CN107909592A - A kind of line drawing drawing generating method for mural painting image - Google Patents
A kind of line drawing drawing generating method for mural painting image Download PDFInfo
- Publication number
- CN107909592A CN107909592A CN201710917202.6A CN201710917202A CN107909592A CN 107909592 A CN107909592 A CN 107909592A CN 201710917202 A CN201710917202 A CN 201710917202A CN 107909592 A CN107909592 A CN 107909592A
- Authority
- CN
- China
- Prior art keywords
- image
- result
- edge
- mural painting
- fdog
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000010422 painting Methods 0.000 title claims abstract description 33
- 238000001914 filtration Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000005516 engineering process Methods 0.000 claims abstract description 9
- 238000009499 grossing Methods 0.000 claims abstract description 8
- 230000002708 enhancing effect Effects 0.000 claims abstract description 4
- 238000003708 edge detection Methods 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims 1
- 238000000605 extraction Methods 0.000 abstract description 8
- 238000005728 strengthening Methods 0.000 abstract 1
- 238000012360 testing method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000003647 oxidation Effects 0.000 description 2
- 238000007254 oxidation reaction Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 208000037656 Respiratory Sounds Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000007591 painting process Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 238000004260 weight control Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a kind of line drawing drawing generating method for mural painting image, comprise the following steps:The artwork of input is smoothly pre-processed using global variation smoothing filter, obtains smoothed image I;Smoothed image I is converted into gray level image I0;Based on Gaussian Blur to I0Carry out High frequency filter lifting, handling result I2Represent;The tangential flow field in edge is constructed, recycles improved FDoG filtered methods to extract the lines figure with enhancing effect, output result is represented with H':The characteristics of on the basis of the tangential flow field in edge is constructed, improving FDoG methods, strengthening mural painting using the lines of high boost extraction, the gray level image I of original input0It is substituted for it and High frequency filter lifting result I2The result I obtained after multiplicationm, define new FDoG wave filters;Processing is filtered to image H ' using median filtering technology.
Description
Technical field
It is related to the line drawing figure generation technique for mural painting image, in particular for the smoothing denoising and wheel of colored mural painting image
Wide extractive technique.
Background technology
Mural painting is one of cultural heritage of China's preciousness, and the line drawing for playing bottom in Dunhuang frescoes is drawn, and is both that Chinese tradition is painted
A kind of distinctive painting mode is drawn, and is a kind of special mural painting salvo, there is important researching value.Traditional line drawing
Drafting relies primarily on manual operations.Due to the particularity of mural painting, painter can only rely on local light to be drawn;With multimedia
The popularization of technology, is changed to that grottoes mural painting is taken pictures or recorded, and painter against digital mural painting again than carrying out line drawing drawing.
It is sufficiently complex by the process for carrying out line drawing drafting by hand and very time-consuming, the speed phase of the threat sprawling faced with mural painting
Than achieving the purpose that to protect the very small of mural painting to utilize computer and image processing techniques with this, line drawing generation being carried out to mural painting
Process, will not both cause damage historical relic, also have the characteristics that efficiently, the conventional method such as reusable do not have it is therefore desirable to pin
Line drawing figure Generation Technology is carried out to digitlization mural painting.
Line drawing figure generation key technology is related to edge extracting.Typical arithmetic operators have Canny [1] etc..But by this
When a little operators directly apply to the extraction of line drawing picture, response can be produced to a road both sides, so that same road occur corresponds to 2 sides
The situation of edge.Global marginal probability detective operators (Glabalized Probability of a Boundary, gPb) [2] tool
There is higher accuracy rate, and one edge is only generated compared with Xi Bi roads for some, but its width is single pixel, continuity
If also poor is handled without skeletal extraction, although the result of gPb operators can embody the width in a road, can not carry
Take the road than comparatively dense.Gaussian difference (FDoG) algorithm [3] based on stream is replaced using the anisotropic filtering device based on stream
Traditional edge detection operator, is filtered vector field with Non-linear Kernel to build edge tangent stream so that prominent edge is protected
Its original direction is stayed, the direction at weak edge is consistent with the direction of prominent edge in its neighborhood, is a kind of effective lines expression side
Method.But phenomena such as influence of nature or human factor, mural painting image generally faces the oxidation of defect crackle, therefore exist in image a large amount of
Noise, or many unnecessary details, therefore must take into consideration and mural painting image is located in advance using smooth or noise-removed technology
Reason operation.
Bibliography:
[1] a kind of effective text image binarization method microcomputer informations of Zhuan Jun, Li Bicheng, Chen Gang, 2005
(8):56-58.
[2] Chen Dan, a bee, a kind of improved text image Binarization methods computer engineering of He Guiming, 2003 (13):
85-86.
