CN101751571A - Practical binary document image tilt angle detection method - Google Patents
Practical binary document image tilt angle detection method Download PDFInfo
- Publication number
- CN101751571A CN101751571A CN200910255753A CN200910255753A CN101751571A CN 101751571 A CN101751571 A CN 101751571A CN 200910255753 A CN200910255753 A CN 200910255753A CN 200910255753 A CN200910255753 A CN 200910255753A CN 101751571 A CN101751571 A CN 101751571A
- Authority
- CN
- China
- Prior art keywords
- pixel
- image
- value
- mark
- carried out
- 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
Links
Images
Landscapes
- Image Analysis (AREA)
- Character Input (AREA)
Abstract
The invention relates to a practical binary document image tilt angle detection method, which belongs to the technical field of image preprocessing. The detection method comprises: extending foreground pixels by traversing the foreground pixels in images to strengthen the characteristics of text lines, and then carrying out filtering smoothing; extracting edge characteristics by thinning algorithm, and then marking the pixels after thinning; counting up the number of pixel points, and filtering and binarizing the edge characteristics of the number of successive pixel points less than the threshold value; and finally, carrying out the Hough transformation for the images after processing, and detecting the tilt angle of the binary document images. The experiment proves that the invention has higher practical value.
Description
Technical field
The present invention relates to a kind of binary document image tilt angle detection method of practicality, belong to image preconditioning technique field.
Background technology
In the acquisition process of file and picture, inclination to a certain degree more or less can appear in scanning image, and the inclination of this image causes difficulty not only can for the cutting of next step character, also influences final character recognition precision.Therefore, image inclination detects with correction and just seems very important, becomes the pretreated important step of image.
For the file and picture inclination angle detection, though real straight line is seldom arranged in the file and picture, in image, very strong directivity is arranged between literal line, therefore can be by detecting the trend of image Chinese one's own profession, thus obtain the angle of inclination.At present, the main method of document image tilt angle detection has: Hough converter technique, least square method, Mathematical Morphology Method etc.All there is a common problem in these methods, promptly are difficult to take into account simultaneously the degree of accuracy and the speed of detection method.Wherein, the Hough conversion is the most frequently used tilt angle detection method, and its advantage is to the insensitive for noise in the image.Its shortcoming is that calculated amount is big, and bigger sensing range and higher detection precision will increase the consumption of time and internal memory sharp, and this is very disadvantageous beyond doubt for the practicability that the batch document image tilt angle detects.
Summary of the invention
At the deficiencies in the prior art, the invention provides a kind of binary document image tilt angle detection method of practicality.
A kind of binary document image tilt angle detection method of practicality, detection method is as follows:
When 1) in file and picture, running into foreground pixel, foreground pixel is expert at and up and down each one-row pixels be the center with the foreground pixel, L pixel of both sides expansion to the left and right, it is 1 that pixel value is set, the L value is 7-10;
2) image after the expansion is carried out length filtering and remove projection isolated in the image, obtain effective text linear feature;
3) employing is carried out smothing filtering based on the method for template to image, at first according to the feature of the image after the expansion, chooses the 3X3 template, then the module that meets template in the image is filled, is removed or connects, and makes the image that obtains at last more level and smooth;
4) refinement: image is traveled through, and when to run into pixel value be 1 pixel, the next line pixel value that the value of this pixel is set to this pixel deducted lastrow pixel value poor of this pixel; Being on duty is at-1 o'clock, and the value of this pixel is set to 0;
5) mark: by row traversal pixel, run into pixel value and be 1 situation, then this pixel is carried out mark with mark, begin by the next column this pixel of line scanning 8-neighborhood from this pixel then, if some pixels are not 0 in the 8-neighborhood next column, then this pixel is carried out mark with mark, and, continue scanning according to the variation of row number, pixel up to 8-neighborhood next column all is 0, add 1 with mark this moment, and return by row and continue scanning, finishes up to the traversal entire image;
6) continuous pixel count of statistics and filtration, binaryzation: comprise following steps: (1) sets up array according to the mark value, the pixel count that statistics links to each other, and deposit in the corresponding array element; (2) traversing graph picture when to run into value be not 0 situation, if its corresponding array element then is changed to 0 with this pixel less than threshold value, otherwise is changed to 1; Threshold value is 2L+1 or 4L+2;
7) Hough conversion: the gained image is carried out the Hough conversion, return the angle of inclination, finish whole angle of inclination testing process.
