the operator must adjust with self environment for use accordingly.Therefore,, must possess experience, not only carry out high accuracy and the adjustment of the best at a high speed, and carry out with the corresponding extremely trickle adjustment of purposes very difficult to the algorithm and the characteristic of localization process in order to carry out trickle parameter adjustment to each object.

Relative therewith, in this execution mode, can automatically obtain the template parameter that meets the demands precision and require productive temp, therefore can suitably carry out trickle parameter adjustment.In addition, owing to be the structure that the operator can indicate precision prescribed and require productive temp, thus also passable even for example how much sacrifice precision, therefore can the corresponding adjustment of purposes such as carry out more at high speed with expectation.

In addition, when the change of parts batch, need to judge whether parameter adjustment, and require correspondingly, carry out parameter adjustment, even therefore also can be updated to proper parameters under the situation of parts material, color, size variation owing to the change of parts batch causes.Therefore, the deterioration of accuracy of detection and productive temp automatically, and can be automatically with its reparation.

(effect)

As stated; In the 1st execution mode; Not only be used to generate the template data that can carry out position probing, and can set be used for generating with situation about handle using in position probing under characteristics such as position detection accuracy, productive temp, the reliability parameter of optimized template data accordingly.In addition; With the precision prescribed of operator's appointment and require productive temp corresponding; Therefore automatically adjust parameter,, also can carry out and the corresponding extremely trickle parameter adjustment of purposes even the operator does not possess experience for the algorithm of shape basis matching treatment etc.

At this moment,, internally apply variation, generate retrieval evaluation map picture, utilize the retrieval evaluation map that generates to look like to carry out parameter adjustment, therefore, can carry out parameter adjustment only through taking a template image to template image.Therefore, can tackle the system that the less system of the capacity of memory of storage photographic images and can't preparing estimates the real image of usefulness.In addition, can be with the true value of the variable quantity that applies for template image as the localization process result, and use as the absolute precision evaluation of estimate of localization process, so the precision evaluation of shape basis matching treatment is easy.

And, when carrying out parameter adjustment, at first, with the meet the demands mode setup parameter of precision of the precision of localization process.At this moment, through extracting the fringe region of template image, the distribution of measuring the local edge value of contrast, edge strength, edge ratio etc., thereby the best rim detection parameter (threshold value of using in the threshold process that detects on the edge of) of setting.As stated, can become best mode setup parameter with the rim detection precision, therefore, the parameter setting of the precision that can meet the demands reliably.

In addition, when carrying out parameter adjustment, maximal margin counted cut down interimly, when each adjustment, the precision of localization process is estimated, guaranteeing that obtaining necessary edge under the critical beat of precision prescribed counts.As stated, maximal margin counted cut down interimly, therefore, the direction gradual change that parameter is shortened to productive temp.In addition, when the direction of productive temp shortening makes the parameter gradual change, the precision of localization process is estimated, obtained best parameter value, therefore can carry out extremely trickle parameter adjustment.

And; The change of detection part batch is aborning judged whether adjusting parameter, therefore; Cause the variation such as material and color of parts in switching owing to parts batch; Produce under the situation of difference between the template data during down with teaching such as the trickle size variation that perhaps causes, can automatically carry out the adjustment once more of parameter, be updated to the parameter of the best owing to foozle.Therefore, can prevent that trickle precision that the change owing to parts batch causes from worsening or the delay of productive temp.

Therefore, in above-mentioned apparatus for mounting component, can carry out high-precision component mounting all the time.

(the 2nd execution mode)

Below, the 2nd execution mode of the present invention is described.

In described the 1st execution mode, utilize the retrieval evaluation map picture that generates by template image, carry out optimal parameter and obtain processing, relative with it, a plurality of retrieval evaluation map pictures that the 2nd execution mode utilizes actual photographed to arrive carry out optimal parameter and obtain processing.Under the situation about using in the initial generation of the template before production, consider generable posture change scope, illumination change scope when producing, prepare to retrieve the evaluation map picture.

(structure)

The plant control unit 11 of this execution mode; As the command parameter relevant with the adjustment of template data; Precision prescribed in localization process, require on the basis of productive temp, with the shooting posture (evaluation posture) of retrieval evaluation map picture, take number, the shooting condition of retrieval evaluation map picture of brightness of illumination etc. when taking sends to image processing apparatus 12.The shooting condition of above-mentioned retrieval evaluation map picture is specified by the operator via man-machine interface 6.

Fig. 5 is that the template of expression the 2nd execution mode generates the flow chart with the image taking handling process.

At first, in step S31, image processing apparatus 12 control is by the camera 5a or the 5b of plant control unit 11 appointments, and the image of the object (electronic unit) 2 of benchmark posture is taken.The image that photographs is by 21 digitlizations of A/D transformation component, and stores in video memory 22 as template image data.

Then, in step S32, whether the photographing process of 12 pairs of retrievals of image processing apparatus evaluation map picture finishes to judge.Here, judge and whether under by the shooting condition of plant control unit 11 appointments, image taking and end are estimated in retrieval, not having to jump to step S33 under the situation about finishing, under situation about finishing, the step S34 that states after jumping to.

