CN101655981A - Method for detecting and adjusting inversion of certificate image - Google Patents

Abstract

The invention discloses a method for detecting and adjusting the inversion of a certificate image, comprising the following steps: (1) acquiring a color certificate image by an image acquisition tool;(2) shrinking the certificate image according to a certain scale factor; (3) converting the RGB color space of the certificate image into HIS color space; (4) converting the HIS image into an binaryimage by utilizing the constraint condition of a human face skin color; (5) carrying out median filtering of the binary image to obtain a binary image only with a human face part; (6) cutting a humanface part image according to the human face part binary image; (7) obtaining a human face edge binary image by an RGB edge extraction method; and (8) utilizing horizontal projection statistics to detect the inversion of a certificate. The method for detecting and adjusting the inversion of the certificate image shrinks the image under the condition of keeping the human face outline information ofthe image constant so as to greatly reduce the pixel participating human face skin color positioning and outline extraction operation, thereby not only enhancing the arithmetic efficiency but also ensuring the high accuracy, robustness and reliability of arithmetic detection.

Description

The method of detecting and adjusting inversion of certificate image

Technical field

The present invention relates to a kind of method that is applied to detecting and adjusting inversion of certificate image, specifically, relate to a kind of face complexion location of utilizing, and add up to determine whether inverted method of image in conjunction with projection.

Background technology

From certificate, obtain relevant information, more and more appear in the various fields of the social people's livelihood, maturation along with computer technology, Flame Image Process and optical character recognizer, utilize computing machine to carry out certificate information and discern typing automatically and become possibility, this appearance of obtaining information approach will improve professional both sides' work efficiency greatly.Certificate information collection system based on computing machine OCR can be widely used in the section industries such as public security, bank, telecommunications, postal service, railway, civil aviaton, security, civil administration, entry and exit, army, Internet bar and hotel, can save great amount of manpower and material resources, have quite wide application prospect.

Certificate information collection system can be divided into image acquisition, image pre-service, certificate location, image segmentation, five key steps of character recognition, detecting and adjusting inversion of certificate image is an important process process between certificate location, image segmentation, certificate image can phenomena of inversion occur owing to the randomness that certificate is placed, this brings difficulty not only can for next step image segmentation, even causes certificate information correctly not discern.

The detecting and adjusting inversion of certificate image is important problem in the certificate information collection system, in order to avoid the inverted problem of certificate imaging as far as possible, can be to collecting device and acquisition step, using method etc. are carried out some corresponding restrictions, but consider the robustness and extensibility (as being adapted to different big or small certificates) of system, also to give the user simultaneously and use degree of freedom to the full extent, so stop relatively difficulty of certificate image imaging phenomena of inversion, but if certificate is not carried out corresponding detecting and adjusting inversion, then can carry out information subgraph piece (as head portrait to certificate image, name, cutting apart message block such as passport NO.) brings very big difficulty, prominent question be exactly can be because information is damaged, factors such as character distortion and cause character recognition and information extraction to make mistakes.

Generally, the image acquisition of certificate is to finish in a dark all around specific environment, it is more far better than the imaging circumstances of common Vehicle License Plate Recognition System, guaranteed that the interference that imaging process is subjected to is had predictability preferably, usually the certificate image that photographs is comparatively clear, effect is relatively good, and object plane also is parallel with lens plane, perspective imaging phenomenon (Perspective Imaging) promptly can not occur.Simultaneously, because the head portrait of certificate such as China second-generation identity card etc. is colored, so for people's face Face Detection, the RGB profile sketches the contours that good basis is provided.

Summary of the invention

At above deficiency, the invention provides a kind of can whether the inversion by fast automatic judgement certificate image, if be inverted, and the method for the detecting and adjusting inversion of certificate image of in time correcting.

The method of detecting and adjusting inversion of certificate image of the present invention comprises:

1) image acquisition: utilize the image acquisition instrument to obtain the coloured image of certificate;

2) conversioning colour space: the rgb color space of the certificate image that above-mentioned steps is obtained converts the HIS color space to;

3) people's face detects: according to face complexion constraint condition the image transitions that above-mentioned steps obtains is become bianry image;

4) people's face cutting: the bianry image according to people face part in the bianry image cuts out the people face part image that step 1) obtains;

5) extract facial contour: utilize RGB edge extracting method, obtain the edge binary images of people's face;

6) the horizontal projection statistics detects the certificate inversion.