In actual applications, text is usually based on horizontally-arranged, the angle of image inclination generally less (in tens degree), at this situation, and the problem that is difficult to take into account simultaneously degree of accuracy and speed in the Hough conversion, the present invention is by the foreground pixel in the traversing graph picture, and it is expanded, strengthen the line of text feature, it is level and smooth to carry out filtering then, and passes through thinning algorithm, extract edge feature, and then by the pixel after the refinement is carried out mark, and statistical pixel counts, and the edge feature that contiguous pixels is counted less than threshold value filters and binaryzation then.At last the image of handling gained is carried out the angle of inclination of Hough change detection binary document image.Experimental results show that the present invention has very high practical value.
Description of drawings
Fig. 1 is a testing process synoptic diagram of the present invention.
Fig. 2 is an expanded text row schematic flow sheet.
Fig. 3 is a length filtering schematic flow sheet.
Fig. 4 is the template and the schematic flow sheet of smothing filtering.
Fig. 5 is the refinement schematic flow sheet.
Fig. 6 is the mark schematic flow sheet.
Fig. 7 is statistics and filtering process synoptic diagram.
Embodiment
Embodiment:
A kind of binary document image tilt angle detection method of practicality, process flow diagram as shown in Figure 1, detection method is as follows:
When 1) in file and picture, running into foreground pixel, foreground pixel is expert at and up and down each one-row pixels be the center with the foreground pixel, L pixel of both sides expansion to the left and right, it is 1 that pixel value is set, the L value is 7-10;
2) image after the expansion is carried out length filtering and remove projection isolated in the image, obtain effective text linear feature;
3) employing is carried out smothing filtering based on the method for template to image, at first according to the feature of the image after the expansion, chooses the 3X3 template, then the module that meets template in the image is filled, is removed or connects, and makes the image that obtains at last more level and smooth;
4) refinement: image is traveled through, and when to run into pixel value be 1 pixel, the next line pixel value that the value of this pixel is set to this pixel deducted lastrow pixel value poor of this pixel; Being on duty is at-1 o'clock, and the value of this pixel is set to 0;
5) mark: by row traversal pixel, run into pixel value and be 1 situation, then this pixel is carried out mark with mark, begin by the next column this pixel of line scanning 8-neighborhood from this pixel then, if some pixels are not 0 in the 8-neighborhood next column, then this pixel is carried out mark with mark, and, continue scanning according to the variation of row number, pixel up to 8-neighborhood next column all is 0, add 1 with mark this moment, and return by row and continue scanning, finishes up to the traversal entire image;
6) continuous pixel count of statistics and filtration, binaryzation: comprise following steps: (1) sets up array according to the mark value, the pixel count that statistics links to each other, and deposit in the corresponding array element; (2) traversing graph picture when to run into value be not 0 situation, if its corresponding array element then is changed to 0 with this pixel less than threshold value, otherwise is changed to 1; Threshold value is 2L+1 or 4L+2;
7) Hough conversion: the gained image is carried out the Hough conversion, return the angle of inclination, finish whole angle of inclination testing process.