In step S33, image processing apparatus 12 sends following order to plant control unit 11, that is, the adsorption mouth 3 that is adsorbed with object (electronic unit 2) is moved to an evaluation posture as shooting condition, and specify the brightness of illumination under this evaluation posture.If receive the response (the shooting order of retrieval evaluation map picture) from plant control unit 11, then the camera 5a or the 5b of 12 pairs of appointments of image processing apparatus control, and the object of estimating under the posture (electronic unit 2) is taken.At this moment, take the number of appointment.If finish a shooting of estimating the retrieval evaluation map picture under the posture, then jump to said step S32.As stated, through carrying out the processing of step S32 and S33 repeatedly, retrieve the evaluation map picture thereby under whole shooting conditions, take.

In step S34, image processing apparatus 12 generates the list of the retrieval evaluation map picture that photographs, and it is stored in video memory 22 as adding data.

Below, explain that the optimal parameter in this execution mode obtains processing.

Fig. 6 is the flow chart that the optimal parameter of expression the 2nd execution mode obtains handling process.

At first, in step S41, image processing apparatus 12 carries out the initial setting of template parameter, and jumps to step S42.Here, it is Cmin value=80 that template parameter is set at default value, minimum edge strength=32, edge ratio=2, edge count=and 1024.The template parameter of setting is stored in operation with in the memory 24.

In addition, in this step S41, also can carry out obtaining processing, carry out the initial setting of template parameter like described optimal parameter shown in Figure 3.In this case, can the initial value of template parameter be made as useful value, therefore, in following processing, can with comparalive ease the template parameter value be converged to optimum value.

In step S42, image processing apparatus 12 is utilized in the template parameter of operation with storage in the memory 24, carries out the generation of template data, and jumps to step S43.

In step S43, image processing apparatus 12 judges whether look like to have carried out retrieval process for the whole retrieval evaluation map that in processing shown in Figure 5, photographs.And, under the situation that does not have to finish in whole retrieval process, jump to step S44, under the situation that whole retrieval process finishes, the step S46 that states after jumping to.

In step S44, image processing apparatus 12 is utilized in the template data that generates among the said step S42, and a plurality of retrieval evaluation map pictures to the object to same evaluation posture photographs carry out retrieval process respectively, obtain the localization process result.

Then, in step S45, image processing apparatus 12 is obtained the statistical disposition data, and is jumped to said step S43 to the localization process result who in said step S44, obtains.Here, by the localization process result data of statistical disposition, for the centre coordinate of template (x, y), (x y), obtains mean value and standard deviation (3 σ) respectively to the solstics coordinate.In addition, for productive temp, also measure these 4 of total process for producing beat, rim detection productive temp, coarse search productive temp, smart retrieval productive temps, and obtain average productive temp, minimum production beat, largest production beat respectively.And, also obtain localization process result (object in the retrieval evaluation map picture is with respect to parallel amount of movement, rotation amount, the ratio variable quantity of template) here.

In addition; In step S46; Image processing apparatus 12 carries out following evaluation, that is, whether the standard deviation (3 σ) of standard deviation of the centre coordinate of the template of in said step S45, obtaining (3 σ) and solstics coordinate drops on respectively in the permissible range with the corresponding regulation of precision prescribed.

And; In step S47; Image processing apparatus 12 drops under the situation in the permissible range being judged as each standard deviation in said step S46 (3 σ), needn't adjust the parameter (precision parameter) relevant with the precision of template data, and the step S52 that states after jumping to.On the other hand, under being judged as situation about dropping on outside the permissible range, the precision parameter of reply template data is adjusted, and jumps to step S48.

The adjustment of precision parameter is to carry out through the minimum edge strength in the template parameter and the minimum edge strength in edge ratio and the search argument and edge ratio are adjusted.In this execution mode, the alter mode of above-mentioned parameter does, from a plurality of candidate values, becomes strict or each 1 grade of ground becomes loose to the each 1 grade of ground of direction that precision improves, and all carries out retrieval process at every turn, and precision is estimated.Carry out this processing repeatedly and till the precision that meets the demands, (carry out retry).

In step S48, whether the number of retries of 12 pairs of precision parameter adjustment of image processing apparatus surpasses predefined maximum times is judged.And, not surpassing under the situation of maximum times, jump to step S49, carry out the precision parameter adjustment and handle.

Fig. 7 is the flow chart that is illustrated in the precision parameter adjustment handling process of carrying out among the step S49.

At first, in step S49a, in many results, extract 2 maximum images of difference for the statistical disposition of retrieval evaluation view data.Here, to the group of same shooting condition (taking posture, brightness of illumination etc.), handle respectively.

Then, in step S49b, 2 images in said step S49a, extracting carry out rim detection respectively, and jump to step S49c.