Described step 1) also comprises step 11): the certificate coloured image that step 1) is obtained carries out convergent-divergent.

Described step 3) also comprises 31): the bianry image of people's face that step 3) is obtained carries out the medium filtering operation, obtains the only bianry image of remaining people face part of a width of cloth.

What the bianry image of people's face was carried out that the medium filtering operation adopts is 9 * 9 template.

The process of utilizing RGB edge extracting method to extract facial contour in the described step 5) comprises:

51) (i, red, green, blue component j) is r to definition protoplast face image pixel f ₁, g ₁, b ₁, (i+1, red, green, blue component j) is r to the f that goes together mutually ₂, g ₂, b ₂, (i, red, green, blue component j+1) is r to same column f ₃, g ₃, b ₃, above-mentioned component increases 1.5 times earlier certainly, image g after obtaining handling according to following formula then (i, red, green, blue component j) is r, g, b:

$r=2\×\sqrt{{(r_1-r_2)}^2+{(r_1-r_3)}^2}$

$g=2\×\sqrt{{(g_1-g_2)}^2+{(g_1-g_3)}^2};$

$b=2\×\sqrt{{(b_1-b_2)}^2+{(b_1-b_3)}^2}$

52) with step 51) r, g, these three components of b of obtaining are converted to gray-scale value with the gray scale formula:

gray＝0.11×b+0.5×g+0.39×r ；

53), be 0 or 255 with this grayvalue transition according to constraint condition:

$\mathrm{binaryvalue}=\left\{\begin{array}{cc}255,& \mathrm{gray}\≤200,\\ 0& \mathrm{gray}>200\end{array}\right.;$

54) traversal view picture facial image repeats step 51)～52) operation, finish until traversal, can obtain the bianry image of a width of cloth facial contour.

Described step 6) horizontal projection statistics detects certificate inversion process and comprises:

61) bianry image with facial contour is divided into upper and lower two parts N ₁And N ₂

61) if N ₁-N ₂＜15, then can conclude image inversion, should be with image Rotate 180 °; Otherwise, judge that then image is normal.

Beneficial effect of the present invention: detecting and adjusting inversion of certificate image method of the present invention under the constant situation of image facial contour information is dwindled image keeping, thereby the profile that participates in face complexion location and image extracts the pixel of operation to significantly reduce, this has not only improved efficiency of algorithm, also guaranteed the accuracy height that algorithm detects, strong robustness, reliability height.

Description of drawings

Fig. 1 is a detecting and adjusting inversion of certificate image method flow diagram of the present invention;

Fig. 2 is certificate original image of the present invention (because certificate relates to personal information, so the information of having carried out is erased in the certificate image subregion, down together);

Fig. 3 detects the bianry image that obtains for the present invention according to face complexion;

Fig. 4 is for utilizing the bianry image after 9 * 9 templates are carried out medium filtering;

The people face part image of Fig. 5 for cutting out according to detected people's face position;

Fig. 6 is the facial contour figure that utilizes RGB profile extraction method to extract;

Fig. 7 is the bianry image that utilizes constraint condition that facial contour figure is converted to;

Fig. 8 rotates normal picture after utilizing the horizontal projection statistic law to detect image inversion.

Embodiment

Below in conjunction with accompanying drawing the present invention is further set forth, but not as a limitation of the invention, dreamboat output of the present invention is that a width of cloth does not have inverted certificate image.

As shown in Figure 1, be inversion of certificate image automatic detecting and correcting method overview flow chart of the present invention, it comprises:

1, image acquisition

At first be to carry out the certificate image collection, can utilize common IP Camera, digital camera, scanner or buy some certificate recognition instrument pickup images.In the present invention, because the imaging background environment is fine, the certificate image that photographs is more clear, is fit to follow-up work of treatment.

In view of certificate all has a width of cloth clour mixing picture, so the present invention adopts based on colour of skin location people's face, come detected image whether to be inverted according to face characteristic then, export normal picture at last.

2, downscaled images

Owing to only need utilize people's face to come detected image whether to be inverted, image is narrowed down to the relevant information that certain degree can't change people's face in the image, then can run counter to desire but image is too small.So with the suitable factor of dwindling it is dwindled behind the certificate color reproduction image that needs above-mentioned steps is obtained, to reduce the operand of algorithm.

