CN104020181A - Method for detecting integrated circuit material by nondestructive detecting apparatus - Google Patents

Method for detecting integrated circuit material by nondestructive detecting apparatus Download PDF

Info

Publication number
CN104020181A
CN104020181A CN201410224033.4A CN201410224033A CN104020181A CN 104020181 A CN104020181 A CN 104020181A CN 201410224033 A CN201410224033 A CN 201410224033A CN 104020181 A CN104020181 A CN 104020181A
Authority
CN
China
Prior art keywords
packing box
measured
gray
pixel
density
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201410224033.4A
Other languages
Chinese (zh)
Other versions
CN104020181B (en
Inventor
林文海
邱颖霞
闵志先
宋夏
刘炳龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CETC 38 Research Institute
Original Assignee
CETC 38 Research Institute
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
Application filed by CETC 38 Research Institute filed Critical CETC 38 Research Institute
Priority to CN201410224033.4A priority Critical patent/CN104020181B/en
Publication of CN104020181A publication Critical patent/CN104020181A/en
Application granted granted Critical
Publication of CN104020181B publication Critical patent/CN104020181B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention provides a method for detecting an integrated circuit material by a nondestructive detecting apparatus, the method is characterized in that ray emitted by a radiation source is passed through a material to be detected in a sealing box of an objective table, an image is formed on an induction imaging system, by comparing with an image formed by a standard material by naked eyes, or comparing the similarity of the standard image and the image of the material to be detected by a signal processing system, through the similarity and a preset threshold, an intelligent determination system is used for determining that under the condition of unsealing of the material, the automatic production requirement of an integrated circuit is satisfied or not. The method for detecting the integrated circuit material by the nondestructive detecting apparatus has the advantages that the ray which is nondestructive to the material and capable of penetrating the sealing box can be taken as a detection source, influences such as disturbance of placing position of a traditional unsealing detection to the material can be avoided, detection and determination of the material used for automatic production of the integrated circuit can be rapidly realized under unsealing condition at high efficiency; a generation link of the material can be simultaneously and accurately positioned, unclear responsibility due to problems during an examination link can be avoided.