In step S49c, at first, be based on the localization process result's (parallel amount of movement, rotation amount, ratio variable quantity) who obtains among the said step S45, edge of image conversion in the coordinate system of template image is estimated in 2 of will in said step S49b, obtain retrieval.And, as shown in Figure 8, overlap with the edge (c) of template image with 2 after conversion retrieval evaluation edge of image (a) and (b), extract inconsistent position (mobile rim).At this moment, the low part of decision threshold with the consistent degree ratio of 3 edge of image that overlap is stipulated proposes as mobile rim.At last, amount to 3, generate the coordinate figure list (mobile rim list) of mobile rim point respectively to 2 retrieval evaluation view data and template image data.At this moment, to the marginal point on the list, rim detection parameter value (minimum edge strength, edge ratio value) is carried out association.

Below, in step S49d, judge in the picture registration process result of said step S49c whether extract mobile rim (difference edge).And, extracting under the situation of mobile rim, jump to step S49e, under the situation that does not extract mobile rim, finish the precision parameter adjustment and handle.

In step S49e, be that minimum edge strength is adjusted to precision parameter, to eliminate mobile rim.To the minimum edge strength of template parameter,,, its value adjusts so that becoming big mode for difference edge deletion with template.In addition, with the minimum edge strength of parameter,, with through the marginal point that extracts is increased, make with the consistent degree at the edge of template image and improve so that the mode that its value diminishes adjusts to retrieval, thus the elimination mobile rim.In addition, the adjustment of minimum edge strength is the rim detection parameter value of storing in the mobile rim list through being based on, and with the strict 1 grade of mode that perhaps becomes loose 1 grade that becomes, from the candidate value shown in the above-mentioned table 1, selects and carries out.

Below, in step S49f,, the edge ratio value as precision parameter is adjusted for the mode that mobile rim is eliminated, finish the precision parameter adjustment and handle.To the edge ratio value of template parameter,,, its value adjusts so that becoming big mode for the difference edge of template is eliminated.In addition,,,, make the consistent degree raising with the edge of template image, thereby point of instability is eliminated with the edge ratio value of parameter to retrieval to increase through the marginal point that makes extraction so that the mode that its value diminishes adjusts.In addition, the adjustment of edge ratio value is the rim detection parameter value of storing in the mobile rim list through being based on, and with the strict 1 grade of mode that perhaps becomes loose 1 grade that becomes, from the candidate value shown in the above-mentioned table 1, selects and carries out.

Return Fig. 6, in step S50, image processing apparatus 12 increases progressively the retry count of the number of retries that is equivalent to the precision parameter adjustment, and jumps to said step S42.

In addition; In said step S48, reach predefined maximum times if be judged as the number of retries of precision parameter adjustment, then jump to step S51; The precision of retrieval process does not reach precision prescribed, and the template parameter of this timing is set as final parameter.And, generate template data based on this parameter, finish optimal parameter and obtain processing.

In step S52,12 pairs of statisticss that in said step S45, obtain of image processing apparatus are resolved, and to rim detection productive temp, coarse search productive temp, smart these 3 of the productive temps of retrieving, judge whether reach the target productive temp respectively.Here, the target productive temp that each is handled is through the detailed ratio to consider that the localization process algorithm is set, and distributes and sets requiring the productive temp value.And, reach in whole productive temps under the situation of target productive temp, successfully carry out obtaining of optimal parameter, jump to step S53.And in step S53, the template parameter that image processing apparatus 12 will this timing is set as final parameter, generate template data based on this parameter after, finish optimal parameter and obtain processing.

On the other hand, in said step S52, at least 1 productive temp in being judged as above-mentioned 3 productive temps does not reach under the situation of target productive temp, jumps to step S54, carries out the productive temp parameter adjustment and handles.

Fig. 9 is the flow chart that is illustrated in the productive temp parameter adjustment process flow of carrying out among the step S54.

At first, in step S54a, whether the rim detection productive temp is reached the target productive temp judge.And, under the situation that does not reach the target productive temp, be judged as and must carry out the adjustment of rim detection productive temp; And jump to step S54b; Under the situation that reaches the target productive temp, be not judged as and need carry out the adjustment of rim detection productive temp, the step S54c that states after jumping to.

In step S54b, carry out the adjustment of rim detection productive temp.The rim detection productive temp depends on the filtering size and the edge candidate is counted.Therefore, here, diminish, and the minimum edge strength of template parameter is increased, thereby the rim detection productive temp is adjusted to the direction that shortens through edge ratio value with template parameter.

The adjustment of each parameter is through to be benchmark with currency, with the strict 1 grade of mode that perhaps becomes loose 1 grade that becomes, from the candidate value shown in the above-mentioned table 1, selects to carry out.In addition, use parameter, can consider and as template parameter, strictly to adjust to retrieval.

Below, in step S54c, whether the coarse search productive temp is reached the target productive temp judge.And, under the situation that does not reach the target productive temp, be judged as and must carry out the adjustment of coarse search productive temp; Jump to step S54d; Under the situation that reaches the target productive temp, be not judged as and need carry out the adjustment of coarse search productive temp, the step S54e that states after jumping to.