Claims (1)
1. the binary document image tilt angle detection method of a practicality is characterized in that, detection method is as follows:
When 1) in file and picture, running into foreground pixel, foreground pixel is expert at and up and down each one-row pixels be the center with the foreground pixel, L pixel of both sides expansion to the left and right, it is 1 that pixel value is set, the L value is 7-10;
2) image after the expansion is carried out length filtering and remove projection isolated in the image, obtain effective text linear feature;
3) employing is carried out smothing filtering based on the method for template to image, at first according to the feature of the image after the expansion, chooses the 3X3 template, then the module that meets template in the image is filled, is removed or connects, and makes the image that obtains at last more level and smooth;
4) refinement: image is traveled through, and when to run into pixel value be 1 pixel, the next line pixel value that the value of this pixel is set to this pixel deducted lastrow pixel value poor of this pixel; Being on duty is at-1 o'clock, and the value of this pixel is set to 0;
5) mark: by row traversal pixel, run into pixel value and be 1 situation, then this pixel is carried out mark with mark, begin by the next column this pixel of line scanning 8-neighborhood from this pixel then, if some pixels are not 0 in the 8-neighborhood next column, then this pixel is carried out mark with mark, and, continue scanning according to the variation of row number, pixel up to 8-neighborhood next column all is 0, add 1 with mark this moment, and return by row and continue scanning, finishes up to the traversal entire image;
6) continuous pixel count of statistics and filtration, binaryzation: comprise following steps: (1) sets up array according to the mark value, the pixel count that statistics links to each other, and deposit in the corresponding array element; (2) traversing graph picture when to run into value be not 0 situation, if its corresponding array element then is changed to 0 with this pixel less than threshold value, otherwise is changed to 1; Threshold value is 2L+1 or 4L+2;
7) Hough conversion: the gained image is carried out the Hough conversion, return the angle of inclination, finish whole angle of inclination testing process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102557536A CN101751571B (en) | 2009-12-28 | 2009-12-28 | Practical binary document image tilt angle detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102557536A CN101751571B (en) | 2009-12-28 | 2009-12-28 | Practical binary document image tilt angle detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101751571A true CN101751571A (en) | 2010-06-23 |
| CN101751571B CN101751571B (en) | 2012-02-29 |
Family
ID=42478532
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009102557536A Expired - Fee Related CN101751571B (en) | 2009-12-28 | 2009-12-28 | Practical binary document image tilt angle detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101751571B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102496018A (en) * | 2011-12-08 | 2012-06-13 | 方正国际软件有限公司 | Document skew detection method and system |
| CN102646194A (en) * | 2012-02-22 | 2012-08-22 | 大连理工大学 | Method for performing printer type evidence obtainment by utilizing character edge features |
| CN102938062A (en) * | 2012-10-16 | 2013-02-20 | 山东山大鸥玛软件有限公司 | Document image slant angle estimation method based on content |
| CN108121983A (en) * | 2016-11-29 | 2018-06-05 | 蓝盾信息安全技术有限公司 | A kind of text image method for correcting error based on Fourier transformation |
| CN108932516A (en) * | 2018-07-11 | 2018-12-04 | 凌云光技术集团有限责任公司 | It is a kind of rotate text image bearing calibration and device |
| CN109166114A (en) * | 2018-08-24 | 2019-01-08 | 武汉轻工大学 | The recognition methods of backbone intervertenral space, equipment, storage medium and device |
| CN111062874A (en) * | 2019-12-12 | 2020-04-24 | 腾讯科技（深圳）有限公司 | Text image display method, device, equipment and storage medium |
| CN113673522A (en) * | 2021-10-21 | 2021-11-19 | 北京世纪好未来教育科技有限公司 | Method, device and equipment for detecting inclination angle of text image and storage medium |
-
2009
- 2009-12-28 CN CN2009102557536A patent/CN101751571B/en not_active Expired - Fee Related
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102496018A (en) * | 2011-12-08 | 2012-06-13 | 方正国际软件有限公司 | Document skew detection method and system |