Description

Method integrated circuit material being detected with nondestructive detecting apparatus
Technical field
The present invention relates to integrated circuit material inspection technology field, be specifically related to method integrated circuit material being detected with nondestructive detecting apparatus.
Background technology
It is the microdevice of being found and navigated to fixed area by chip mounter by its vision system that the robotization of the microdevice of high density integrated circuit mounts, and draws and mount on substrate from component cases.Need microdevice placing direction consistent, spacing rule.
Due to the needs of supply of material producer transport, the packing box of the microdevice of high density integrated circuit is generally possesses some strength and flexibility, can compress device, inside accommodates the cavity of a large amount of microdevices, and these requirements cause packing box conventionally cannot realize with transparent material.Due to the needs that robotization mounts, microdevice is conventionally put into array-like and is arranged, and putting position is had to accuracy requirement; And the type material arrives assembling and needs inspection when enterprise, with quantification and whether neatly put.The common method of inspection is to open box visual inspection by quality or product control personnel, confirms after errorless carefully to cover back again.But because integrated circuit automated production often needs the microdevice of enormous amount, this operation irregularity is heavy, in this process, also easily occur that misoperation causes microdevice to jump out cavity or hamper is not tight, in moving process, microdevice spills etc. problem from seam; More seriously due to exist these potential may, supplier easily when checking self human error to serve as reasons to shirk self packaging error or the problem of packing box own cause microdevice to be put not meeting the responsibility of automated production needs.And often little, the thin edge of precise structure, volume of microdevice itself is manually placed in pick process and tends to break or pollute device surface, waste or goods return and replacement difficulty are so just caused.
Existing producer starts to adopt nondestructive detecting apparatus to detect integrated circuit material for this reason, and the material that becomes shadow technology to observe package interior by ray is put situation, under the prerequisite of avoiding material to be broken or to pollute, carries out the examination of goods.But existing disposal route is: irradiate one by one the goods packing box on objective table by radiographic source, directly on display screen, manifest by induction image forming system the image receiving, carry out artificial judgment, the testing process of the method needs staff all the time, the efficiency of artificial cognition is low first, second the time of artificial cognition longer, be responsible for differentiate the easier fatigue of operator and erroneous judgement causes decrease in efficiency.As improvement, there is enterprise to be connected with signal processing system and intelligent decision system in induction image forming system rear end; By image recognition software, intelligent decision system is all identified and deposited in to the image of qualified goods and goods to be measured by signal processing system and carry out GTG comparison between two, judge that the arrangement in the arrangement in goods to be measured and qualified goods is consistent if the image gray-scale level of the image gray-scale level of goods to be measured and qualified goods is identical, thereby significantly improved the efficiency of intelligent decision.But because the imaging of goods exists the difference of the depth, and though the GTG of image cross dark or cross shallow all can cause intelligent decision system in the time judge easily to partially sternly or partially pine of the entirety judgement of goods, the precision judging is inadequate.If employing human assistance, cannot reach intelligence, quick, precisely and without the requirement of staff on duty, cannot meet the needs that enterprise produces.
Summary of the invention
In view of above content, the invention provides a kind of method integrated circuit material being detected with nondestructive detecting apparatus.Particular content is as follows:
Method integrated circuit material being detected with nondestructive detecting apparatus, described nondestructive detecting apparatus comprises: radiographic source 1, objective table 2, induction image forming system 4, signal processing system 5 and intelligent decision system 6.Wherein, on the end face of objective table 2, be provided with the travelling belt that horizontal transport object is used.Above the travelling belt of objective table 2, be provided with radiographic source 1, radiographic source 1 is one can send ray and penetrate the controlled ray emission pipe of box-packed material.Below the travelling belt of objective table 2, be provided with induction image forming system 4, induction image forming system 4 is corresponding with radiographic source 1, and the imaging surface area of induction image forming system 4 is greater than the ray that radiographic source 1 launches and impinges upon the area on objective table 2 workplaces.The signal output part of induction image forming system 4 is connected with the signal input part of signal processing system 5, and the signal output part of signal processing system 5 is connected with the signal input part of intelligent decision system 6.Induction image forming system 4 contains display.Induction image forming system 4 is responsible for the ray signal receiving to be converted into the gray level image on k rank, is presented on display and transfers to signal processing system 5.Signal processing system 5 comprises a storage unit, and the gray level image that signal processing system 5 is responsible for induction image forming system 4 to generate is extracted as density image feature, and deposits storage unit in.The density image feature that intelligent decision system 6 is responsible for that signal processing system 5 is extracted is carried out similarity relatively and is judged.The concrete grammar that intelligent decision system 6 is carried out nondestructive test is as described below:
Step 1: standard pack box is placed on objective table 2, and described standard pack box is that the packing box of piling up neat material is housed.Make radiographic source 1 produce sense radiation, and converge to the top of the packing box that block pattern row pattern material is housed, described sense radiation is wavelength coverage is less than 0.01nm to the X ray between 10nm or wavelength gamma-rays at 0.01nm.Sense radiation is radiated in induction image forming system 4 after passing the packing box that block pattern row pattern material is housed.Make induction image forming system 4 receive the density image Fig that is designated as standard material and packing box through the k rank gray level image that after the sense radiation signal of standard pack box, conversion generates 0, subsequently by the density image Fig of standard material and packing box 0be passed to signal processing system 5.Signal processing system 5 is by the density image Fig of the standard material receiving and packing box 0carry out density feature extraction, obtain the density image feature S of standard material and packing box 0.The density image characteristic image feature S of standard material and packing box 0be stored in the storage unit of signal processing system 5 and wait for and calling.Subsequently, the packing box that block pattern row pattern material is housed is taken off on objective table 2;