In step S54d, carry out the adjustment of coarse search productive temp.Because coarse search is handled and to be based on characteristic point object is retrieved, and counts, is distance between characteristic point so the coarse search productive temp depends on characteristic.Therefore,, become big here, thereby, the coarse search productive temp is adjusted to the direction that shortens characteristic minimizings of counting through distance between the characteristic point that will set as template parameter.The adjustment of distance between characteristic point is to carry out with addition value through the adjustment that currency is added predefined regulation.In addition, at this moment, to consider apart from the mode that is no more than predefined higher limit between characteristic point and to adjust.

Then, in step S54e, whether essence retrieval productive temp is reached the target productive temp judge.And, under the situation that does not reach the target productive temp, be judged as the adjustment that must carry out smart retrieval productive temp; Jump to step S54f; Under the situation that reaches the target productive temp, be judged as the adjustment that need not carry out smart retrieval productive temp, the parameter adjustment of completed product run beat is handled.

In step S54f, carry out the adjustment of smart retrieval productive temp after, the parameter adjustment of completed product run beat is handled.Smart retrieval productive temp depends on the edge of in template data, logining and counts.Therefore,,, thereby essence is retrieved productive temp to the direction adjustment of shortening here through minimizing that the maximal margin of template parameter is counted.The adjustment that maximal margin is counted is to carry out with the subtraction value through the adjustment that deducts predefined regulation to currency.Its with marginal point on the edge of on the line equably the mode of configuration to handle the reduction of carrying out marginal point at interval of equal value by elongating, and through make elongate count at interval with 1,2,3 ... Order increase, thereby the stage make the edge minimizing of counting.

But, smart retrieval process as stated, through with template and searching object picture registration, carry out template repeatedly parallelly move, rotation is moved and ratio changes, thereby retrieve with retrieving images in the highest position of consistent degree of object.

When certain characteristic point is positioned near edge on the searching object image, from image energy (with the concentration data of image at x, on the y direction respectively 2 rank differential and obtain) receive power, attract to the edge.Power from this image energy receives is the parallel mobile vector of this characteristic point.Shown in figure 10; With the parallel mobile vector g of characteristic point, straight line (center of gravity normal) direction that links with the center of gravity with this characteristic point and template is a benchmark, when vertical composition and the decomposition of horizontal composition; Level becomes to be divided into to amplify dwindles vector h, and vertical one-tenth is divided into rotating vector v.In this execution mode; Through the correction coefficient that vector h utilizes regulation is respectively dwindled in parallel mobile vector g, rotating vector v and amplification; Be transformed to amount of movement (parallel amount of movement, the anglec of rotation, ratio variable quantity), thereby calculate the amount of movement of the template in 1 circular treatment.Amount of movement calculates based on following (2)~(4) formula.

m=「·γg·∑g………(2)

θ=「·γθ·∑v/∑Rd………(3)

scl=「·γscl·∑h/∑Rd………(4)

Here, m is that (unit: pixel), θ is that (unit: rad), scl is that (unit: multiplying power), Rd is a distance between center of gravity-characteristic point to the ratio variable quantity to the anglec of rotation to parallel amount of movement.In addition, Γ is that (default value: 0.2), γ g is that (default value: 1.0), γ θ is that (default value: 1.0), γ scl is for amplifying the correction coefficient (default value: 0.007) that dwindles for the correction coefficient of rotating to parallel mobile correction coefficient to whole correction coefficient.

And the energy summation of dwindling until parallel mobile vector g, rotating vector v and amplification after the big or small addition of vector h becomes till the convergence decision threshold that is less than or equal to regulation, carries out moving of template repeatedly.

Correction coefficient (Γ, γ g, γ θ, γ scl) when therefore, smart retrieval productive temp depends on the cycle-index of smart retrieval process and mobile vector is transformed to amount of movement.Therefore, here, through above-mentioned cycle-index is reduced, and above-mentioned correction coefficient is increased, thereby essence is retrieved productive temp to the direction adjustment of shortening.The adjustment of cycle-index and correction coefficient is through deducting or add that to currency the adjusted value of regulation carries out.This adjusted value and environment for use are set accordingly, and the mode that can not reach predefined critical value with cycle-index and correction coefficient is considered and adjusted.

Return Fig. 6, in step S55, image processing apparatus 12 is utilized in the template parameter of current timing setting, carries out the generation of template data.The template data that generates is stored in parameter storage part 25.

Then, in step S56, image processing apparatus 12 is utilized in the template data that generates among the said step S55, judges and whether finishes retrieval process to whole retrieval evaluation map pictures.And, being judged as under the situation that whole retrieval process do not have to finish, jump to step S57, be judged as under the situation that whole retrieval process finishes the step S60 that states after jumping to.

In step S57, image processing apparatus 12 is utilized in the template data that generates among the said step S55, and a plurality of retrieval evaluation map pictures for the object to same evaluation posture photographs carry out retrieval process respectively, obtain the localization process result.

Then, in step S58, the total process for producing beat in 12 pairs of retrieval process in said step S57 of image processing apparatus, rim detection productive temp, coarse search productive temp, smart these 4 of productive temps of retrieval are measured, and jump to step S59 then.