| CN102496018B (en) * | 2011-12-08 | 2013-07-24 | 方正国际软件有限公司 | Document skew detection method and system |
| CN102646194A (en) * | 2012-02-22 | 2012-08-22 | 大连理工大学 | Method for performing printer type evidence obtainment by utilizing character edge features |
| CN102938062A (en) * | 2012-10-16 | 2013-02-20 | 山东山大鸥玛软件有限公司 | Document image slant angle estimation method based on content |
| CN102938062B (en) * | 2012-10-16 | 2015-08-19 | 山东山大鸥玛软件有限公司 | A kind of content-based file image inclination angular estimation method |
| CN108121983A (en) * | 2016-11-29 | 2018-06-05 | 蓝盾信息安全技术有限公司 | A kind of text image method for correcting error based on Fourier transformation |
| CN108932516A (en) * | 2018-07-11 | 2018-12-04 | 凌云光技术集团有限责任公司 | It is a kind of rotate text image bearing calibration and device |
| CN109166114A (en) * | 2018-08-24 | 2019-01-08 | 武汉轻工大学 | The recognition methods of backbone intervertenral space, equipment, storage medium and device |
| CN111062874A (en) * | 2019-12-12 | 2020-04-24 | 腾讯科技（深圳）有限公司 | Text image display method, device, equipment and storage medium |
| CN111062874B (en) * | 2019-12-12 | 2023-03-31 | 腾讯科技（深圳）有限公司 | Text image display method, device, equipment and storage medium |
| CN113673522A (en) * | 2021-10-21 | 2021-11-19 | 北京世纪好未来教育科技有限公司 | Method, device and equipment for detecting inclination angle of text image and storage medium |
| CN113673522B (en) * | 2021-10-21 | 2022-04-19 | 北京世纪好未来教育科技有限公司 | Method, device and equipment for detecting inclination angle of text image and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101751571B (en) | 2012-02-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101751571B (en) | Practical binary document image tilt angle detection method | |
| US8548246B2 (en) | Method and system for preprocessing an image for optical character recognition | |
| CN101923741B (en) | Paper currency number identification method based on currency detector | |
| US8194983B2 (en) | Method and system for preprocessing an image for optical character recognition | |
| CN103258198B (en) | Character extracting method in a kind of form document image | |
| CN101515325B (en) | Character extracting method in digital video based on character segmentation and color cluster | |
| CN104732183A (en) | One-dimensional barcode identification method based on image sampling line grey scale information analysis | |
| CN101458770B (en) | Character recognition method and system | |
| Jusoh et al. | Application of freeman chain codes: An alternative recognition technique for Malaysian car plates | |
| CN103136528B (en) | A kind of licence plate recognition method based on dual edge detection | |
| WO2017016448A1 (en) | Qr code feature detection method and system | |
| CN104182750A (en) | Extremum connected domain based Chinese character detection method in natural scene image | |
| CN103258201A (en) | Form line extraction method integrating global information and local information | |
| CN110516673B (en) | Yi-nationality ancient book character detection method based on connected component and regression type character segmentation | |
| CN106504225A (en) | A kind of recognition methodss of regular polygon and device | |
| Yadav et al. | Text extraction in document images: highlight on using corner points | |
| CN102054275B (en) | Real-time detection method for sea sky line of gray level image | |
| CN109447117A (en) | The double-deck licence plate recognition method, device, computer equipment and storage medium | |
| Vithlani et al. | Structural and statistical feature extraction methods for character and digit recognition | |
| CN115438682B (en) | Method and device for determining decoding direction and decoding equipment | |
| CN106127118A (en) | A kind of English word recognition methods and device | |
| CN102542273A (en) | Detection method and system for complex formula areas in document image | |
| CN115564673A (en) | Method and system for extracting three-dimensional point cloud underground garage column and automatically generating vector | |
| CN102332088B (en) | Vote symbolic machine visual identification method based on run length feature | |
| CN112418123B (en) | Hough transformation-based engineering drawing line and line type identification method |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120229 Termination date: 20121228 |