CN204576547U - A kind of chip character check device - Google Patents
A kind of chip character check device Download PDFInfo
- Publication number
- CN204576547U CN204576547U CN201520278488.4U CN201520278488U CN204576547U CN 204576547 U CN204576547 U CN 204576547U CN 201520278488 U CN201520278488 U CN 201520278488U CN 204576547 U CN204576547 U CN 204576547U
- Authority
- CN
- China
- Prior art keywords
- chip
- character
- bearing platform
- processing apparatus
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The utility model discloses a kind of chip character check device.Comprise lighting device, chip bearing platform, image collecting device, image processing apparatus and warning device, described chip bearing platform, for batch carries chips, and move relative to image collecting device according to chip placement order; Described lighting device, is arranged between chip bearing platform and image collecting device; Described image collecting device, arranges and directly over chip bearing platform chip to be imaged, for treating imager chip imaging, and view data is passed to image processing apparatus; Described image processing apparatus, for storage standards character string, receiving chip view data, and verify by chip character check method, error result is exported to warning device; Described warning device, with image processing apparatus data cube computation, for pointing out the chip verifying and make mistakes.Described apparatus structure is simple, and recognition efficiency is high.
Description
Technical field
The utility model belongs to integrated circuit field of quality control, more specifically, relates to a kind of chip character check device.
Background technology
The vital effect of integrated circuit (IC) chip performer in electron trade, the extensive manufacture of IC chip reduces the price of electronic product.In order to ensure the quality of each IC chip, needing in IC chip production process to detect chip package quality, mainly comprising and checking size and typographic(al) mark.The object that IC chip character detects in real time is according to known reference number (given reference number), whether the numerical character detecting laser printing on chip is correct, because the general volume of IC chip is little, line speed is fast, be difficult to extract character outline and character feature clearly, want the character on real-time extraction IC chip and and standard character compare, appropriate character extraction will be carried out according to different character features, recognition strategy.
Existing chip character check apparatus structure is complicated, and recognition efficiency is low.
Utility model content
For above defect or the Improvement requirement of prior art, the utility model provides a kind of chip character check device, comprises lighting device, chip bearing platform, image collecting device, image processing apparatus and warning device;
Described chip bearing platform, for batch carries chips, and moves relative to image collecting device according to chip placement order;
Described lighting device, is arranged between chip bearing platform and image collecting device;
Described image collecting device, arranges and directly over chip bearing platform chip to be imaged, for treating imager chip imaging, and view data is passed to image processing apparatus;
Described image processing apparatus, for storage standards character string, receiving chip view data, and verify by chip character check method, error result is exported to warning device;
Described warning device, with image processing apparatus data cube computation, for pointing out the chip verifying and make mistakes.
Preferably, described chip character check device, its lighting device is ring light, arrange with chip bearing platform chip to be imaged directly over.
In general, the above technical scheme conceived by the utility model compared with prior art, can obtain following beneficial effect:
The device that the utility model provides, can obtain good IC chip character picture, and batch check and correction IC chip mark problem, substantially increases production efficiency.
Accompanying drawing explanation
Fig. 1 is embodiment 1 chips character recognition and calibration equipment image;
Fig. 2 is the image in embodiment 1 after binaryzation;
Fig. 3 is row projective distribution figure in embodiment 1;
Fig. 4 is the image in embodiment 1 after space segmentation;
Fig. 5 is a line image in embodiment 1;
Fig. 6 is the row projected image of this row in embodiment 1;
Fig. 7 is the pre-character picture of this row in embodiment 1;
Fig. 8 is image after the fractionation of adhesion character in embodiment 1;
Fig. 9 is image after the merging of embodiment 1 Fracture character.
In all of the figs, identical Reference numeral is used for representing identical element or structure, and wherein: 1 is lighting device, 2 is chip bearing platform, and 3 is image collecting device, and 4 is image processing apparatus, and 5 is warning device.
Embodiment
In order to make the purpose of this utility model, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the utility model is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the utility model, and be not used in restriction the utility model.In addition, if below in described each embodiment of the utility model involved technical characteristic do not form conflict each other and just can mutually combine.
The chip character check device that the utility model provides, as shown in Figure 9, comprises lighting device, chip bearing platform, image collecting device, image processing apparatus and warning device;
Described chip bearing platform, for batch carries chips, and moves relative to image collecting device according to chip placement order;
Described lighting device, is arranged between chip bearing platform and image collecting device; Preferred employing ring light, arrange with chip bearing platform chip to be imaged directly over.
The annular light source adopted can make 360 ° of uniform irradiations on the ic chip, and do not have diplopia to produce, the image definition photographed is high.