In step S59, image processing apparatus 12 is obtained the statistical disposition data, and is jumped to said step S56 to the localization process result who in said step S57, obtains.Here, for the centre coordinate of template (x, y), (x y), obtains mean value and standard deviation (3 σ) respectively to the solstics coordinate.In addition, also in said step S58, measure 4 productive temps, obtain average productive temp, minimum production beat, largest production beat respectively.

In addition; In step S60; Image processing apparatus 12 carries out following evaluation, that is, whether the standard deviation (3 σ) of standard deviation of the centre coordinate of the template of in said step S59, obtaining (3 σ) and solstics coordinate drops on respectively in the permissible range with the corresponding regulation of precision prescribed.

And in step S61, image processing apparatus 12 is judged as each standard deviation (3 σ) and drops under the situation in the permissible range in said step S60, proceeds the productive temp parameter adjustment and handles, and jumps to said step S52.On the other hand, under being judged as situation about dropping on outside the permissible range, jump to step S62, before productive temp reaches the target productive temp, make the precision of retrieval process be lower than precision prescribed, the template parameter of this timing is set as final parameter.And, generate template data based on this parameter, finish optimal parameter and obtain processing.

In addition, in Fig. 5, it is corresponding that step S31 and template image are obtained the unit, step S32~S34 with estimate that to obtain the unit with image corresponding.In addition, in Fig. 6, step S42~S50 is corresponding with the parameter setting unit, and step S52~S61 is corresponding with the parameter update unit.And in Fig. 7, step S49a is corresponding with extraction unit, and step S49b and S49c are corresponding with the mobile rim extraction unit, and step S49d~49f is corresponding with parameter gradual change unit.

(action)

Below, the action of the 2nd execution mode is described.

If parts batch change, then plant control unit 11 is notified this information to image processing apparatus 12.At this moment, plant control unit 11 also sends the precision prescribed of localization process and requires productive temp to image processing apparatus 12.

If image processing apparatus 12 receives the change notice of parts batch via interface 28 slave unit control device 11; Then identical ground with described the 1st execution mode; Precision and productive temp that image processing apparatus 12 measurement and positionings are handled, and with operation with memory 24 in the value of maintenance compare.And; The precision that measures and productive temp surpass permissible range with respect to the value that keeps in memory 24 in operation and situation about worsening under; Be judged as template data and the retrieval that reply stores and adjust in parameter storage part 25, carry out optimal parameter and obtain processing with parameter.

Obtain in the processing in optimal parameter, with template parameter and retrieval with parameter adjustment to best value.In this execution mode, template image is taken (the step S31 of Fig. 5), and under the shooting condition of regulation, a plurality of retrieval evaluation maps are looked like to take (step S32 and S33).And, use the template data of initial condition, look like to carry out retrieval process to a plurality of retrieval evaluation maps that photograph, measure repeatable accuracy, the precision of the localization process precision that whether meets the demands is judged (the step S43 of Fig. 6~S47).Under the situation of the precision that do not meet the demands,, adjust the parameter relevant (minimum edge strength, edge ratio) (step S49) with the precision of localization process for the precision that meets the demands.

Specifically; From a plurality of retrieval evaluation map pictures, extract 2 maximum retrieval evaluation map pictures (the step S49a of Fig. 7) of opposed edges variable quantity; Through with these 2 retrieval evaluation map pictures and template image coincidences, thereby extract mobile rim (step S49b and S49c).And; Minimum edge strength and edge ratio to template parameter are adjusted; Eliminating with the marginal portion of the corresponding template of mobile rim; And, retrieval is adjusted with the minimum edge strength and the edge ratio of parameter, to utilize retrieval process that mobile rim is stablized and can retrieve (step S49e, S49f).

If the adjustment of the parameter relevant with precision finishes, then carry out afterwards the adjustment of the parameter relevant with productive temp.Here; The productive temp productive temp that whether meets the demands is judged (the step S52 of Fig. 6); Under the situation of the productive temp that do not meet the demands; The parameter relevant with the productive temp of localization process (distance between minimum edge strength, edge ratio value, characteristic point, maximal margin counts, cycle-index, amount of movement correction coefficient) adjusted, with the productive temp that meets the demands (step S54).

Specifically, minimum edge strength and edge ratio value are changed gradually to the direction that the rim detection productive temp shortens, will guarantee that the critical value of precision prescribed is made as best parameter value.In addition, make the direction variation that distance shortens to the coarse search productive temp gradually between characteristic point, will guarantee that the critical value of precision prescribed is made as best parameter value.And, make that maximal margin is counted, cycle-index and amount of movement correction coefficient change to the direction that essence retrieval productive temp shortens, and will guarantee that the critical value of precision prescribed is made as best parameter value.As stated, with in the mode that can guarantee to carry out under the critical productive temp of precision prescribed rim detection, coarse search and smart retrieval, each parameter value is adjusted (step S55~S61).

(effect)

As stated, in above-mentioned the 2nd execution mode, under a plurality of shooting conditions of the excursion of representing the actual retrieval object images; The retrieval evaluation map is looked like to take, and the retrieval evaluation map picture based on photographing carries out parameter adjustment; Therefore, can generate the higher template data of reliability.Therefore, needn't just can template be launched other equipment adjustment.