[3]J.Bernsen.Dynamic Thresholding of Gray level.Internal Conference
on Pattern Recognition,1986:1251-1255.
The content of the invention
The present invention proposes a kind of line drawing drawing generating method for mural painting image, can be automatically converted to colored mural painting image
Corresponding line drawing figure, has the characteristics that strong antijamming capability, applied widely.Technical solution is as follows:
A kind of line drawing drawing generating method for mural painting image, comprises the following steps:
(1) artwork of input is smoothly pre-processed using global variation smoothing filter, obtains smoothed image I;
(2) smoothed image I is converted into gray level image I0, its red, green, blue triple channel image uses I respectivelyR、IGAnd IBTable
Show;
(3) based on Gaussian Blur to I0Carry out High frequency filter lifting, handling result I2Represent;
(4) the tangential flow field in edge is constructed, recycles improved FDoG filtered methods to extract with enhancing effect
Lines figure, output result are represented with H', and contours extract is carried out using following process:
1st step:Edge tangent flow field is built from smoothed image I, t (x) represents local edge direction, in its vertical direction
There are maximum-contrast, i.e. gradient direction;
2nd step:On the basis of the tangential flow field in edge is constructed, FDoG methods are improved, the line extracted using high boost
Bar strengthens the characteristics of mural painting, the gray level image I of original input0It is substituted for it and High frequency filter lifting result I2Obtained after multiplication
As a result Im, define new FDoG wave filter H (x);
(5) processing is filtered to image H ' using median filtering technology.
Beneficial effects of the present invention have:1) there is stronger anti-noise jamming ability;2) the line drawing figure of generation can retain wall
The main structure and artistic style of picture.
Brief description of the drawings
Fig. 1 institutes extracting method block diagram
Fig. 2 institutes extracting method processing procedure exemplary plot
(a) input picture (b) sharpening result (c) gray-scale map
(d) extraction of gray scale increment figure (e) high boost result (f) line drawing and denoising result
Three width mural painting line drawing of mural painting image (b) the generation result of three width of Fig. 3 part of test results (a) input
Embodiment
The extraction of line drawing figure and generation method proposed by the present invention towards mural painting image, including smoothing processing, High frequency filter,
Edge tangent stream structure, contours extract and medium filtering and etc..Global variation is carried out to artwork mural painting image first smoothly to locate
Reason, rejects unnecessary detailed information, then carries out gray proces, then carries out the High frequency filter based on Gaussian Blur and is lifted come simple
Change the distribution of background pixel gray value, next build edge tangent stream (ETF), and obtain using improved FDoG extraction profiles
Abstract lines figure is finally handled result using medium filtering means.
Mainly include:Smooth pretreatment, gray processing, High frequency filter lifting, stream difference of Gaussian contours extract and medium filtering
And etc..Fig. 1 gives the block diagram of institute's extracting method.
1st, smooth pretreatment
The problems such as mural painting damaged, oxidation are caused due to nature or human factor, thus we first have to mural painting image into
Row smoothing processing, excludes noise jamming as far as possible.Institute's extracting method uses global variation smoothing filter.Detailed process is as follows:
Algorithm 1:Global variation is smooth
1st step:Input signal g, exports signal f, and object function is defined as formula (1):
λ is the proportion of a weight control between the two, and actually it is a smoothing parameter, when its value is more big more flat
Non-zero gradient number and λ are in monotonic increase relation in sliding images
2nd step:Calculate the L0 norm c (f) of image gradient, i.e. constraints such as formula (2).
C (f)=# p | | fp-fp-1|≠0} (2)
Wherein p and p-1 is adjacent element in image, | fp-fp-1| it is exactly image gradient, # { } represents to count, therefore the table
The non-zero number number that gradient is not zero in other words for representing output up to formula is equal to k.Processing result image is represented with I.
2nd, gray processing
The smoothed image obtained in previous step is converted into gray level image.Smoothed image is represented with I, its red, green, blue threeway
Road image uses I respectivelyR、IGAnd IBRepresent.Gray level image is obtained using formula (3), and uses I0Represent, that is, have:
3rd, High frequency filter is lifted
Usual mural painting image Zhong Bi road pixels and background pixel grey value difference in subrange are larger, but complete
Intensity value ranges have coincidence to be easier to be extracted for Shi Bi roads in the range of office, while keep the contrast of a road and background, herein
Based on Gaussian Blur to I0Carry out High frequency filter lifting, handling result I2Represent.Detailed process is as follows:
Algorithm 2:High boost
1st step:To gray level image I0Carry out Gaussian Blur, i.e. Gassian low-pass filter.The definition of wave filter such as formula (4) institute
Show, handling result I1Represent.