Step 2: manually set the tolerance A of intelligent decision system 6 and the gray scale gray scale rank tolerance j in judging.Wherein, gray scale rank tolerance j, refer to that when needing the gray-scale value of the pixel on image relatively and being less than numerical value j in same position and as the difference of the gray-scale value of the pixel on the image of standard of comparison, the pixel on this image that need to compare is considered as identical with the pixel on the image as standard of comparison.Tolerance A is gray scale image similarity threshold, refer to when accounting for while needing the number percent of pixel sum in image to be relatively greater than numerical value A be considered as identical pixel as standard of comparison image in the image needing relatively, judge this image that need to compare and the image similarity as standard of comparison;
Step 3: start the travelling belt of objective table 2, packing box to be measured is passed through from the below of radiographic source 1 one by one.
Whenever a packing box to be measured is in the time that the below of radiographic source 1 is passed through, generate corresponding density image Fig by induction image forming system 4 nand transfer to signal processing system 5, and wherein, N gets 1 to n, and the corresponding gray level image of first packing box to be measured is designated as the density image Fig of the first material to be measured and packing box 1, second corresponding gray level image of packing box to be measured is designated as the density image Fig of the second material to be measured and packing box 2, by that analogy, n the corresponding gray level image of packing box to be measured is designated as the density image Fig of n material to be measured and packing box n.
Induction image forming system 4 is transmitted next material to be measured and the density image Fig of packing box by signal processing system 5 ncarry out one by one density feature extraction, obtain the density feature image S of material to be measured and packing box n, be followed successively by the density feature image S of the first material and packing box 1, the second material and packing box density feature image S 2..., n material and packing box density feature image S n.The density feature image Fig of a said n material to be measured and packing box nexist in signal processing system 5 and wait for and calling.
Intelligent decision system 6 is by the density feature image S of the material to be measured of signal processing system 5 interior storages and packing box nthe standard material obtaining with step 1 respectively and the density image feature S of packing box 0carry out similarity threshold comparison:
3.1 by the density feature image S of material to be measured and packing box ndirectly and the density image feature S of standard material and packing box 0relatively similarity.The material to be measured of same position and the density feature image S of packing box will be positioned at nin the gray-scale value of pixel and the density image feature S of standard material and packing box 0in the gray-scale value of pixel do poorly, if when difference is less than the gray scale rank tolerance j setting in step 2, judge above-mentioned two pixels (density feature image S of material to be measured and packing box npixel and standard material in correspondence position and the density image feature S of packing box 0pixel) put identical.Otherwise, judge that above-mentioned two pixels are not identical.
By the density feature image S of material to be measured and packing box nin the gray-scale value of pixel and the density image feature S of the standard material of correspondence position and packing box 0in the gray-scale value of pixel compare one by one, the density image feature S of note and standard material and packing box 0the density feature image S of identical material to be measured and packing box ninterior pixel accounts for the density feature image S of material to be measured and packing box nthe percent value of pixel sum is similarity threshold: in the time that this similarity threshold is greater than the tolerance A setting in step 2, judge the density image feature S of this material to be measured and packing box nin corresponding packing box to be measured material put the requirement that meets reception, exit the judgement comparison to this packing box to be measured, carry out the judgement comparison of packing box next to be measured.Otherwise, enter next step.
3.2 taking 1 rank as span, heighten the density feature image S of material to be measured and packing box ngTG, by the comparative approach of step 3.1 again with the density image feature S of standard material and packing box 0relatively similarity.
If heighten material to be measured after gray-scale value and the density image feature S of packing box nsimilarity threshold while being greater than the tolerance A setting in step 2, judge that the material in corresponding packing box to be measured meets the requirements, exit the judgement comparison to this packing box to be measured, carry out subsequently the judgement comparison of packing box next to be measured.
If the density image feature S of material to be measured and packing box ngTG value while heightening to 255 rank, heighten the material to be measured of gray-scale value and the density image feature S of packing box nsimilarity threshold be still not more than the tolerance A setting in step 2, enter next step.
3.3. taking 1 rank as span, turn down the density feature image S of material to be measured and packing box ngTG, by the comparative approach of step 3.1 again with the density image feature S of standard material and packing box 0relatively similarity.
If turn down the material to be measured of gray-scale value and the density image feature S of packing box nbe greater than with the similarity threshold of standard material the tolerance A setting in step 2, judge that in corresponding packing box to be measured, putting of material meets reception requirement, exit the judgement comparison to this packing box to be measured, carry out subsequently the judgement comparison of packing box next to be measured.
If the density image feature S of material to be measured and packing box ngTG value GTG value while being transferred to 1 rank, turn down the material to be measured of gray-scale value and the density image feature S of packing box nsimilarity threshold is not more than the tolerance A setting in step 2, judges the density image feature S of this material to be measured and packing box nthe material in corresponding packing box there is a volume defect or arrange undesirable, intelligent decision system 6 records the serial number of this packing box to be measured and points out this material to belong to unacceptable product, exit the judgement comparison to this packing box to be measured, carry out subsequently the judgement comparison of packing box next to be measured.
useful technique effect
When the present invention utilizes ray to pass the material on objective table, because of the method that element forms or density variation is carried out imaging and differentiation of material, to the undisturbed of material own and damage.The imaging link of induction image forming system 4 of the present invention and the identification link of intelligent decision system 6 are not only without manual operation, the more important thing is, this method has increased on the basis of existing technology goods image gray-scale level to be measured and has had the relatively judgement under GTG depth difference condition, the image gray-scale level of having avoided adopting single standard packing box is evaluated criterion and the erroneous judgement that causes the most, reduce the frequency of manually checking, guarantee efficiency and precision, be conducive to tackle fast and efficiently the also material of needs inspection of enormous amount in automated production process.