In addition, when carrying out parameter adjustment, to more weak edge of contrast etc., become unsettled position, adjust parameter with the mode of from the object of retrieval process, removing according to the shooting condition edge detection results.Therefore, can the meet the demands parameter setting of the best of precision.

At this moment, the minimum edge strength in the template parameter is increased, perhaps the edge ratio value is dwindled, therefore when from template image, generating template data, can not extract mobile rim, and mobile rim is removed from the retrieval process object reliably.In addition, likewise, retrieval is reduced with the minimum edge strength in the parameter; The edge ratio value is increased; Therefore, can when from the searching object image, extracting the edge, not extract mobile rim, and can mobile rim be removed from the object of retrieval process reliably.

And, when carrying out parameter adjustment, can adjust to the direction that the productive temp that each is handled shortens respectively handling with edge detection process, coarse search and the relevant parameter of productive temp of smart retrieval process.Therefore, can further carry out best parameter adjustment.

At this moment, minimum edge strength is increased, perhaps the edge ratio value is dwindled, therefore can shorten the rim detection productive temp reliably.In addition, distance between characteristic point is increased, therefore, can reliably the coarse search productive temp be shortened.And with the maximal margin minimizing of counting, perhaps the cycle-index with smart retrieval process reduces, and the correction coefficient when perhaps will be in smart retrieval process mobile vector being transformed to amount of movement increases, and therefore, can reliably essence be retrieved productive temp and shorten.

Claims (14)