Wherein, σ is the standard deviation of Gaussian Profile, can represent the degrees of expansion of distribution curve.
2nd step:The increment of gray-scale map is calculated using formula (5), as a result with Δ I0Represent, that is, have
3rd step:Carry out high boost.As a result I is used2Represent, that is, have
I2(x, y)=min (255, I0(x,y)+ΔI0(x,y)) (6)
4th, contours extract
The tangential flow field in edge is constructed, recycles improved FDoG filtered methods to extract the lines with enhancing effect
Figure.Output result is represented with H'.The present invention carries out contours extract using procedure below:
Algorithm 3:Flow difference of Gaussian filtering
1st step:Edge tangent stream is built from smoothed image I using classical ETF (Edge Tangent Flow) method
t(x).T (x) represents local edge direction, it is meant that has maximum-contrast, i.e. gradient direction in its vertical direction.
2nd step:On the basis of the tangential flow field in edge is constructed, the FDoG (Flow-Based of classics are improved
Difference Of Gaussian) method, using high boost extraction lines strengthen mural painting the characteristics of, original input ash
Spend figure image I0It is substituted for it and high boost result I2The result I obtained after multiplicationm.Represent that defining new FDoG filters with H (x)
Ripple device, it is defined as follows:
In formula, GσFor a unilateral amount, variance σ2Gaussian function, see formula (4).σmDetermine the length of stream core,
To σmAs soon as assigning a definite value, the size of gradient direction S is determined.F (s) is defined as follows:
Im(x, y)=I0(x,y)·Im(x,y)/255(9)
ls(t) represent in straight line lsPoint at upper t, t are arc length parameters, its value is between [- T, T].Im(ls(t)) represent
Input picture ImIn ls(t) value.The noise rank that ρ controls detect, generally in [0.97,1] regional change.
3rd step:Go out the lines image of black and white using threshold extraction, as a result with H'(x) represent.
Wherein, tanh (H (x)) is hyperbolic tangent function, and τ is the threshold value of [0,1] scope, controls the sensitive journey of edge detection
Degree.
In above-mentioned algorithm, ρ=0.998, τ=0.5, σ are takens=1.6 σc, σcValue can be given by user and is adjusted.
5th, medium filtering
In view of the above method generation result in can still include partial noise, using median filtering technology to image H ' into
Row filtering process, reduces influence of noise.The image after bilateral filtering is handled is represented with G.
Using the Visual C++2010 under Windows7 systems as experiment simulation platform.The scanning for selecting oneself to gather
File and picture amounts to 120 width images as test set.Horizontal/vertical resolution is 300dpi, and pixel number is 800 × 680.Adopt
Test image is handled with institute's extracting method of the present invention, has obtained good treatment effect, average treatment speed is 10s, place
Reason speed disclosure satisfy that requirement.
Fig. 2 show institute's extracting method processing procedure example.Fig. 3 show more handling results, wherein (a) is input
Mural painting image, (b) for institute's extracting method of the present invention handling result.
The present invention may be summarized to be following method and step;
Step 1:Using algorithm 1, convolution (1) and formula (2), input color mural painting image is smoothly pre-processed, place
Reason result is represented with I.
Step 2:Using formula (3), gray proces, handling result I are carried out to I0Represent.
Step 3:Use algorithm 2, convolution (4), formula (5) and formula (6), to I0Carry out High frequency filter lifting, handling result
Use I2Represent.
Step 4:Using algorithm 3, convolution (7) to formula (10), builds edge tangent stream, to ImCarry out the Gauss based on stream
Difference lines extract, and handling result H ' is represented.
Step 5:Medium filtering is carried out to H ', removes noise, handling result is represented with G.
Claims (2)
1. a kind of line drawing drawing generating method for mural painting image, comprises the following steps:
(1) artwork of input is smoothly pre-processed using global variation smoothing filter.Obtain smoothed image I;
(2) smoothed image I is converted into gray level image I0, its red, green, blue triple channel image uses I respectivelyR、IGAnd IBRepresent;
(3) based on Gaussian Blur to I0Carry out High frequency filter lifting, handling result I2Represent;
(4) the tangential flow field in edge is constructed, recycles improved FDoG filtered methods to extract the lines with enhancing effect
Figure, output result are represented with H', and contours extract is carried out using following process:
1st step:Edge tangent flow field is built from smoothed image I, t (x) represents local edge direction, has most in its vertical direction
Big contrast, i.e. gradient direction;
2nd step:On the basis of the tangential flow field in edge is constructed, FDoG methods are improved, the lines extracted using high boost are strong
The characteristics of changing mural painting, the gray level image I of original input0It is substituted for it and High frequency filter lifting result I2The result obtained after multiplication
Im, define new FDoG wave filter H (x);
3rd step:The lines image of black and white is extracted using threshold method, as a result with H'(x) represent;
(5) processing is filtered to image H ' using median filtering technology.