Open compared with box inspection technology with existing, the inventive method has overcome the impact of the GTG depth, without staff on duty, without evaluation criterion is repeatedly proofreaded and adjusted, outstanding feature of the present invention is, in the time that intelligence automatically judges that whether material to be measured is qualified, the overall intensity difference being caused for light path by intelligent decision system (conventionally adopting an industrial computer) has and repeatedly compensates recognition function, can effectively avoid erroneous judgement.Do not need to open box inspection, can not close box and cause the material change in location of arranging because holding box, then cause the identification to automated machine to cause difficulty.Manufacturing enterprise does not open the material image that in box situation, inspection obtains, responsibility investigation clear and definite after supplier provides material simultaneously.
Carry out image gray-scale level while comparing to determine in intelligent decision system, when continuing the GTG entirety of packing box to be detected of one batch higher or lower than the image gray-scale level of standard pack box, pass through the display screen display reminding information of induction image forming system 4 by intelligent decision system, under operating personnel's instruction, the travelling belt of intelligent decision system drive objective table rises or declines, to improve identification efficiency and accuracy.
Brief description of the drawings
Fig. 1 is the simple view of nondestructive detecting apparatus used in the present invention.
Fig. 2 is the process flow diagram of the inventive method.
Fig. 3 is when substandard product detected, by the shown package interior condition diagram of induction image forming system 4.
sequence number in figure is:radiographic source 1, objective table 2, box-packed material 3, induction image forming system 4, signal processing system 5, intelligent decision system 6.
Embodiment
With reference to the accompanying drawings, the invention will be further described in conjunction with specific embodiments.
Referring to Fig. 1, with nondestructive detecting apparatus, integrated circuit material is carried out the method for Non-Destructive Testing, described nondestructive detecting apparatus comprises: radiographic source 1, objective table 2, induction image forming system 4, signal processing system 5 and intelligent decision system 6.Wherein, on the end face of objective table 2, be provided with the travelling belt that horizontal transport object is used.Above the travelling belt of objective table 2, be provided with radiographic source 1, radiographic source 1 is one can send ray and penetrate the controlled ray emission pipe of box-packed material.Below the travelling belt of objective table 2, be provided with induction image forming system 4, induction image forming system 4 is corresponding with radiographic source 1, and the imaging surface area of induction image forming system 4 is greater than the ray that radiographic source 1 launches and impinges upon the area on objective table 2 workplaces.The signal output part of induction image forming system 4 is connected with the signal input part of signal processing system 5, and the signal output part of signal processing system 5 is connected with the signal input part of intelligent decision system 6.Induction image forming system 4 contains display.Induction image forming system 4 is responsible for the ray signal receiving to be converted into the gray level image on k rank, is presented on display and transfers to signal processing system 5.
Signal processing system 5 comprises a storage unit, and the gray level image that signal processing system 5 is responsible for induction image forming system 4 to generate is extracted as density image feature, and deposits storage unit in.The density image feature that intelligent decision system 6 is responsible for that signal processing system 5 is extracted is carried out similarity relatively and is judged.
Referring to Fig. 2, the concrete grammar that intelligent decision system 6 is carried out nondestructive test as described below:
Step 1: standard pack box is placed on objective table 2, and described standard pack box is that the packing box of piling up neat material is housed;
Make radiographic source 1 produce sense radiation, and converge to the top of the packing box that block pattern row pattern material is housed, described sense radiation is wavelength coverage is less than 0.01nm to the X ray between 10nm or wavelength gamma-rays at 0.01nm; Sense radiation is radiated in induction image forming system 4 after passing the packing box that block pattern row pattern material is housed; Make induction image forming system 4 receive the density image Fig that is designated as standard material and packing box through the k rank gray level image that after the sense radiation signal of standard pack box, conversion generates 0, subsequently by the density image Fig of standard material and packing box 0be passed to signal processing system 5;
Signal processing system 5 is by the density image Fig of the standard material receiving and packing box 0carry out density feature extraction, obtain the density image feature S of standard material and packing box 0.The density image characteristic image feature S of standard material and packing box 0be stored in the storage unit of signal processing system 5 and wait for and calling.Subsequently, the packing box that block pattern row pattern material is housed is taken off on objective table 2;
Step 2: manually set the tolerance A of intelligent decision system 6 and the gray scale gray scale rank tolerance j in judging.Wherein, gray scale rank tolerance j, refer to that when needing the gray-scale value of the pixel on image relatively and being less than numerical value j in same position and as the difference of the gray-scale value of the pixel on the image of standard of comparison, the pixel on this image that need to compare is considered as identical with the pixel on the image as standard of comparison.Tolerance A is gray scale image similarity threshold, refer to when accounting for while needing the number percent of pixel sum in image to be relatively greater than numerical value A be considered as identical pixel as standard of comparison image in the image needing relatively, judge this image that need to compare and the image similarity as standard of comparison;
Step 3: start the travelling belt of objective table 2, packing box to be measured is passed through from the below of radiographic source 1 one by one.
Whenever a packing box to be measured is in the time that the below of radiographic source 1 is passed through, generate corresponding density image Fig by induction image forming system 4 nand transfer to signal processing system 5, and wherein, N gets 1 to n, and the corresponding gray level image of first packing box to be measured is designated as the density image Fig of the first material to be measured and packing box 1, second corresponding gray level image of packing box to be measured is designated as the density image Fig of the second material to be measured and packing box 2, by that analogy, n the corresponding gray level image of packing box to be measured is designated as the density image Fig of n material to be measured and packing box n.
Induction image forming system 4 is transmitted next material to be measured and the density image Fig of packing box by signal processing system 5 ncarry out one by one density feature extraction, obtain the density feature image S of material to be measured and packing box n, be followed successively by the density feature image S of the first material and packing box 1, the second material and packing box density feature image S 2..., n material and packing box density feature image S n.The density feature image Fig of a said n material to be measured and packing box nexist in signal processing system 5 and wait for and calling.