1. the adjusting device of shape basis match parameter, it is adjusted the parameter of in the matching treatment of shape basis, using, and it is characterized in that having:
Designating unit, the precision prescribed of its designated shape basis matching treatment and require productive temp;
Evaluation obtains the unit with image, and it will apply the image of the object of the posture after any variation for the benchmark posture, obtains with image as the evaluation of said parameter; And
The parameter adjustment unit, it uses image based on obtaining the said evaluation that obtains the unit by said evaluation with image, said parameter is adjusted,
Said parameter adjustment unit has:
The parameter setting unit, the said parameter of its automatic setting is so that the precision of shape basis matching treatment satisfies the said precision prescribed by said designating unit appointment; And
The parameter update unit; It is that productive temp does not satisfy when requiring productive temp by said designating unit appointment said utilizing the result who said evaluation is carried out shape basis matching treatment with image by the said parameter of said parameter setting unit automatic setting; The direction gradual change that the said parameter set by said parameter setting unit is shortened to said productive temp; On one side the precision of shape basis matching treatment being estimated, is the critical parameters value that said precision can be guaranteed said precision prescribed with said parameter update.
2. the adjusting device of shape according to claim 1 basis match parameter is characterized in that having:
Take the unit, it is taken said object; And
Template image is obtained the unit, and it utilizes said shooting unit that the said object of benchmark posture is taken, and the image of the said object that will photograph obtains as template image,
Said evaluation obtains the unit based on obtained the said template image that the unit is obtained by said template image with image, generates the correction image that has applied any variable quantity for this template image, and said correction image is obtained with image as said evaluation.
3. the adjusting device of shape according to claim 2 basis match parameter is characterized in that,
Said parameter setting unit has:
The fringe region extraction unit, it extracts the fringe region of being obtained the said template image of obtaining the unit by said template image; And
Rim detection parameter setting unit; It is based on the distribution of the local edge value in the said fringe region that is extracted by said fringe region extraction unit; With the parameter relevant in the said parameter, be set at the value that obtains best rim detection precision in predefined a plurality of parameter candidate value with the rim detection precision.
4. the adjusting device of shape according to claim 2 basis match parameter is characterized in that,
The precision of shape basis matching treatment is utilized absolute precision evaluation, and this absolute precision is to obtain said variable quantity that the unit applies to said template image as true value by said evaluation with image.
5. the adjusting device of shape according to claim 3 basis match parameter is characterized in that,
The precision of shape basis matching treatment is utilized absolute precision evaluation, and this absolute precision is to obtain said variable quantity that the unit applies to said template image as true value by said evaluation with image.
6. the adjusting device of shape according to claim 1 basis match parameter is characterized in that having:
Take the unit, it is taken said object; And
Template image is obtained the unit, and it utilizes said shooting unit that the said object of benchmark posture is taken, and the image of the said object that will photograph obtains as template image,
Said evaluation obtains the said shooting of unit by using unit with image, repeatedly takes being directed to the said object that the benchmark posture applied the posture of any variation, and a plurality of images that photograph are obtained with image as said evaluation.
7. the adjusting device of shape according to claim 6 basis match parameter is characterized in that,
Said parameter setting unit has:
Extraction unit, it extracts 2 maximum evaluations of opposed edges variable quantity and uses image from being obtained a plurality of evaluations under the same shooting condition of obtaining the unit with image by said evaluation with the image;
The mobile rim extraction unit, it extracts regulation is compared in 2 evaluations that extracted by said extraction unit with edge of image unanimity degree the low mobile rim of decision threshold; And
Parameter gradual change unit; It makes said parameter to the direction gradual change that will from the object of shape basis matching treatment, remove with the corresponding edge of mobile rim that is extracted by said mobile rim extraction unit on one side; On one side the precision of shape basis matching treatment is estimated, this precision is satisfied the parameter value of said precision prescribed and confirm as said parameter.
8. the adjusting device of shape according to claim 7 basis match parameter is characterized in that,
Said parameter comprises: template parameter, and it uses when utilizing said template image to generate the template data that is made up of the coordinate of the marginal point of said object and edge taper vector; And retrieve and use parameter, it uses when from the searching object image, detecting said object based on said template data,
Said parameter gradual change unit, the direction to will from said template data, removing with the corresponding edge of mobile rim that is extracted by said mobile rim extraction unit makes said template parameter gradual change.
9. the adjusting device of shape according to claim 7 basis match parameter is characterized in that,
Said parameter comprises: template parameter, and it uses when utilizing said template image to generate the template data that is made up of the coordinate of the marginal point of said object and edge taper vector; And retrieve and use parameter, it uses when from the searching object image, detecting said object based on said template data,
Said parameter gradual change unit, to be difficult to with the corresponding edge of mobile rim that extracts by said mobile rim extraction unit detected direction from said searching object image, make said retrieval use the parameter gradual change.
10. the adjusting device of shape according to claim 8 basis match parameter is characterized in that,
Said parameter comprises: template parameter, and it uses when utilizing said template image to generate the template data that is made up of the coordinate of the marginal point of said object and edge taper vector; And retrieve and use parameter, it uses when from the searching object image, detecting said object based on said template data,
Said parameter gradual change unit, to be difficult to with the corresponding edge of mobile rim that extracts by said mobile rim extraction unit detected direction from said searching object image, make said retrieval use the parameter gradual change.
11. the adjusting device according to each described shape basis match parameter in the claim 6 to 10 is characterized in that,
The precision of shape basis matching treatment is utilized repeatable accuracy evaluation, and this repeatable accuracy is carried out statistical disposition to the result who has carried out shape basis matching treatment with image to said a plurality of evaluations and obtained.
12. the adjusting device according to each described shape basis match parameter in the claim 1 to 11 is characterized in that,
Said shape basis matching treatment is carried out coarse search for the result who carries out edge detection process to the searching object image and is handled, and based on this coarse search process result, carries out smart retrieval process,
Said parameter update unit; On one side to said edge detection process, said coarse search are handled and the productive temp of said smart retrieval process shortens respectively direction; Make the said parameter gradual change of in this each processing, using; On one side the precision of shape basis matching treatment is estimated, said parameter update can be guaranteed the critical parameters value of said precision prescribed for this precision.
13. the method for adjustment of a shape basis match parameter, it is adjusted the parameter of in the matching treatment of shape basis, using, and it is characterized in that having following step:
The precision prescribed of designated shape basis matching treatment and require productive temp;
To apply the image of the object of the posture after any variation for the benchmark posture, obtain with image as the evaluation of said parameter;
Set said parameter, so that the precision of shape basis matching treatment satisfies said precision prescribed;
Use the said parameter of setting, carry out shape basis matching treatment to said evaluation with image, and measure productive temp; And
Do not satisfy said when requiring productive temp in the productive temp that measures; The direction that shortens to this productive temp on one side makes said parameter gradual change; On one side the precision of shape basis matching treatment is estimated, said parameter update can be guaranteed the critical parameters value of said precision prescribed for this precision.
14. an apparatus for mounting component, it utilizes adsorption mouth attract electrons parts, and the assigned position on substrate is installed this electronic unit, it is characterized in that having:
The adjusting device of each described shape basis match parameter in said claim 1 to 12;
Memory cell, it is stored in the parameter of using in the matching treatment of shape basis;
The searching object image is obtained the unit, and it is obtained the said image that photographs being taken by the image of the electronic unit of said adsorption mouth absorption as the searching object image;
Positioning unit, it is utilized in stored parameters in the said memory cell to obtaining the searching object image of obtaining the unit by said searching object image, carries out shape basis matching treatment, carries out the lift-launch location of said electronic unit;
Detecting unit, its change to parts batch detects; And
Judging unit, it judges whether and need adjust stored parameters in said memory cell when the change that is gone out said parts batch by said detection,
Be to adjust said parameter the time by said judgment unit judges, utilizing the adjusting device of said shape basis match parameter to carry out the adjustment of said parameter.
CN201110288632.9A 2010-09-24 2011-09-22 The shape basis adjusting device of match parameter, method of adjustment and apparatus for mounting component CN102421279B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2010214489A JP5759133B2 (en) 2010-09-24 2010-09-24 Shape-based matching parameter adjustment device and component mounting device
JP2010-214489 2010-09-24

Publications (2)

Publication Number Publication Date
CN102421279A true CN102421279A (en) 2012-04-18
CN102421279B CN102421279B (en) 2016-04-13

Family

ID=45945431

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201110288632.9A CN102421279B (en) 2010-09-24 2011-09-22 The shape basis adjusting device of match parameter, method of adjustment and apparatus for mounting component