2. according to the method described in claim 1, it is characterized in that, the formula of the lines image of black and white is extracted using threshold method
For:
Wherein, τ is the threshold value of [0,1] scope, controls the sensitivity of edge detection.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710917202.6A CN107909592A (en) | 2017-09-28 | 2017-09-28 | A kind of line drawing drawing generating method for mural painting image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710917202.6A CN107909592A (en) | 2017-09-28 | 2017-09-28 | A kind of line drawing drawing generating method for mural painting image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107909592A true CN107909592A (en) | 2018-04-13 |
Family
ID=61841043
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710917202.6A Pending CN107909592A (en) | 2017-09-28 | 2017-09-28 | A kind of line drawing drawing generating method for mural painting image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107909592A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108921862A (en) * | 2018-05-28 | 2018-11-30 | 天津科技大学 | Mural painting line drawing figure generating algorithm based on convolutional neural networks |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030059113A1 (en) * | 2001-09-27 | 2003-03-27 | Walton William C. | System and method for creating unclassified line drawings from classified NTM images |
CN102930576A (en) * | 2012-10-15 | 2013-02-13 | 中国科学院软件研究所 | Feature flow-based method for generating abstract line drawing |
CN103914863A (en) * | 2014-03-10 | 2014-07-09 | 西藏民族学院 | Method for abstractly drawing color image |
CN106485765A (en) * | 2016-10-13 | 2017-03-08 | 中国科学院半导体研究所 | A kind of method of automatic description face stick figure |
-
2017
- 2017-09-28 CN CN201710917202.6A patent/CN107909592A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030059113A1 (en) * | 2001-09-27 | 2003-03-27 | Walton William C. | System and method for creating unclassified line drawings from classified NTM images |
CN102930576A (en) * | 2012-10-15 | 2013-02-13 | 中国科学院软件研究所 | Feature flow-based method for generating abstract line drawing |
CN103914863A (en) * | 2014-03-10 | 2014-07-09 | 西藏民族学院 | Method for abstractly drawing color image |
CN106485765A (en) * | 2016-10-13 | 2017-03-08 | 中国科学院半导体研究所 | A kind of method of automatic description face stick figure |
Non-Patent Citations (2)
Title |
---|
吴宗胜: "建筑物图像的风格化增强技术研究", 《中国优秀硕士学位论文全文数据库.信息科技辑》 * |
许迪: "面向中国古代壁画的线描画生成方法", 《计算机工程》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108921862A (en) * | 2018-05-28 | 2018-11-30 | 天津科技大学 | Mural painting line drawing figure generating algorithm based on convolutional neural networks |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Savant | A review on edge detection techniques for image segmentation | |
CN105654436A (en) | Backlight image enhancement and denoising method based on foreground-background separation | |
US9042649B2 (en) | Color document image segmentation and binarization using automatic inpainting | |
CN108510491A (en) | Blur the filter method of skeleton critical point detection result under background | |
CN103226829A (en) | Image edge detection method based on edge enhancement operator | |
CN104637067A (en) | Method for detecting defect of textured surface | |
CN105894474A (en) | Non-linear image enhancement method, and edge detection method using the same | |
CN103942756B (en) | A kind of method of depth map post processing and filtering | |
CN106327464A (en) | Edge detection method | |
CN106846271A (en) | A kind of method of reticulate pattern in removal identity card picture | |
JPH05216989A (en) | Method and device for extracting area of color picture | |
CN103455985A (en) | Road crack enhancement method based on Hessian structural analysis | |
Zhao et al. | Analysis of image edge checking algorithms for the estimation of pear size | |
CN114332081A (en) | Textile surface abnormity determination method based on image processing | |
CN107909592A (en) | A kind of line drawing drawing generating method for mural painting image | |
Asadov et al. | Using hough transformation for detection of linear geological structures in satellite images | |
CN109448010B (en) | Automatic four-side continuous pattern generation method based on content features | |
CN115131336B (en) | Dark line defect detection method under four-color background picture of display screen | |
CN105354823A (en) | Tree-ring image edge extraction and segmentation system | |
Kaur et al. | Contour detection of gradient images using morphological operator and transform domain filtering | |
CN106600662A (en) | Drawing method and drawing device for main lines in image | |
Zhang et al. | Moving shadow removal algorithm based on HSV color space | |
Wang et al. | Research on thangka image edge detection based on differential operator | |
Dari | Identifikasi Deteksi Tepi Pada Pola Wajah Menerapkan Metode Sobel, Roberts dan Prewitt | |
CN104463898A (en) | Image identification counting system for logistics platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180413 |