3, conversioning colour space

The rgb color space of this image is converted to the HIS color space, and conversion formula is:

$S=1-\frac{3}{(R+G+B)}\left[\min (R,G,B)\right]$

$I=\frac{1}{3}(R+G+B)$

Wherein, R, G, B represent to normalize to redness, green, the blue three-channel color component of the rgb color space of [0,1] scope; H, I, S represent colourity, brightness, the saturation degree component of HIS color space; $\mathrm{\θ}=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{1/2}}\right\},$ Angle between the red axle of represent pixel point and HIS color space.

4, people's face detects

Is bianry image according to following face complexion constraint condition with image transitions:

0.003＜H＜0.174

0.040＜S＜0.352

0.352＜I＜1

Also there are some interference noises in image after the conversion, must these interference noises of filtering.

5, medium filtering

Utilize 9 * 9 template that bianry image is carried out the medium filtering operation, can be with the interference noise filtering, thus obtain the only bianry image of remaining people face part of a width of cloth.

6, people's face cutting

According to people's face binary map the people face part image that the front dwindles figure is cut out, this moment picture size will littler, the participation computing pixel quantity also still less, thereby subsequent operation speed can be greatly enhanced.

7, extract facial contour

Utilize RGB edge extracting method, obtain the edge binary images of people's face, method is as follows:

Step1: (i, red, green, blue component j) is r to establish protoplast's face image pixel f ₁, g ₁, b ₁, (i+1, red, green, blue component j) is r to the f that goes together mutually ₂, g ₂, b ₂, (i, red, green, blue component j+1) is r to same column f ₃, g ₃, b ₃, above-mentioned component increases 1.5 times earlier certainly, image g after obtaining handling according to following formula then (i, red, green, blue component j) is r, g, b, these three components can calculate by following formula:

$r=2\×\sqrt{{(r_1-r_2)}^2+{(r_1-r_3)}^2}$

$g=2\×\sqrt{{(g_1-g_2)}^2+{(g_1-g_3)}^2}$

$b=2\×\sqrt{{(b_1-b_2)}^2+{(b_1-b_3)}^2}$

Step2: r, g, these three components of b that Step1 is obtained are converted to gray-scale value with the gray scale formula:

gray＝0.11×b+0.5×g+0.39×r

Step3:, be 0 or 255 with this grayvalue transition according to constraint condition:

$\mathrm{binaryvalue}=\left\{\begin{array}{cc}255,& \mathrm{gray}\≤200,\\ 0& \mathrm{gray}>200\end{array}\right.$

Step4: the traversal entire image, repeat Step1～Step3 operation, finish until traversal, can obtain the bianry image of the width of cloth facial contour this moment.

8, the horizontal projection statistics detects the certificate inversion

People's face is divided into upper and lower two parts N ₁, N ₂According to face characteristic as can be known, the lines pixel sum of band eyes, eyebrow part is always than only the lines pixel sum with nose, lip part is many, therefore can adopt the pixel sum of horizontal projection statistical method statistics top and the bottom image, determine according to both sizes whether image is inverted then: if N ₁-N ₂＜15, promptly the difference of top and lower part profile lines pixel is less than 15, then can conclude image inversion (because the people on the face half part comprise eyes, eyebrow, the nose that comprises than people's face the latter half, the contour pixel of face is much more.), should be with image Rotate 180 °; Otherwise, judge that then image is normal.

So far, tilt and the end of inversion detection algorithm, export the normal certificate image of a width of cloth.

Specific embodiment:

Use a USB IP Camera, resolution is 640 * 480, the second generation residential identity certificate image that the width of cloth level of taking is put.

As seen from Figure 2, there is phenomena of inversion in certificate image, can locate the people face part by face complexion, then according to the facial contour feature, utilizes the horizontal projection statistic law to come detected image whether to have phenomena of inversion.

Fig. 2 is reduced into 1/4 of original size, to reduce operand, then according to face complexion constraint condition, with image binaryzation, as shown in Figure 3.

Exist a large amount of interference noise in the binary map this moment, and it is carried out 9 * 9 medium filtering, can be with noise filtering, as shown in Figure 4.

As shown in Figure 4, the image cut of people face part is come out in only remaining people face part in this moment binary map, as shown in Figure 5, further reduces operand.

Can obtain the profile diagram 6 of people face part according to formula RGB edge extracting method, by the constraint condition of gray-scale value, profile diagram is converted to bianry image again, as shown in Figure 7.

Fig. 7 is divided into upper and lower two halves N ₁, N ₂, the summation of utilizing the horizontal projection statistic law that these two parts are added up profile lines pixel respectively respectively obtains N ₁=52, N ₂=85, by N ₁-N ₂＜15 as can be known source images be inverted, so ° can be normal with image restoring with Fig. 2 Rotate 180, as shown in Figure 8.

Whether can correctly detect the certificate image by whole algorithm and the visible the present invention of result thereof fell Put, thereby with it reduction, give the use free degree of user Geng Gao. Because this algorithm takes full advantage of Dwindle the color of rear image and people's face colour of skin, the constant characteristics of people's face profile feature, greatly reduced The operand of algorithm has improved the speed of service of program, thereby reaches efficient, reliable, robustness By force, the designing requirement of zmodem.

Claims (6)

1, a kind of method of detecting and adjusting inversion of certificate image is characterized in that, it comprises:

1) image acquisition: utilize the image acquisition instrument to obtain the coloured image of certificate;

2) conversioning colour space: the rgb color space of the certificate image that above-mentioned steps is obtained converts the HIS color space to;

3) people's face detects: according to face complexion constraint condition the image transitions that above-mentioned steps obtains is become bianry image;

4) people's face cutting: the bianry image according to people face part in the bianry image cuts out the people face part image that step 1) obtains;

5) extract facial contour: utilize RGB edge extracting method, obtain the edge binary images of people's face;

6) the horizontal projection statistics detects the certificate inversion.

2, the method for detecting and adjusting inversion of certificate image according to claim 1 is characterized in that, described step 1) also comprises step 11): the certificate coloured image that step 1) is obtained carries out convergent-divergent.

3, the method for detecting and adjusting inversion of certificate image according to claim 1 and 2 is characterized in that, described step 3) also comprises 31): step 3) is obtained.

4, the method for detecting and adjusting inversion of certificate image according to claim 3 is characterized in that, described bianry image to people's face carries out that medium filtering operation adopts is 9 * 9 template.

According to the method for claim 1 or 4 described detecting and adjusting inversion of certificate image, it is characterized in that 5, the process of utilizing RGB edge extracting method to extract facial contour in the described step 5) comprises:

51) (i, red, green, blue component j) is r to definition protoplast face image pixel f ₁, g ₁, b ₁, (i+1, red, green, blue component j) is r to the f that goes together mutually ₂, g ₂, b ₂, (i, red, green, blue component j+1) is r to same column f ₃, g ₃, b ₃, above-mentioned component increases 1.5 times earlier certainly, image g after obtaining handling according to following formula then (i, red, green, blue component j) is r, g, b:

$r=2\×\sqrt{{(r_1-r_2)}^2+{(r_1-r_3)}^2}$

$g=2\×\sqrt{{(g_1-g_2)}^2+{(g_1-g_3)}^2};$

$b=2\×\sqrt{{(b_1-b_2)}^2+{(b_1-b_3)}^2}$

52) with step 51) r, g, these three components of b of obtaining are converted to gray-scale value with the gray scale formula:

gray＝0.11×b+0.5×g+0.39×r；

53), be 0 or 255 with this grayvalue transition according to constraint condition:

$\mathrm{binaryvalue}=\left\{\begin{array}{cc}255,& \mathrm{gray}\≤200,\\ 0& \mathrm{gray}>200\end{array}\right.;$

54) traversal view picture facial image repeats step 51)～52) operation, finish until traversal, can obtain the bianry image of a width of cloth facial contour.

According to the method for claim 1 or 4 described detecting and adjusting inversion of certificate image, it is characterized in that 6, described step 6) horizontal projection statistics detects certificate inversion process and comprises:

61) bianry image with facial contour is divided into upper and lower two parts N ₁And N ₂

61) if N ₁-N ₂＜15, then can conclude image inversion, should be with image Rotate 180 °; Otherwise, judge that then image is normal.

Images