Described image collecting device, arranges and directly over chip bearing platform chip to be imaged, for treating imager chip imaging, and view data is passed to image processing apparatus;
Described image processing apparatus, for storage standards character string, receiving chip view data, and verify according to chip character check method, error result is exported to warning device;
Preferably, concrete chip character check step is as follows:
(1) chip character picture obtains and pre-service: obtain chip character picture, and be processed into binary image;
(1-1) chip image obtains: the chip image that shooting obtains is carried out filtering process and obtains filtered image;
Filtering method can adopt gaussian filtering, medium filtering, bilateral filtering etc.
(1-2) image binaryzation process: adopt the Da-Jin algorithm based on piecemeal, carries out binaryzation to filtered image in step (1-1), obtains binary image.
Little at chip motion, uneven illumination and chip character volume, the fuzzy incompleteness of image chips of collection can be caused, the problems such as picture gray-scale value is uneven can be there is simultaneously, adopt single Da-Jin algorithm to use the threshold value of the overall situation to carry out binary conversion treatment to chip like this and chip Character segmentation can be caused imperfect.Da-Jin algorithm based on piecemeal adopts local threshold, picture is divided into several pieces, uses the threshold value of every block, carry out binary conversion treatment to entire image.
(2) chip Character segmentation:
(2-1) row segmentation: the binary image obtained in step (1) is lined by line scan, the number of the white pixel point in the every a line of statistical picture, exceedes the number of wherein white pixel point the continuous print presetting threshold value in the ranks capable of a character row;
(2-2) column split: each character row of (2-1) middle acquisition is in steps scanned by column, the number of the white pixel point in each row of statistics character row, exceedes the continuous print of threshold value between predetermined word and arranges as a pre-character picture using the number of white pixel point;
(2-3) Character segmentation: for each the pre-character picture obtained in step (2-2), add up its columns, using the pre-character picture of columns within standard character columns scope as single character picture; By the pre-character picture of columns lower than standard character columns range lower limit, merge successively according to order from left to right, until its columns merging rear character picture is within standard character columns scope, then the character picture after merging is as single character picture; By the pre-character picture of columns higher than standard character columns range limit, split according to the intercharacter threshold value preset, until the columns of each character picture after each fractionation is within standard character columns scope, using each character picture after splitting as single character picture;
(3) chip character recognition verification: the single character picture obtained in step (2) is identified as single character respectively, character group is synthesized character string.
(3-1) normalized: by the single character picture obtained in step (2), adopts contiguous differential technique, keeps character length breadth ratio constant, is scaled the normalization character picture of unified size;
The nearest adjoint point gray-scale value of 4 adjoint point middle distances of interpolation point is as the gray-scale value of this point.If interpolation point (i, j) is dk (k=1,2 to the distance of periphery 4 adjoint point fk (i, j) (k=1,2,3,4), 3,4), then: g (i, j)=fk (i, j), dl=min{d1, d2, d3, d4}, l=1,2,3,4.The object of this step is by reducing picture, reduces the computing time of character picture feature, meets the requirement of real-time of IC chip identification and verification.
(3-2) machine learning identification: the method adopting pattern-recognition, is characterized as the aspect of model with the grid search-engine of character and geometric moment, carries out pattern-recognition, is identified as character by the normalization character picture obtained in step (3-1);
(3-3) combine: by the character in step (3-2) according to it after the initial position of negative image is combined into character string, with preset character string compare, obtain check results.
Described warning device, with image processing apparatus data cube computation, for pointing out the chip verifying and make mistakes.The type of alarm of warning device mainly contains blinking red lamp, warning horn sounds.
Be below embodiment:
Embodiment 1
A kind of chip character check device, as shown in Figure 1, comprises lighting device 1, chip bearing platform 2, image collecting device 3, image processing apparatus 4 and warning device 5;
Described chip bearing platform 2, for batch carries chips, and moves relative to image collecting device 3 according to chip placement order;
Described lighting device 1, is arranged between chip bearing platform 2 and image collecting device 3; Adopt two strip light, arrange with chip bearing platform 2 chip to be imaged directly over.
The strip source adopted can strengthen the contrast of IC chip identification, makes the image definition that photographs high.
Described image collecting device 3, arranges and directly over chip bearing platform 2 chip to be imaged, for treating imager chip imaging, and view data is passed to image processing apparatus 4;
Described image processing apparatus 4, for storage standards character string, receiving chip view data, and following chip character check method verifies, and error result is exported to warning device 5;
Chip character check method concrete steps are as follows:
(1) chip character picture obtains and pre-service: obtain chip character picture, and be processed into binary image;
(1-1) chip image obtains: gather IC chip picture to be identified, carry out gaussian filtering to it.
(1-2) image binaryzation process: use Da-Jin algorithm to carry out binary conversion treatment to filtered image, obtain bianry image, as shown in Figure 2.In image, prospect (character area) and background (white space) are distinguished by different colours, and in this image, white pixel represents character area, and black picture element represents white space.