Country Status (2)

Country Link
JP (1) JP5759133B2 (en)
CN (1) CN102421279B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103984362A (en) * 2013-02-07 2014-08-13 佳能株式会社 Position and orientation measuring apparatus, information processing apparatus and information processing method

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014002813A1 (en) * 2012-06-26 2014-01-03 コニカミノルタ株式会社 Image processing device, image processing method, and image processing program
JP6418739B2 (en) * 2014-01-09 2018-11-07 ヤマハ発動機株式会社 Evaluation device, surface mounter, evaluation method
JP6413071B2 (en) * 2014-07-17 2018-10-31 パナソニックIpマネジメント株式会社 Component mounting method and component mounting system
EP3404583A1 (en) 2017-05-19 2018-11-21 MVTec Software GmbH System and method for model adaptation

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002140693A (en) * 2000-10-31 2002-05-17 Keyence Corp Image processing parameter determination device, its method and recording medium with recorded image processing parameter determination program
JP2005217325A (en) * 2004-01-30 2005-08-11 Nikon Corp Reference pattern determination method and equipment therefor, position detection method and equipment therefor, and exposure method and equipment therefor
JP2007004829A (en) * 2006-09-11 2007-01-11 Mitsubishi Electric Corp Image processing device

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000260699A (en) * 1999-03-09 2000-09-22 Canon Inc Position detector and semiconductor aligner employing the same
JP2003042970A (en) * 2001-07-31 2003-02-13 Matsushita Electric Ind Co Ltd Visual inspection device and visual inspection method for mounting board
JP3834297B2 (en) * 2003-05-12 2006-10-18 ファナック株式会社 Image processing device
JP5160366B2 (en) * 2008-10-10 2013-03-13 Juki株式会社 Pattern matching method for electronic parts

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002140693A (en) * 2000-10-31 2002-05-17 Keyence Corp Image processing parameter determination device, its method and recording medium with recorded image processing parameter determination program
JP2005217325A (en) * 2004-01-30 2005-08-11 Nikon Corp Reference pattern determination method and equipment therefor, position detection method and equipment therefor, and exposure method and equipment therefor
JP2007004829A (en) * 2006-09-11 2007-01-11 Mitsubishi Electric Corp Image processing device

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103984362A (en) * 2013-02-07 2014-08-13 佳能株式会社 Position and orientation measuring apparatus, information processing apparatus and information processing method
US9704248B2 (en) 2013-02-07 2017-07-11 Canon Kabushiki Kaisha Position and orientation measuring apparatus, information processing apparatus and information processing method
CN103984362B (en) * 2013-02-07 2018-02-02 佳能株式会社 Position and orientation measurement apparatus, information processor and information processing method

Also Published As

Publication number Publication date
CN102421279B (en) 2016-04-13
JP5759133B2 (en) 2015-08-05
JP2012069003A (en) 2012-04-05

Similar Documents

Publication Publication Date Title
Gobakken et al. Comparing biophysical forest characteristics estimated from photogrammetric matching of aerial images and airborne laser scanning data
US9013576B2 (en) Aerial photograph image pickup method and aerial photograph image pickup apparatus
US8934709B2 (en) Dynamic object classification
US7804996B2 (en) Method for associating stereo image and three-dimensional data preparation system
CN103582119B (en) The finger print data base construction method of WiFi indoor locating system
CN104361770B (en) Precision approach autocontrol method for traffic information collection unmanned plane
TWI483215B (en) Augmenting image data based on related 3d point cloud data
KR20140076562A (en) Method and system for detecting and correcting problematic advanced process control parameters
JP6573354B2 (en) Image processing apparatus, image processing method, and program
US9445744B2 (en) Methods, systems, and devices for spine centrum extraction and intervertebral disk dividing
Deng et al. UAV-based multispectral remote sensing for precision agriculture: A comparison between different cameras
Udalski et al. The optical gravitational lensing experiment. final reductions of the OGLE-III Data
CN104704384A (en) Image processing method, particularly used in a vision-based localization of a device
CN105025440B (en) Indoor and outdoor scene detection method and equipment
CN104463117B (en) A kind of recognition of face sample collection method and system based on video mode
CN103792760B (en) The location Calculation of a kind of automatic focusing topworks and method for correcting position
US7835879B2 (en) Method and device for calibrating a magnetic sensor
JP5758728B2 (en) Charged particle beam equipment
US20050205780A1 (en) Scanning electron microscope and a method for evaluating accuracy of repeated measurement using the same
US10462351B2 (en) Fast auto-focus in imaging
DE102010029668A1 (en) Calibration of a three-axis magnetic field sensor
US7151258B2 (en) Electron beam system and electron beam measuring and observing methods
JP6188695B2 (en) Method and apparatus for classifying wrinkles using surface height attributes
US20120029812A1 (en) Method and system for automatically planning and scheduling a remote sensing satellite mission
WO2012164950A1 (en) Human attribute estimation system, human attribute estimation apparatus and human attribute estimation method

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
C14 Grant of patent or utility model