Intelligent decision system 6 is by the density feature image S of the material to be measured of signal processing system 5 interior storages and packing box nrespectively with the density image feature S of standard material and packing box 0carry out similarity threshold comparison:
3.1 by the density feature image S of material to be measured and packing box ndirectly and the density image feature S of standard material and packing box 0relatively similarity.The material to be measured of same position and the density feature image S of packing box will be positioned at nin the gray-scale value of pixel and the density image feature S of standard material and packing box 0in the gray-scale value of pixel do poorly, if when difference is less than the gray scale rank tolerance j setting in step 2, judge the density feature image S of this material to be measured and packing box npixel and standard material in correspondence position and the density image feature S of packing box 0pixel identical.Otherwise, judge that above-mentioned two pixels are not identical; By the density feature image S of material to be measured and packing box nin the gray-scale value of pixel and the density image feature S of the standard material of correspondence position and packing box 0in the gray-scale value of pixel compare one by one, the density image feature S of statistics and standard material and packing box 0the density feature image S of identical material to be measured and packing box nin pixel account for the density feature image S of material to be measured and packing box nthe number percent of pixel sum: in the time that this number percent is greater than the tolerance A setting in step 2, judge the density image feature S of this material to be measured and packing box nmaterial in corresponding packing box to be measured meets the requirements, and exits the judgement comparison to this packing box to be measured, carries out subsequently the judgement comparison of packing box next to be measured.Otherwise, enter next step.
3.2 heighten the density feature image S of material to be measured and packing box ngTG, then with the density image feature S of standard material and packing box 0relatively similarity.
Taking 1 rank as span, by the density image feature S of material to be measured and packing box nthe GTG value of interior whole pixel is that benchmark is successively heightened, and will heighten the material to be measured of pixel GTG value and the density image feature S of packing box by the comparative approach of step 3.1 nagain with the density image feature S of standard material and packing box 0compare:
If heighten material to be measured after gray-scale value and the density image feature S of packing box ndensity image feature S with standard material and packing box 0similarity threshold while being greater than the tolerance A setting in step 2, judge the density image feature S of this material to be measured and packing box nmaterial in corresponding packing box to be measured meets the requirements, and exits the judgement comparison to this packing box to be measured, carries out subsequently the judgement comparison of packing box next to be measured.
If the density image feature S of material to be measured and packing box ngTG value while heightening to 255 rank, heighten the material to be measured of gray-scale value and the density image feature S of packing box ndensity image feature S with standard material and packing box 0similarity threshold be still less than the tolerance A setting in step 2, enter next step.
3.3. turn down the density feature image S of material to be measured and packing box ngTG, then with the density image feature S of standard material and packing box 0relatively similarity.
Taking 1 rank as span, by the density image feature S of material to be measured and packing box nthe GTG value of interior whole pixel is that benchmark is successively turned down, and will heighten the material to be measured of pixel GTG value and the density image feature S of packing box by the comparative approach of step 3.1 nagain with the density image feature S of standard material and packing box 0compare:
If turn down the material to be measured of gray-scale value and the density image feature S of packing box ndensity image feature S with standard material and packing box 0similarity threshold be greater than the tolerance A setting in step 2, judge the density image feature S of this material to be measured and packing box nmaterial in corresponding packing box to be measured meets the requirements, and exits the judgement comparison to this packing box to be measured, carries out subsequently the judgement comparison of packing box next to be measured.
If the density image feature S of material to be measured and packing box ngTG value GTG value while being transferred to 1 rank, turn down the material to be measured of gray-scale value and the density image feature S of packing box ndensity image feature S with standard material and packing box 0similarity threshold be still less than the tolerance A setting in step 2, judge the density image feature S of this material to be measured and packing box nthere is a volume defect or arrange undesirable in the material in corresponding packing box to be measured, intelligent decision system 6 records the serial number of this packing box to be measured and points out this material to belong to unacceptable product, exit the judgement comparison to this packing box to be measured, carry out subsequently the judgement comparison of packing box next to be measured.
Furtherly, the travelling belt on objective table 2 end faces rises or declines under intelligent decision system 6 drives, thus the density image Fig of adjustment criteria material and packing box 0and the density image Fig of material to be measured and packing box ncontrast.Travelling belt ensures that by moving up and down the radial energy that sends of radiographic source 1 is all the time to box-packed material focusing to be measured.
Furtherly, there is a volume defect or arrange undesirable in material in intelligent decision system 6 is pointed out monitored packing box to be measured, whether the gray level image of corresponding packing box to be measured is shown by the display of induction image forming system 4, turn by the existing individuality of material or the defect of arranging in this packing box to be measured of artificial judgment and can accept.
Furtherly, place in packing box by neat discharge, the box-packed or band dress ic component that can identify and pick up by vision system, mount for automatic machinery.
Furtherly, ic component is the box-packed silicon bare chip of wafer, SiGe bare chip, gallium arsenide bare chip, chip capacity or miniature inductance.
Furtherly, standard material induction image forming system 4 being transmitted by knowledge figure software signal disposal system 5 and the density image of packing box carry out position mark and gray scale mark to each pixel respectively, due to the density image Fig of standard material and packing box 0contain x*y pixel, contain an x capable y row standard material of pixel and the density image Fig of packing box therefore signal processing system 5 reads 0in the gray-scale value of each pixel, this x*y gray-scale value is according to pixels put to first left and then right, the first up and then down order in position and arranges, form the density image feature S of standard material and packing box 0=S 0-11,s 0-12 ...,s 0-1y,s 0-21,s 0-22 ...,s 0-2y ...,s 0-xy, S 0-xyrepresent to be positioned at the density image Fig of standard material and packing box 0the gray-scale value of the capable y row of x pixel.Subsequently, the density image feature S of standard material and packing box 0be stored in signal processing system 5 stand-by.
Signal processing system 5 will contain an x capable y row material to be detected of pixel and the density feature image S of packing box nby the sequencing numbering of radiographic source 1 irradiation area, read the density feature image Fig of each material to be detected and packing box by corresponding box to be packaged nin each pixel gray-scale value and arrange by first left and then right, first up and then down sequence of positions by this x*y pixel, form the density image feature S of material to be detected and packing box n=S n-11,s n-12 ...,s n-1y,s n-21,s n-22 ...,s n-2y ...,s n-xy, S n-xyrepresent the density image Fig of N standard material and packing box nin be positioned at the gray-scale value of the capable y row of x pixel, N gets 1 to n.Wherein, the density feature image Fig of the first material and packing box 1in the gray-scale value of each pixel be designated as successively S 1=S 1-11,s 1-12 ...,s 1-1y,s 1-21,s 1-22 ...,s 1-2y ...,s 1-xy, the density feature image Fig of the second material and packing box 2in the gray-scale value of each pixel be designated as successively S 2=S 2-11,s 2-12 ...,s 2-1y,s 2-21,s 2-22 ...,s 2-2y ...,s 2-xy, by that analogy, the density feature image Fig of n material and packing box nin the gray-scale value of each pixel be designated as successively S n=S n-11,s n-12 ...,s n-1y,s n-21,s n-22 ...,s n-2y ...,s n-xy, the density image feature S of said n group material to be detected and packing box nbe stored in signal processing system 5 and wait for and calling.
Furtherly, tolerance A is gray scale image similarity threshold, and tolerance A is the density feature image S of material to be measured and packing box nin the density variation characteristic image S of each pixel gray scale and standard material and packing box 0in the identical number percent of each pixel gray scale.
Furtherly, gray scale rank tolerance j refers to the density feature image S of material to be measured and packing box ndensity image feature S with standard material and packing box 0in carry out similarity threshold relatively time, when being positioned at the material to be measured of same position and the density feature image S of packing box nin the gray-scale value of pixel and the density image feature S of standard material and packing box 0in the difference of gray-scale value of pixel while being less than j, judge that the GTG value of two pixels in correspondence position is identical.
Furtherly, the judgement detailed step of intelligent decision system 6 is as follows:
Step 3.1 is by the density image feature S of standard material and packing box 0in the gray-scale value S of each pixel 0-11,s 0-12 ...,s 0-1y,s 0-21,s 0-22 ...,s 0-2y ...,s 0-xydensity image feature S with material to be measured and packing box nin the gray-scale value S of each pixel n-11,s n-12 ...,s n-1y,s n-21,s n-22 ...,s n-2y ...,s n-xycontrast respectively: S 0-11with S n-11difference be less than gray scale rank tolerance j this to pixel S 0-11with S n-11equate S 0-12with S n-12difference be less than gray scale rank tolerance j this to pixel S 0-12with S n-12equate, by that analogy, the number percent of the pixel that statistics gray scale is identical, if the identical ratio of pixel gray scale exceedes tolerance A, thinks that the interior material loading of this packing box is qualified, carries out the detection of next packing box.Otherwise, enter next step.
Step 3.2 is by the density image feature S of material to be measured and packing box nin all pixel gray-scale value Integral liftings 1 rank, by the method for step 3.1 again with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box.
Otherwise, by the density image feature S of material to be measured and packing box nin all pixel gray-scale value Integral liftings 2 rank, by the method for step 3.1 again with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box.
By that analogy, successively improve the density image feature S of material to be measured and packing box as span taking 1 rank nin all pixels gray-scale value and compare by the method for step 3.3, if the number percent of identical pixel is greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box.
As the density image feature S of material to be measured and packing box nin the gray-scale value of all pixels while being promoted to the high-order 255 of gray level image, if the percentage of identical pixel is still less than tolerance A, enter next step.
Step 3.3 is by the density image feature S of material to be measured and packing box nin the gray-scale value of all pixels reduce by 1 rank, then with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box.
Otherwise, by the density image feature S of material to be measured and packing box nin all pixel gray-scale values entirety reduce by 2 rank, by the method for step 3.1 again with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box.
By that analogy, successively reduce the density image feature S of material to be measured and packing box as span taking 1 rank nin all pixels gray-scale value and compare by the method for step 3.1, if the number percent of identical pixel is greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box.
If as the density image feature S of material to be measured and packing box nin the gray-scale value of all pixels while being reduced to the lowest-order 1 of gray level image, if the number percent of identical pixel is still less than tolerance A, judge this this material S to be measured nmaterial is defective material, has volume defect or arranges undesirablely, carries out the detection of next packing box.
In other words, intelligent decision system 6 arranges gray scale image similarity threshold A, and A is the density feature image S of material to be measured and packing box nrespectively with the density variation characteristic image S of standard material and packing box 0in the identical number percent of each pixel gray scale, the density variation characteristic image S of the accurate material of bidding and packing box 0for k rank gray scale, when gray difference thinks that in j order range the identical j of two pixel grey scales is much smaller than k.By the density feature image S of the material to be measured of signal processing system 5 interior storages and packing box nrespectively with the density image feature S of standard material and packing box 0carry out similarity threshold comparison: first by the gray scale S of the each pixel of density variation characteristic image of standard material and packing box 0-11,s 0-12s 0-1ys 0-xywith material S to be measured nin the gray scale S of each pixel n-11,s n-12s n-1ys n-xycontrast respectively, if the identical ratio of pixel gray scale exceedes A, think that material to be measured is qualified.As the density image feature S of material to be measured and packing box ndensity image feature S with standard material and packing box 0similarity while being less than A, by S nin former figure, all pixel gray scale Integral liftings 1 rank, compare again, as similarity is greater than A, judge this material S to be measured nfor qualified material.As being still less than A, similarity again promotes S nmiddle gray scale 1 rank, compare repeatedly.In the time being promoted to the high-order k of gray scale, similarity is still less than A, at S nthe all pixel gray scales of former figure reduce by 1 rank, again compare.In the time being reduced to gray scale lowest-order 1, similarity is still less than A, judges this this material S to be measured nmaterial is defective material, has a volume defect or arranges undesirable.Intelligent decision system 6 records the sequence number of this packing box to be measured and points out this material to belong to unacceptable product.
Fig. 3, for there being partial material incomplete or put at randomly, does not meet robotization and criticizes the sectional drawing of the material of production material standard.

Claims (4)

1. the method with nondestructive detecting apparatus, integrated circuit material being detected, described nondestructive detecting apparatus comprises radiographic source (1), objective table (2), induction image forming system (4), signal processing system (5) and intelligent decision system (6); Be provided with radiographic source (1) in the top of the travelling belt of objective table (2); Be provided with induction image forming system (4) in the below of objective table (2); Induction image forming system (4) is connected with intelligent decision system (6) through signal processing system (5); It is characterized in that, it is specific as follows that intelligent decision system (6) is carried out the method for nondestructive test:
Step 1: standard pack box is placed on objective table (2); Start radiographic source (1) and produce sense radiation; The gray level image that induction image forming system (4) generates the sense radiation conversion receiving, this gray level image is designated as the density image Fig of standard material and packing box 0; Signal processing system (5) is by the density image Fig of standard material and packing box 0carry out density feature extraction, obtain the density image feature S of standard material and packing box 0;
Step 2: manually set tolerance A and gray scale rank tolerance j, and input intelligent decision system (6);
Step 3: packing box to be measured is passed through from the below of radiographic source (1) one by one;
Whenever a packing box to be measured is in the time that the below of radiographic source (1) is passed through, generated the density image Fig of material to be measured and packing box by induction image forming system (4) n, by signal processing system (5) by the density image Fig of material to be measured and packing box ncarry out density feature extraction, obtain the density feature image S of material to be measured and packing box n; Intelligent decision system (6) is by the density feature image S of material to be measured and packing box nthe standard material obtaining with step 1 respectively and the density image feature S of packing box 0carry out similarity threshold comparison:
3.1 by the density feature image S of material to be measured and packing box ndirectly and the density image feature S of standard material and packing box 0relatively similarity: will be positioned at the material to be measured of same position and the density feature image S of packing box nin the gray-scale value of pixel and the density image feature S of standard material and packing box 0in the gray-scale value of pixel do poorly, if difference is less than gray scale rank tolerance j, judge that above-mentioned two pixels are identical; Otherwise, judge not identical;
The density image feature S of note and standard material and packing box 0the density feature image S of identical material to be measured and packing box ninterior pixel accounts for the density feature image S of material to be measured and packing box nthe percent value of pixel sum is similarity threshold; When described similarity threshold is greater than tolerance A, judge that in corresponding packing box to be measured, putting of material meets reception requirement, detects next packing box; Otherwise, enter next step;
3.2 taking 1 rank as span, heighten the density feature image S of material to be measured and packing box ngTG, by the comparative approach of step 3.1 again with the density image feature S of standard material and packing box 0relatively similarity: if the similarity threshold of heightening after gray-scale value is greater than tolerance A, judge that the material of putting in corresponding packing box to be measured is qualified, detects next packing box to be measured; If the density image feature S of material to be measured and packing box nthe similarity threshold of GTG value while heightening to 255 rank be still not more than tolerance A, enter next step;
Taking 1 rank as span, turn down the density feature image S of material to be measured and packing box ngTG, by the comparative approach of step 3.1 again with the density image feature S of standard material and packing box 0relatively similarity: if the similarity threshold of turning down after gray-scale value is greater than tolerance A, judge that the material of putting in corresponding packing box to be measured is qualified, detect next packing box to be measured; If the density image feature S of material to be measured and packing box nthe similarity threshold of GTG value GTG value while being transferred to 1 rank be still not more than tolerance A, judge that the material of putting in corresponding packing box to be measured exists a volume defect or arranges undesirable, intelligent decision system (6) records the serial number of this packing box to be measured and points out this material to belong to unacceptable product, detects the next packing box to be measured of row.
2. method integrated circuit material being detected with nondestructive detecting apparatus as claimed in claim 1, is characterized in that: signal processing system (5) reads and contains an x capable y row standard material of pixel and the density image Fig of packing box 0in the gray-scale value of each pixel, this x*y gray-scale value is according to pixels put to first left and then right, the first up and then down order in position and arranges, form the density image feature S of standard material and packing box 0=(S 0-11,s 0-12 ...,s 0-1y,s 0-21,s 0-22 ...,s 0-2y ...,s 0-xy), S 0-xyrepresent to be positioned at the density image Fig of standard material and packing box 0the gray-scale value of the capable y row of x pixel; Subsequently, the density image feature S of standard material and packing box 0be stored in signal processing system (5) stand-by;
Signal processing system (5) will contain an x capable y row material to be detected of pixel and the density feature image S of packing box nby the sequencing numbering of radiographic source (1) irradiation area, read the density feature image Fig of each material to be detected and packing box by corresponding box to be packaged nin each pixel gray-scale value and arrange by first left and then right, first up and then down sequence of positions by this x*y pixel, form the density image feature S of material to be detected and packing box n=(S n-11,s n-12 ...,s n-1y,s n-21,s n-22 ...,s n-2y ...,s n-xy), S n-xyrepresent the density image Fig of N standard material and packing box nin be positioned at the gray-scale value of the capable y row of x pixel, N gets 1 to n; The density feature image Fig of n material and packing box nin the gray-scale value of each pixel be designated as successively S n=(S n-11,s n-12 ...,s n-1y,s n-21,s n-22 ...,s n-2y ...,s n-xy), the density image feature S of said n group material to be detected and packing box nbe stored in signal processing system (5) and wait for and calling.
3. method integrated circuit material being detected with nondestructive detecting apparatus as claimed in claim 2, is characterized in that: the judgement detailed step of intelligent decision system (6) is as follows:
Step 3.1 is by the density image feature S of standard material and packing box 0in the gray-scale value S of each pixel 0-11,s 0-12 ...,s 0-1y,s 0-21,s 0-22 ...,s 0-2y ...,s 0-xydensity image feature S with material to be measured and packing box nin the gray-scale value S of each pixel n-11,s n-12 ...,s n-1y,s n-21,s n-22 ...,s n-2y ...,s n-xycontrast respectively: S 0-11with S n-11difference be less than gray scale rank tolerance j this to pixel S 0-11with S n-11equate S 0-12with S n-12difference be less than gray scale rank tolerance j this to pixel S 0-12with S n-12equate, by that analogy, the number percent of the pixel that statistics gray scale is identical, if the identical ratio of pixel gray scale exceedes tolerance A, thinks that the interior material loading of this packing box is qualified, carries out the detection of next packing box; Otherwise, enter next step;
Step 3.2 is by the density image feature S of material to be measured and packing box nin all pixel gray-scale value Integral liftings 1 rank, by the method for step 3.1 again with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box;
Otherwise, by the density image feature S of material to be measured and packing box nin all pixel gray-scale value Integral liftings 2 rank, by the method for step 3.1 again with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box;
By that analogy, successively improve the density image feature S of material to be measured and packing box as span taking 1 rank nin all pixels gray-scale value and compare by the method for step 3.3, if the number percent of identical pixel is greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box;
As the density image feature S of material to be measured and packing box nin the gray-scale value of all pixels while being promoted to the high-order 255 of gray level image, if the percentage of identical pixel is still less than tolerance A, enter next step;
Step 3.3 is by the density image feature S of material to be measured and packing box nin the gray-scale value of all pixels reduce by 1 rank, then with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box;
Otherwise, by the density image feature S of material to be measured and packing box nin all pixel gray-scale values entirety reduce by 2 rank, by the method for step 3.1 again with the density image feature S of standard material and packing box 0in the gray-scale value of each pixel compare, as the number percent of identical pixel is now greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box;
By that analogy, successively reduce the density image feature S of material to be measured and packing box as span taking 1 rank nin all pixels gray-scale value and compare by the method for step 3.1, if the number percent of identical pixel is greater than tolerance A, think that the material loading in this packing box is qualified, carry out the detection of next packing box;
If as the density image feature S of material to be measured and packing box nin the gray-scale value of all pixels while being reduced to the lowest-order 1 of gray level image, if the number percent of identical pixel is still less than tolerance A, be judged to be defective material, there is a volume defect or arrange undesirable.
4. method integrated circuit material being detected with nondestructive detecting apparatus as claimed in claim 1, is characterized in that: the span of tolerance A is between 90.0% to 100%, and the span of gray scale rank tolerance j is between 1 to 60.
CN201410224033.4A 2014-05-26 2014-05-26 Method integrated circuit material detected with nondestructive detecting apparatus Active CN104020181B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410224033.4A CN104020181B (en) 2014-05-26 2014-05-26 Method integrated circuit material detected with nondestructive detecting apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410224033.4A CN104020181B (en) 2014-05-26 2014-05-26 Method integrated circuit material detected with nondestructive detecting apparatus

Publications (2)

Publication Number Publication Date
CN104020181A true CN104020181A (en) 2014-09-03
CN104020181B CN104020181B (en) 2016-06-29

Family

ID=51437056

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410224033.4A Active CN104020181B (en) 2014-05-26 2014-05-26 Method integrated circuit material detected with nondestructive detecting apparatus

Country Status (1)

Country Link
CN (1) CN104020181B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105842262A (en) * 2016-04-13 2016-08-10 云南电网有限责任公司电力科学研究院 Quality inspection method of high-voltage current-limiting fuse based on X-ray
CN108564572A (en) * 2018-04-02 2018-09-21 滁州晨润工贸有限公司 A kind of product quality control method based on characteristics of image
CN111389747A (en) * 2020-04-26 2020-07-10 成都复星联创科技有限公司 Bamboo strip intelligent sorting system and method based on visual identification

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4653855A (en) * 1984-10-09 1987-03-31 Quantum Diagnostics Ltd. Apparatus and process for object analysis by perturbation of interference fringes
JPH0851131A (en) * 1994-08-08 1996-02-20 Kobe Steel Ltd X-ray inspecting method
JP2004093279A (en) * 2002-08-30 2004-03-25 Hitachi Kenki Fine Tech Co Ltd X-ray inspection apparatus and x-ray inspection program
CN1715894A (en) * 2004-06-30 2006-01-04 欧姆龙株式会社 Inspection method and system for and method of producing component mounting substrate
CN101706458A (en) * 2009-11-30 2010-05-12 中北大学 Automatic detection system and detection method of high resolution printed circuit board
CN102150036A (en) * 2008-09-10 2011-08-10 欧姆龙株式会社 X-ray inspection device and method for X-ray inspection
CN102265143A (en) * 2008-12-22 2011-11-30 欧姆龙株式会社 X-ray inspection method and x-ray inspection apparatus

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4653855A (en) * 1984-10-09 1987-03-31 Quantum Diagnostics Ltd. Apparatus and process for object analysis by perturbation of interference fringes
JPH0851131A (en) * 1994-08-08 1996-02-20 Kobe Steel Ltd X-ray inspecting method
JP2004093279A (en) * 2002-08-30 2004-03-25 Hitachi Kenki Fine Tech Co Ltd X-ray inspection apparatus and x-ray inspection program
CN1715894A (en) * 2004-06-30 2006-01-04 欧姆龙株式会社 Inspection method and system for and method of producing component mounting substrate
CN102150036A (en) * 2008-09-10 2011-08-10 欧姆龙株式会社 X-ray inspection device and method for X-ray inspection
CN102265143A (en) * 2008-12-22 2011-11-30 欧姆龙株式会社 X-ray inspection method and x-ray inspection apparatus
CN101706458A (en) * 2009-11-30 2010-05-12 中北大学 Automatic detection system and detection method of high resolution printed circuit board

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
李峰峰: "电子元器件的外观检测系统的研究与开发", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105842262A (en) * 2016-04-13 2016-08-10 云南电网有限责任公司电力科学研究院 Quality inspection method of high-voltage current-limiting fuse based on X-ray
CN105842262B (en) * 2016-04-13 2018-10-12 云南电网有限责任公司电力科学研究院 A kind of high-voltage and current-limitation fuse quality determining method based on X-ray
CN108564572A (en) * 2018-04-02 2018-09-21 滁州晨润工贸有限公司 A kind of product quality control method based on characteristics of image
CN111389747A (en) * 2020-04-26 2020-07-10 成都复星联创科技有限公司 Bamboo strip intelligent sorting system and method based on visual identification

Also Published As

Publication number Publication date
CN104020181B (en) 2016-06-29

Similar Documents

Publication Publication Date Title
KR102102874B1 (en) Automatic optical inspection system and operating method thereof
EP3234861B1 (en) Screening of electronic components for detection of counterfeit articles using automated inspection system
CN106670127A (en) Full-automatic visual inspection system for screen defects
US9316687B2 (en) Method and apparatus for testing light-emitting device
KR100882252B1 (en) Semiconductor substrate defects detection device and method of detection of defects
US10215695B1 (en) Inspection system and method for detecting defects at a materials interface
JP2014038045A (en) Inspection device, illumination, inspection method, program and substrate producing method
CN101327483B (en) Device and method for classifying chips
CN104020181A (en) Method for detecting integrated circuit material by nondestructive detecting apparatus
CN107672325A (en) It is automatically positioned the Laser Jet method of product
KR100629921B1 (en) Test handler and method for test of semiconductor package
KR20120089109A (en) Apparatus and Method for Detecting Coordinate of Unit
TWI536011B (en) Electronic parts inspection device
CN103606529A (en) Method and device for improving defect classification accuracy
JP2017138193A (en) Article inspection device
KR20210031811A (en) Die bonding device and method of manufacturing semiconductor device
JP2018054464A (en) Electronic component conveyance device and electronic component inspection device
CN204102921U (en) LED chip-detecting apparatus
CN207224865U (en) Vision positioning Laser Jet mark detection device
KR20160031812A (en) Apparatus for Investigating LED Package with X-ray
CN105486698A (en) AOI automatic detection device and detection method
KR102572968B1 (en) Wafer inspection method using the deep learning and wafer inspection device for implementing this
JP5205224B2 (en) Component mounting state inspection device
CN112974289B (en) Sorting method, sorting apparatus, computer-readable storage medium, and electronic device
KR200368650Y1 (en) Test handler for test of semiconductor package

Legal Events

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