(2) chip Character segmentation:
(2-1) row segmentation: the binary image obtained in step (1) is lined by line scan, the number of the white pixel point in the every a line of statistical picture, exceedes the number of wherein white pixel point the continuous print presetting threshold value in the ranks capable of a character row; Concrete operations are: carry out number of white pixels statistics to each row of image, obtain white pixel horizontal direction projection histogram, the projection value of every a line is the white pixel point number in this figure in corresponding row, select suitable threshold value, such as 0, white pixel point is from left to right scanned, if n-th line projection value is 0, (n+1)th row projection value >0, then the (n+1)th row is as the starting point of a line; If n-th line projection value >0, the (n+1)th row projection value=0, then n-th line is as the terminal of a line.According to this kind of method, image is from left to right analyzed, the perspective view obtained as shown in Figure 3, row segmentation after figure as shown in Figure 4, through space segmentation after have three single line of text.
(2-2) column split: each character row of (2-1) middle acquisition is in steps scanned by column, the number of the white pixel point in each row of statistics character row, exceedes the continuous print of threshold value between predetermined word and arranges as a pre-character picture using the number of white pixel point;
Concrete operations are: carry out number of white pixels statistics to each row of image, obtain white pixel horizontal direction projection histogram, the projection value of each row is the white pixel point number in this figure in corresponding row, select suitable threshold value, such as 0, row are entered to white pixel point and from left to right scans, if the n-th row projection value is 0, (n+1)th row projection value >0, then the (n+1)th row are as the starting point of a pre-character picture; If the n-th row projection value >0, the (n+1)th row projection value=0, then the n-th row are as the terminal of a pre-character picture.According to this kind of method, Fig. 5 is from left to right analyzed.Its vertical projection figure as shown in Figure 6, pre-character picture, as shown in Figure 7.
(2-3) Character segmentation: for each the pre-character picture obtained in step (2-2), add up its columns, using the pre-character picture of columns within standard character columns scope as single character picture; By the pre-character picture of columns lower than standard character columns range lower limit, merge successively according to order from left to right, until its columns merging rear character picture is within standard character columns scope, then the character picture after merging is as single character picture; By the pre-character picture of columns higher than standard character columns range limit, split according to the intercharacter threshold value preset, until the columns of each character picture after each fractionation is within standard character columns scope, using each character picture after splitting as single character picture;
Concrete operations are as follows:
A1: as Fig. 7 obtains pre-character block, and be labeled as numbering 1-9, each character block width is CSW, according to standard character width C W, pre-character picture is divided into 3 classes, if 0.8CW<CSW<1.2CW, then this pre-character block is standard character block; If CSW>1.6CW, then this pre-character block is adhesion character block; If CSW<0.6CW, then this pre-character block is fracture character block;
A2: analyze from left to right, for adhesion character block, takes fractionation mode, and as being numbered the character block of 1 in Fig. 6, its width 1.6CW<CSW<2.4CW, be then divided into two character blocks, as shown in Figure 8;
A3: for fracture character block, adopt merging mode, as being numbered the character block of 5,6 in Fig. 6, its width is all CSW<0.6CW, be then merged into a character block, as shown in Figure 9
A4: complete in full line image, the fractionation of adhesion character block and the merging of fracture character block, result is as shown in Figure 9;
(3) chip character recognition verification: the single character picture obtained in step (2) is identified as single character respectively, character group is synthesized character string.
(3-1) normalized: by the single character picture obtained in step (2), adopts contiguous differential technique, keeps character length breadth ratio constant, is scaled the normalization character picture of unified size;
The nearest adjoint point gray-scale value of 4 adjoint point middle distances of interpolation point is as the gray-scale value of this point.If interpolation point (i, j) is dk (k=1,2 to the distance of periphery 4 adjoint point fk (i, j) (k=1,2,3,4), 3,4), then: g (i, j)=fk (i, j), dl=min{d1, d2, d3, d4}, l=1,2,3,4.The object of this step is by reducing picture, reduces the computing time of character picture feature, meets the requirement of real-time of IC chip identification and verification.
(3-2) machine learning identification: the method adopting pattern-recognition, the aspect of model is characterized as with the grid search-engine of character and geometric moment, use SVM (support vector machine) to carry out identification and carry out pattern-recognition, the normalization character picture obtained in step (3-1) is identified as character;
(3-3) combine: by the character in step (3-2) according to it after the initial position of negative image is combined into character string, compare with the character string preset, obtain check results: if recognition result and standard results inconsistent, then export misprint mark, if recognition result is consistent with standard results, carry out the inspection of next chip.
Described warning device 5, with image processing apparatus 4 data cube computation, for pointing out the chip verifying and make mistakes.The employing warning horn of warning device 5.
Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present utility model; not in order to limit the utility model; all do within spirit of the present utility model and principle any amendment, equivalent to replace and improvement etc., all should be included within protection domain of the present utility model.
Claims (2)
1. a chip character check device, it is characterized in that, comprise lighting device (1), chip bearing platform (2), image collecting device (3), image processing apparatus (4) and warning device (5);
Described chip bearing platform (2), for batch carries chips, and moves relative to image collecting device according to chip placement order;
Described lighting device (1), is arranged between chip bearing platform (2) and image collecting device (3);
Described image collecting device (3), arranges and directly over chip bearing platform (2) chip to be imaged, for treating imager chip imaging, and view data is passed to image processing apparatus (4);
Described image processing apparatus (4), for storage standards character string, receiving chip view data, and verify according to chip character check method, error result is exported to warning device (5);
Described warning device (5), with image processing apparatus (4) data cube computation, for pointing out the chip verifying and make mistakes.
2. device as claimed in claim 1, it is characterized in that, described lighting device (1) is ring light, arrange with chip bearing platform (2) chip to be imaged directly over.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201520278488.4U CN204576547U (en) | 2015-04-29 | 2015-04-29 | A kind of chip character check device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201520278488.4U CN204576547U (en) | 2015-04-29 | 2015-04-29 | A kind of chip character check device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN204576547U true CN204576547U (en) | 2015-08-19 |
Family
ID=53869153
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201520278488.4U Active CN204576547U (en) | 2015-04-29 | 2015-04-29 | A kind of chip character check device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN204576547U (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110907796A (en) * | 2018-09-14 | 2020-03-24 | 长鑫存储技术有限公司 | Integrated circuit measurement result imaging analysis method and system |
| CN112435222A (en) * | 2020-11-11 | 2021-03-02 | 深圳技术大学 | Circuit board detection method and device and computer readable storage medium |
| CN113433068A (en) * | 2021-07-02 | 2021-09-24 | 合肥图迅电子科技有限公司 | Detection device for detecting chip characters covered by film |
-
2015
- 2015-04-29 CN CN201520278488.4U patent/CN204576547U/en active Active
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110907796A (en) * | 2018-09-14 | 2020-03-24 | 长鑫存储技术有限公司 | Integrated circuit measurement result imaging analysis method and system |
| CN112435222A (en) * | 2020-11-11 | 2021-03-02 | 深圳技术大学 | Circuit board detection method and device and computer readable storage medium |
| CN113433068A (en) * | 2021-07-02 | 2021-09-24 | 合肥图迅电子科技有限公司 | Detection device for detecting chip characters covered by film |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104881665A (en) | Chip character identification and verification method and apparatus | |
| CN105046196B (en) | Front truck information of vehicles structuring output method based on concatenated convolutional neutral net | |
| CN109101924B (en) | Machine learning-based road traffic sign identification method | |
| CN109726717B (en) | Vehicle comprehensive information detection system | |
| CN101872416B (en) | Vehicle license plate recognition method and system of road image | |
| Zheng et al. | HLU 2-Net: a residual U-structure embedded U-Net with hybrid loss for tire defect inspection | |
| CN110008909B (en) | Real-name system business real-time auditing system based on AI | |
| WO2022121039A1 (en) | Bankcard tilt correction-based detection method and apparatus, readable storage medium, and terminal | |
| CN106610969A (en) | Multimodal information-based video content auditing system and method | |
| CN104809481A (en) | Natural scene text detection method based on adaptive color clustering | |
| Wang et al. | License plate segmentation and recognition of Chinese vehicle based on BPNN | |
| CN106446894A (en) | Method for recognizing position of spherical object based on contour | |
| CN109961399B (en) | Optimal suture line searching method based on image distance transformation | |
| CN105354573B (en) | A kind of container licence plate recognition method and system | |
| CN108334881A (en) | A kind of licence plate recognition method based on deep learning | |
| CN103310211A (en) | Filling mark recognition method based on image processing | |
| CN204576547U (en) | A kind of chip character check device | |
| CN102147867A (en) | Method for identifying traditional Chinese painting images and calligraphy images based on subject | |
| CN106295528A (en) | A kind of vehicle checking method based on multi-part spatial relation GMM modeling | |
| CN104112138A (en) | Object color classification method and device | |
| CN110852315A (en) | Method and device for quickly identifying annular characters | |
| CN112740223A (en) | Tire sidewall imaging method | |
| CN115294483A (en) | Small target identification method and system for complex scene of power transmission line | |
| CN111695373B (en) | Zebra stripes positioning method, system, medium and equipment | |
| CN115082776A (en) | Electric energy meter automatic detection system and method based on image recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |