CN105139404A - Identification camera capable of detecting photographing quality and photographing quality detecting method - Google Patents

Abstract

The invention discloses an identification camera capable of detecting photographing quality and a photographing quality detecting method. The identification camera comprises an information acquiring module, a quality detecting module, a scoring module, a quality score computing module, and a photography suggesting module. The method comprises steps of: acquiring the file attribute information, the portrait characteristic information, the background information, the brightness information, the chrominance information, and the sharpness information of an identification photo; detecting the quality of scoring dimensions one by one according to the acquired file attribute information, the portrait characteristic information, the background information, the brightness information, the chrominance information, and the sharpness information; scoring the quality detection results of the scoring dimensions in order to obtain the scores of the scoring dimensions; according to the scores of the scoring dimensions and the weights of the scores, computing the total quality score of the identification photo; and according to the total quality score of the identification photo, giving a professional photography suggestion. The identification camera may guarantee that the identification photos passing the detection comply with a standard requirement correlative to the identification photos.

Description

A kind of license camera and shooting quality detection method detecting shooting quality

Technical field

The present invention relates to a kind of the license camera and the shooting quality detection method that detect shooting quality, belong to certificate photograph shooting and processing technology field.

Background technology

Certificate photograph refers to there is the requirement of its standard by the photo that certificate making as legal in I.D., passport, Hong Kong pass, Macao's pass, the Taiwan pass, exit permit, residence permit, social security card etc. uses, as:

No.2 residence card digital photo technical standard: " technical requirement of GA461-2004 ID cards digital photo ";

Motor vehicle driving license digital photo technical standard: " GA482-2012 People's Republic of China (PRC) motor vehicle driving certificate ";

Exit and entry certificates digital photo technical standard: " technical requirement of GA/T1180-2014 exit and entry certificates digital photo ".

The project of these standard-requireds is numerous and in small, broken bits, and people are generally difficult to understand completely these standards, and perform very difficult especially, shooting effect varies with each individual especially, and photographic quality is uneven.But, be all by the personnel of specialty, picture quality is manually checked at present, not yet there is a software automatically can complete quality testing to these photos on the market, therefore, whether the photo that domestic consumer but to have no way of learning captured by it out after intelligent terminal photographs photo meets the operative norm of certificate photograph.

Summary of the invention

The object of the invention is the defect in order to solve above-mentioned prior art, provide a kind of license camera detecting shooting quality, this license camera is by obtaining the information of the rear certificate photograph of shooting, quality testing item by item being carried out to every dimensions, guaranteeing that the certificate photograph by detecting can meet the requirement of certificate photograph relevant criterion.

Another object of the present invention is to provide a kind of portrait pose detection method.

Object of the present invention can reach by taking following technical scheme:

Detect a license camera for shooting quality, described license camera comprises:

Data obtaining module, for the certificate photograph according to shooting imaging, obtains file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information;

Quality detection module, for according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carries out quality testing item by item to every dimensions; Wherein, every dimensions is the examination criteria according to government's public security industry standard formulation;

Grading module, for the default score value according to every dimensions, marks to the quality measurements of every dimensions, obtains the score value of every dimensions;

Quality overall score computing module, for the score value according to every dimensions, and scoring weight, calculate the quality overall score of certificate photograph;

Shooting suggestion module, for the quality overall score according to certificate photograph, provides the shooting suggestion of specialty.

Further, described data obtaining module comprises:

File attribute information acquiring unit, for obtaining the file attribute information of certificate photograph; Wherein, described file attribute information comprises shooting time, colored figure place, file suffixes, file size and image pixel;

Portrait characteristic acquisition unit, for obtaining the portrait characteristic information of certificate photograph; Wherein, described portrait characteristic comprises eyes, nose, ear, face, facial contour, trunk profile, portrait profile, the crown and chin characteristic;

Background information acquiring unit, for when background is not replaced, is converted to Lab color space image by certificate photograph, to Lab color space image, read each pixel of foreground area, set up the mixed Gauss model of prospect colourity; To each pixel of background area, the histogram of background extraction colourity, calculates the overlapping possibility of the histogram of background colourity and the mixed Gauss model of prospect colourity, to the accurate mean square value P of overlapping possibility label taking; Or for when background replaces to normal background, adopt edge detecting technology to detect portrait profile to certificate photograph, obtain the background information beyond certificate photograph portrait profile;

Monochrome information acquiring unit, for obtaining the brightness histogram of certificate photograph, carries out data analysis to brightness histogram, calculates the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA;

Chrominance information acquiring unit, for certificate photograph being converted to Lab color space image, adds up a axle of each pixel at Lab color space image and colourity mean value Da and Db of b axle, according to colourity mean value Da and Db, calculates colour cast COEFFICIENT K ₂;

Sharpness information acquiring unit, for certificate photograph is converted to gray level image, calculates sharpness factor DR.

Further, described quality detection module comprises:

Shooting time detecting unit, for obtaining shooting time from file attribute information, judging whether to take in the recent times preset, if so, then determining whether photochrome, if so, then meet the requirements;

Colored figure place detecting unit, for obtaining colored figure place from file attribute information, judging whether to be not less than 24, if so, then meeting the requirements;

Compressed article quality detection unit, for obtaining file suffixes from file attribute information, judges whether file is compressed file, when file is compressed file, determines whether that JPEG compresses, and if so, judges whether compression quality meets the requirement of corresponding certificate photo;

File size detecting unit, for obtaining file size from file attribute information, judges whether file size is greater than the preset value of corresponding certificate photo, if so, then meets the requirements;

Image pixel detecting unit, for obtaining image pixel from file attribute information, judges whether image pixel meets the requirement of corresponding certificate photo;

Background detection unit, for when background is not replaced, obtain standard mean square value P from background information, whether criterion mean square value P is greater than default mean square value, if not, then meets the requirements; Or for when background replaces to normal background, judge whether the background information obtained meets the requirement of corresponding certificate photo;

Bounding box features detecting unit, for adopting edge detecting technology, detecting the whether well-regulated bounding box features of certificate photograph surrounding, if nothing, then meeting the requirements;

Brightness detection unit, for obtaining the abnormal COEFFICIENT K of brightness from monochrome information ₁with average luminance shift value DA, if the abnormal COEFFICIENT K of brightness ₁be less than 1, then brightness is normal, meets the requirements; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is greater than 0, then brightness is excessively bright; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is less than 0, then too dark brightness;

Colorimetric detection units, for obtaining colour cast COEFFICIENT K from chrominance information ₂, colourity mean value Da and Db, if colour cast COEFFICIENT K ₂be less than 1, then colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₂be more than or equal to 1, when colourity mean value Da is greater than 0, represent partially red, when colourity mean value Da is less than 0, represent partially green, when colourity mean value Db is greater than 0, represent partially yellow, when colourity mean value Db is less than 0, represent partially blue;

Sharpness detecting unit, for obtaining sharpness factor DR from sharpness information, if sharpness factor DR is greater than default sharpness factor value, then it is clear to illustrate, meets the requirements;

Face detecting unit, for obtaining eyes, nose, ear, face, the crown and chin characteristic from portrait characteristic information, by the eyes of acquisition, nose, ear, face, the crown and chin characteristic, compare with the standard feature preset, judge that whether face are complete, if so, then meet the requirements;

Facial contour detecting unit, for obtaining facial contour feature data from portrait characteristic information, detects the diversity factor between facial contour feature data and default facial contour feature, if diversity factor is greater than preset value, then distort, if diversity factor is less than or equal to preset value, then meet the requirements; Wherein, the diversity factor between described facial contour feature data and default facial contour feature is by the ratio calculation of non-overlapped area and overlapping area;

Portrait posture detecting unit, for obtaining eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic from portrait characteristic information, judges whether portrait attitude is rectified, and if so, then meets the requirements;

Head portrait size and location detecting unit, for obtaining eyes, ear, the crown and chin characteristic from portrait characteristic information, relevance calculating is carried out to eyes, ear, the crown and chin characteristic, obtaining the wide pixel of head, head high pixel, eyes distance pixel, eyes position distance up/down Edge Distance pixel and the crown gains fame and fortune apart from certificate photograph coboundary pixel, the parameter of these data and corresponding certificate photo is compared, judges whether to meet the requirements.

Further, described portrait posture detecting unit comprises:

Face state-detection subelement, for according to the eyes obtained, nose, face and facial contour feature data, judges whether face is rectified, and if so, then meets the requirements;

Shoulder state-detection subelement, for according to the trunk contour feature data obtained, judges whether shoulder flushes, and if so, then meets the requirements;

Eyes open width detection subelement, for according to the eye feature data obtained, calculate the width opening part in the middle of eyes, if width is greater than preset value, then meet the requirements;

Face is closed opens detection sub-unit, for according to the face characteristic obtained, adopts edge detecting technology, judges on face whether be a curve in the middle of lower lip, if so, then illustrates that face closes, and meets the requirements, and if not, then illustrates that face opens;

Detection sub-unit without a hat on, for according to the portrait profile characteristic obtained, detects the diversity factor between portrait profile characteristic and default portrait profile feature, if diversity factor is less than preset value, and head color is without obvious segmentation, then illustrates without a hat on, meet the requirements;

Redeye detection subelement, for according to the eye feature data obtained, calculates the colour cast COEFFICIENT K of eyes ₃with colourity mean value Da1, if colour cast COEFFICIENT K ₃be less than 1, then eyes colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₃be more than or equal to 1, and when colourity mean value Da1 is greater than 0, then there is blood-shot eye illness;

The color detection sub-unit of adornment, for the facial contour according to acquisition, eyes and face characteristic, detect color error ratio value, the color error ratio value of eye face position, the color error ratio value at face position of face, if the color error ratio value of one of them is greater than default deviate, then illustrate there is heavy make-up; If the color error ratio value of three is all less than or equal to default deviate, then illustrates without heavy make-up, meet the requirements;

Face detection subelement, for obtaining face mask characteristic, detect the diversity factor between face mask characteristic and default face mask feature, if diversity factor is less than preset value, then hair does not cover face, meets the requirements.

Further, described face state-detection subelement judges whether face is rectified, and specifically comprises:

According to eye feature data, generate two lines, calculate the angle of two lines and certificate photograph horizontal direction, judge whether this angle exceeds preset range, if not, then illustrate and face camera;

According to eyes, nose and face characteristic, obtain the center point coordinate of two lines, obtain the centre coordinate of nose, obtain the centre coordinate of face, fitting a straight line is done to three points and obtains straight line LC, calculate the angle of this straight line LC and certificate photograph vertical direction, judge whether this angle exceeds preset range, if not, then illustrate that head is vertical;

According to facial contour feature data, obtain the center line FLC of human face region vertical direction, calculate the center line FLC of human face region vertical direction and the distance DC of straight line LC, calculate the ratio of this distance DC and human face region width, judge whether this ratio exceeds preset range, if not, then face horizontal center is described;

When facing the vertical and face horizontal center of camera, head, judging that face is proper, meeting the requirements.

Further, described shoulder state-detection subelement judges whether shoulder flushes, and specifically comprises:

To trunk profile coordinate array, fit to a polygon, calculate the coordinate (X1 that polygonal center of gravity obtains center of gravity, Y1), calculate the distance Dweight of X1 and certificate photograph vertical center of gravity line, calculate the ratio of distance Dweight and certificate photograph width W F, judge whether this ratio exceeds preset range;

To trunk profile coordinate array, adopt minimum enclosed rectangle algorithm, obtain a minimum rectangle, calculate the angle of rectangular vertical center of gravity line and certificate photograph vertical center of gravity line, judge whether this angle exceeds preset range;

If the angle of the ratio of distance Dweight and certificate photograph width W F, rectangular vertical center of gravity line and certificate photograph vertical center of gravity line does not all exceed preset range, then shoulder flushes, and meets the requirements.

Another object of the present invention can reach by taking following technical scheme:

A kind of shooting quality detection method, be applied in license camera, described method comprises:

Described license camera, according to the certificate photograph of shooting imaging, obtains file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information;

Described license camera, according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carries out quality testing item by item to every dimensions; Wherein, every dimensions is the examination criteria according to government's public security industry standard formulation;

Described license camera, according to the default score value of every dimensions, is marked to the quality measurements of every dimensions, obtains the score value of every dimensions;

Described license camera is according to the score value of every dimensions, and scoring weight, calculates the quality overall score of certificate photograph;

Described license camera, according to the quality overall score of certificate photograph, provides the shooting suggestion of specialty.

Further, described license camera, according to the certificate photograph of shooting imaging, obtains file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, specifically comprises:

Obtain the file attribute information of certificate photograph; Wherein, described file attribute information comprises shooting time, colored figure place, file suffixes, file size and image pixel;

Obtain the portrait characteristic information of certificate photograph; Wherein, described portrait characteristic comprises eyes, nose, ear, face, facial contour, trunk profile, portrait profile, the crown and chin characteristic;

When background is not replaced, certificate photograph is converted to Lab color space image, to Lab color space image, reads each pixel of foreground area, set up the mixed Gauss model of prospect colourity; To each pixel of background area, the histogram of background extraction colourity, calculates the overlapping possibility of the histogram of background colourity and the mixed Gauss model of prospect colourity, to the accurate mean square value P of overlapping possibility label taking; Or when background replaces to normal background, adopt edge detecting technology to detect portrait profile to certificate photograph, obtain the background information beyond certificate photograph portrait profile;

Obtain the brightness histogram of certificate photograph, data analysis is carried out to brightness histogram, calculate the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA;

Certificate photograph is converted to Lab color space image, adds up a axle of each pixel at Lab color space image and colourity mean value Da and Db of b axle, according to colourity mean value Da and Db, calculate colour cast COEFFICIENT K ₂;

Certificate photograph is converted to gray level image, calculates sharpness factor DR.

Further, described license camera, according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carries out quality testing item by item to every dimensions, specifically comprises:

From file attribute information, obtain shooting time, judge whether to take in the recent times preset, if so, then determine whether photochrome, if so, then meet the requirements;

From file attribute information, obtain colored figure place, judge whether to be not less than 24, if so, then meet the requirements;

From file attribute information, obtain file suffixes, judge whether file is compressed file, when file is compressed file, determines whether that JPEG compresses, if so, judge whether compression quality meets the requirement of corresponding certificate photo;

From file attribute information, obtain file size, judge whether file size is greater than the preset value of corresponding certificate photo, if so, then meets the requirements;

From file attribute information, obtain image pixel, judge whether image pixel meets the requirement of corresponding certificate photo;

When background is not replaced, obtain standard mean square value P from background information, whether criterion mean square value P is greater than default mean square value, if not, then meets the requirements; Or for when background replaces to normal background, judge whether the background information obtained meets the requirement of corresponding certificate photo;

Adopt edge detecting technology, detect the whether well-regulated bounding box features of certificate photograph surrounding, if nothing, then meet the requirements;

The abnormal COEFFICIENT K of brightness is obtained from monochrome information ₁with average luminance shift value DA, if the abnormal COEFFICIENT K of brightness ₁be less than 1, then brightness is normal, meets the requirements; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is greater than 0, then brightness is excessively bright; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is less than 0, then too dark brightness;

Colour cast COEFFICIENT K is obtained from chrominance information ₂, colourity mean value Da and Db, if colour cast COEFFICIENT K ₂be less than 1, then colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₂be more than or equal to 1, when colourity mean value Da is greater than 0, represent partially red, when colourity mean value Da is less than 0, represent partially green, when colourity mean value Db is greater than 0, represent partially yellow, when colourity mean value Db is less than 0, represent partially blue;

From sharpness information, obtain sharpness factor DR, if sharpness factor DR is greater than default sharpness factor value, then it is clear to illustrate, meets the requirements;

Eyes, nose, ear, face, the crown and chin characteristic is obtained from portrait characteristic information, by the eyes of acquisition, nose, ear, face, the crown and chin characteristic, compare with the standard feature preset, judge that whether face are complete, if so, then meet the requirements;

From portrait characteristic information, obtain facial contour feature data, detect the diversity factor between facial contour feature data and default facial contour feature, if diversity factor is greater than preset value, then distort, if diversity factor is less than or equal to preset value, then meet the requirements; Wherein, the diversity factor between described facial contour feature data and default facial contour feature is by the ratio calculation of non-overlapped area and overlapping area;

From portrait characteristic information, obtain eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic, judge whether portrait attitude is rectified, and if so, then meets the requirements;

Eyes, ear, the crown and chin characteristic is obtained from portrait characteristic information, relevance calculating is carried out to eyes, ear, the crown and chin characteristic, obtaining the wide pixel of head, head high pixel, eyes distance pixel, eyes position distance up/down Edge Distance pixel and the crown gains fame and fortune apart from certificate photograph coboundary pixel, the parameter of these data and corresponding certificate photo is compared, judges whether to meet the requirements.

Further, describedly judge whether portrait attitude is rectified, and specifically comprises:

According to the eyes obtained, nose, face and facial contour feature data, judge whether face is rectified, and if so, then meets the requirements;

According to the trunk contour feature data obtained, judge whether shoulder flushes, and if so, then meets the requirements;

According to the eye feature data obtained, calculate the width opening part in the middle of eyes, if width is greater than preset value, then meet the requirements;

According to the face characteristic obtained, adopt edge detecting technology, judge on face whether be a curve in the middle of lower lip, if so, then illustrate that face closes, and meets the requirements, if not, then illustrate that face opens;

According to the portrait profile characteristic obtained, detect the diversity factor between portrait profile characteristic and default portrait profile feature, if diversity factor is less than preset value, and head color is without obvious segmentation, then illustrate without a hat on, meet the requirements;

According to the eye feature data obtained, calculate the colour cast COEFFICIENT K of eyes ₃with colourity mean value Da1, if colour cast COEFFICIENT K ₃be less than 1, then eyes colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₁be more than or equal to 1, and when colourity mean value Da1 is greater than 0, then there is blood-shot eye illness;

According to the facial contour obtained, eyes and face characteristic, detect color error ratio value, the color error ratio value of eye face position, the color error ratio value at face position of face, if the color error ratio value of one of them is greater than default deviate, then illustrate there is heavy make-up; If the color error ratio value of three is all less than or equal to default deviate, then illustrates without heavy make-up, meet the requirements;

Obtain face mask characteristic, detect the diversity factor between face mask characteristic and default face mask feature, if diversity factor is less than preset value, then hair does not cover face, meets the requirements.

Further, describedly judge whether face is rectified, and specifically comprises:

According to eye feature data, generate two lines, calculate the angle of two lines and certificate photograph horizontal direction, judge whether this angle exceeds preset range, if not, then illustrate and face camera;

According to eyes, nose and face characteristic, obtain the center point coordinate of two lines, obtain the centre coordinate of nose, obtain the centre coordinate of face, fitting a straight line is done to three points and obtains straight line LC, calculate the angle of this straight line LC and certificate photograph vertical direction, judge whether this angle exceeds preset range, if not, then illustrate that head is vertical;

According to facial contour feature data, obtain the center line FLC of human face region vertical direction, calculate the center line FLC of human face region vertical direction and the distance DC of straight line LC, calculate the ratio of this distance DC and human face region width, judge whether this ratio exceeds preset range, if not, then face horizontal center is described;

When facing the vertical and face horizontal center of camera, head, judging that face is proper, meeting the requirements.

Further, describedly judge whether shoulder flushes, and specifically comprises:

To trunk profile coordinate array, fit to a polygon, calculate the coordinate (X1 that polygonal center of gravity obtains center of gravity, Y1), calculate the distance Dweight of X1 and certificate photograph vertical center of gravity line, calculate the ratio of distance Dweight and certificate photograph width W F, judge whether this ratio exceeds preset range;

To trunk profile coordinate array, adopt minimum enclosed rectangle algorithm, obtain a minimum rectangle, calculate the angle of rectangular vertical center of gravity line and certificate photograph vertical center of gravity line, judge whether this angle exceeds preset range;

If the angle of the ratio of distance Dweight and certificate photograph width W F, rectangular vertical center of gravity line and certificate photograph vertical center of gravity line does not all exceed preset range, then shoulder flushes, and meets the requirements.

The present invention has following beneficial effect relative to prior art:

1, license camera of the present invention and method according to government's public security industry standard formulation every dimensions, and set up the standard database of dissimilar certificate photo, by obtaining the information (as file attribute information, portrait characteristic information, background information, monochrome information, chrominance information, sharpness information) of the rear certificate photograph of shooting, quality testing item by item being carried out to every dimensions, guaranteeing that the certificate photograph by detecting can meet the requirement of certificate photograph relevant criterion.

2, the dimensions that license camera of the present invention and method are formulated has 14, respectively: whether be recent photochrome, whether be not less than 24 RGB true color, whether be JPEG compress technique and compress quality whether meet the requirements, whether file size meets the requirements, whether the wide pixel of image image height meets the requirements, whether background color meets the requirements, photo surrounding whether Rimless, whether image is clear, whether face are visible, whether facial contour is without obviously distortion, whether brightness is evenly sufficient, image whether no color differnece, whether attitude is rectified, whether head portrait size and location meet the requirements, the visible complete detection achieved certificate photograph after shooting, make the requirement that the quality testing of certificate photograph reaches objective and accurate.

3, license camera of the present invention and method can detect the shooting quality of the certificate photograph after taking imaging automatically, by giving a mark and taking proposed mechanism, allow the certificate photograph do not met the demands block at source place, save the time of subsequent artefacts's one by onechecking, improve detection efficiency.

Accompanying drawing explanation

Fig. 1 is the license camera function module map of the embodiment of the present invention 1.

Fig. 2 is the portrait posture detecting unit structural drawing of the embodiment of the present invention 1.

Fig. 3 is the shooting quality detection method process flow diagram of the embodiment of the present invention 2.

Embodiment

Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited thereto.

Embodiment 1:

License camera is mounted in the APP (Application on intelligent terminal, application program), it can have Android version and iOS version, be responsible for shooting a width and meet the former figure of license that certificate making uses, the former figure of license refers to by user by license camera captured in real-time gained, according to the useful region of license standard cutting image, and the image informations such as color, brightness, background, personage's biological characteristic are without the view data of any process.

The license camera of the present embodiment for be the certificate photograph having taken imaging, quality testing can be carried out to certificate photograph, main according to government's public security industry standard formulation every dimensions, and set up the standard database of dissimilar certificate photo, the certificate photograph data message of acquisition and the data of database are compared, whether one by onechecking assessment meets the requirements, and calculate quality overall score according to testing result, when the score value of quality overall score does not reach preset value (as 60 points), then advise again taking.

Therefore, the license camera of the present embodiment comprises data obtaining module, quality detection module, grading module, quality overall score computing module and shooting suggestion module, as shown in Figure 1; The concrete function of modules is as follows:

Described data obtaining module, for the certificate photograph according to shooting imaging, obtains file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information; This data obtaining module comprises:

File attribute information acquiring unit, for obtaining the file attribute information of certificate photograph; Wherein, described file attribute information comprises shooting time, colored figure place, file suffixes, file size and image pixel;

Portrait characteristic acquisition unit, for obtaining the portrait characteristic information of certificate photograph; Wherein, described portrait characteristic comprises eyes, nose, ear, face, facial contour, trunk profile, portrait profile, the crown and chin characteristic; The extraction of portrait characteristic information can as follows:

1) certificate photograph is converted to gray level image;

2) upper part of the body characteristic region detection is carried out to gray level image, when upper part of the body characteristic region being detected, this area pixel information is saved as region rectangle data structure above the waist, and ROI region is above the waist arranged to gray level image, be designated as ROIB;

3) detection of facial contour feature data area is carried out to ROIB, when facial contour feature data area being detected, this area pixel information is saved as facial contour region rectangle data structure, and facial contour ROI region is arranged to gray level image, be designated as ROIF;

4) eyes characteristic region detection is carried out to the specific region of ROIF, when eyes characteristic region being detected, this area pixel information is saved as eyes region rectangle data structure, and obtain the coordinate of eyes, with the ROI region top that the Y-axis minimum point of eyes is new, facial contour feature data area is got to the region of below eyes, be set to nose target ROI region, be designated as ROIN;

5) ears characteristic region detection being carried out to ROIF, when ears characteristic region being detected, this area pixel information being saved as ears region rectangle data structure;

6) nose characteristic region detection is carried out to ROIN, when nose characteristic region being detected, this area pixel information is saved as nasal area rectangle data structure, using the base of nasal area rectangle data structure as the top margin of new ROI region, nose is got with lower area to facial contour feature data area, be set to face target ROI region, be designated as ROIM;

7) face characteristic region detection being carried out to ROIM, when face characteristic region being detected, this area pixel information being saved as face region rectangle data structure;

8) to the region of ROIB, cut out the region removing ROIF, obtain trunk contour feature data area ROIMB, this area pixel information is saved as trunk contour area rectangle data structure;

9) rim detection is carried out to ROIF, obtain the line segment at the sign mutation edge in contouring head regional extent, then carry out head convex closure contour detecting to this head edge line segment result, the coordinate of the key point obtained by head convex closure contour detecting saves as contouring head array;

10) rim detection is carried out to ROIMB, obtain the line segment at the sign mutation edge within the scope of trunk contour area, then carry out trunk convex closure contour detecting to this trunk edge line segment result, the coordinate of the key point obtained by trunk convex closure contour detecting saves as trunk profile array;

11) contouring head array and trunk number of contours group are combined, form portrait profile array.

By the extraction of above-mentioned portrait feature, namely eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic is obtained, crown characteristic can be obtained according to eye position, this part has detailed introduction in other patent document, do not repeat them here, the characteristic of chin is also obtain according to same principle.

Background information acquiring unit, for when background is not replaced, is converted to Lab color space image by certificate photograph, to Lab color space image, read each pixel of foreground area, set up the mixed Gauss model of prospect colourity; To each pixel of background area, the histogram of background extraction colourity, calculates the overlapping possibility of the histogram of background colourity and the mixed Gauss model of prospect colourity, to the accurate mean square value P of overlapping possibility label taking; Or for when background replaces to normal background, adopt edge detecting technology to detect portrait profile to certificate photograph, obtain the background information beyond certificate photograph portrait profile;

Monochrome information acquiring unit, for obtaining the brightness histogram of certificate photograph, carries out data analysis to brightness histogram, calculates the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA, specifically comprise:

1) from the image array of certificate photograph, read the brightness value (Y-component, Y (n)) of each pixel, add up the number of pixels of each brightness 0 ~ 255,128 is intermediate value, obtains the brightness histogram of certificate photograph;

2) data analysis is carried out to brightness histogram, calculates brightness integrated value:

$SumDA=\underset{0\&le;row<rows}{\underset{0\&le;col<cols}{\&Sigma;}}\left(Y\right(col,row)-128)$

Wherein, constant 128 is intensity deviation value; Col is the row of the image array of certificate photograph, and cols is maximum row; Row is the row of the image array of certificate photograph, and rows is maximum row; Certain any brightness value in the image array that Y (row, col) is certificate photograph; SumDA is the intensity deviation summation of each pixel relative 128;

3) the intensity deviation summation of trying to achieve is averaging sum of all pixels, obtains average luminance shift value, be expressed as DA:

DA＝SumDA/(cols*rows)

Be averaged the absolute value of intensity deviation value DA, remember D ₁:

D ₁＝abs(DA)

4) brightness histogram is an array HIST [i], under be designated as brightness value, scope is 0 to 255, and cell value is the number of pixels of this brightness value, obtains:

$Mda={\&Sigma;}_{i=0}^{255}abs(i-128-DA)*HIST\&lsqb;i\&rsqb;$

Wherein, Mda is the intensity deviation value summation of each brightness degree of brightness histogram;

5) Mda is taken absolute value, and averages:

M ₁＝abs(Mda)/(cols*rows)

6) according to M ₁, D ₁, try to achieve the abnormal coefficient of brightness:

K ₁＝D ₁/M ₁

Wherein, K ₁for the abnormal coefficient of brightness.

Chrominance information acquiring unit, for certificate photograph being converted to Lab color space image, adds up a axle of each pixel at Lab color space image and colourity mean value Da and Db of b axle, according to colourity mean value Da and Db, calculates colour cast COEFFICIENT K ₂, specifically comprise:

1) certificate photograph is converted to Lab color space image, adopts following formula to add up each pixel at a axle of Lab color space image and the colourity mean value Da (red/green estimated value partially) of b axle and Db (yellow/blue estimated value partially):

$Da=\left(\underset{0\&le;row<rows}{\underset{0\&le;col<cols}{\&Sigma;}}\right(Vec3b(row,col)\&lsqb;1\&rsqb;-128\left)\right)/(cols*rows)$

$Db=\left(\underset{0\&le;row<rows}{\underset{0\&le;col<cols}{\&Sigma;}}\right(Vec3b(row,col)\&lsqb;2\&rsqb;-128\left)\right)/(cols*rows)$

Wherein, the structure on the Lab color space that Vec3b (row, col) is a reading pixel, its first is brightness value, and second is a axle value of Lab, and the 3rd is the b axle value of Lab; Col is the row of Lab color space image array, and cols is maximum row; Row is the row of Lab color space image, and rows is maximum row; Constant 128 is intensity deviation value;

2) the variance yields D of a axle and b axle mean value is calculated ₂:

$D_2=\sqrt{{\mathrm{Da}}^2+{\mathrm{Db}}^2}$

3) brightness histogram HIST_A and HIST_B of computed image on a axle and b axle, two groups of brightness histograms have 256 statistical values respectively;

4) mean value Ma and Mb of a axle and b axle aberration is asked:

$Ma={\&Sigma;}_{i=0}^{255}(abs(i-128-Da)*HIST\_A\&lsqb;i\&rsqb;)/cols*rows$

$Mb={\&Sigma;}_{i=0}^{255}(abs(i-128-Db)*HIST\_B\&lsqb;i\&rsqb;)/cols*rows$

5) colour cast total value M is asked ₂:

$M_2=\sqrt{{\mathrm{Ma}}^2+{\mathrm{Mb}}^2}$

6) according to variance yields D ₂with colour cast total value M ₂, try to achieve colour cast coefficient:

K ₂＝D ₂/M ₂

Wherein, K ₂for colour cast coefficient.

Sharpness information acquiring unit, for certificate photograph is converted to gray level image, calculates sharpness factor DR, specifically comprises:

1) certificate photograph is converted to gray level image, sharpness meter calculates based on the basis of gray scale is implemented at last, sharpness also can be expressed as sharpness of vision, the sensation of people to the sharp keen degree of the image seen, through a large amount of experiments with research and analyse, the concept of generally acknowledging is that sharpness of vision is made up of resolution and object edge profile contrast two factors.Sharpness of vision is the steepness (slope) of transition, equals the change of change divided by position of output brightness;

In order to simplify calculating, adopt the excessive slope of brightness of two pixels and the mean value of the brightness interpolating of vertical four adjacent pixels that calculated level is adjacent, as the measurement index evaluating sharpness;

2) for a gray level image, suppose that pixel is following matrix:

ABX

CDX

XXX

Consecutive point are A, B, C and D;

3) two right-angle sides being right-angle triangle with the luminance difference of vertical direction in the horizontal direction with adjacent 4, calculate at adjacent 4 and change curvature with the brightness of vertical direction in the horizontal direction:

$\mathrm{\&theta;}1=\sqrt{{(B-A)}^2+{(C-A)}^2}$

And the luminance difference of adjacent with A 2:

Δ1＝abs(B-A)+abs(C-A)

4) following formula is adopted to calculate the luminance difference mean value of each point and its consecutive point in gray-scale map matrix:

$DR=\left({\&Sigma;}_{{}_{0\&le;row<rows}^{0\&le;col<cols}}\right(\mathrm{\&theta;}(row,col)+\mathrm{\&Delta;}(row,col)\left)\right)/(cols*rows)$

Wherein, θ (row, col) for adjacent 4 in the horizontal direction with the brightness of vertical direction change curvature, Δ (row, col) be and certain any luminance difference of adjacent 2, and col is gray-scale map matrix column, and cols is maximum row; Row is the row of gray-scale map matrix, and rows is maximum row, and DR is sharpness factor.

Described quality detection module, for according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carries out quality testing item by item to every dimensions; Whether wherein, every dimensions is the examination criteria according to government's public security industry standard formulation, comprises 14 sports, and the detection ordering of every dimensions can be arranged voluntarily by user, respectively: be 1) recent photochrome; 2) 24 RGB true color whether are not less than; 3) whether be JPEG compress technique, and whether compression quality meet the requirements; 4) whether file size meets the requirements; 5) whether the wide pixel of image image height meets the requirements; 6) whether background color meets the requirements; 7) photo surrounding whether Rimless; 8) whether focusing is accurate, and namely whether image is clear; 9) whether clear, the levels are rich of face, namely whether face are visible; 10) whether facial contour is without obviously distortion; 11) whether brightness is evenly sufficient, face's whether shadow-free, speck; 12) whether image color is balanced, and whether skin presents true tone, i.e. whether no color differnece; 13) whether attitude is rectified; 14) whether head portrait size and location meet the requirements; Therefore, this quality detection module comprises:

Shooting time detecting unit, for obtaining shooting time from file attribute information, judge whether that the recent times presetting (is generally set as 6 months, section sets is 3 months) interior shooting, if so, then photochrome is determined whether, if, then illustrate it is recent photochrome, meet the requirements;

Colored figure place detecting unit, for obtaining colored figure place from file attribute information, judging whether to be not less than 24 (being namely more than or equal to 24), if so, then meeting the requirements;

Compressed article quality detection unit, for obtaining file suffixes from file attribute information, judges whether file is compressed file, when file is compressed file, determines whether that JPEG compresses, and if so, judges whether compression quality meets the requirement of corresponding certificate photo, such as:

[I.D./residence permit/social security card] compressed article prime factor >=85;

[driver's license] compressed article prime factor >=85;

[entry and exit card] maximum compression multiple is no more than 25 times;

If meet above-mentioned standard, then prove that compression quality meets the requirements.

File size detecting unit, for obtaining file size from file attribute information, judges whether file size is greater than the preset value of corresponding certificate photo, if so, then meets the requirements; The requirement of different certificate photo is as follows:

[I.D./residence permit/social security card] photo files capacity > 30K;

[driver's license] photo files capacity > 30K;

[entry and exit card] photo files capacity > 80K.

Image pixel detecting unit, for obtaining image pixel from file attribute information, judges whether image pixel meets the requirement of corresponding certificate photo; The following is the requirement of different certificate photo:

[I.D./residence permit/social security card] wide > of image image height 441 (height) × 358 (wide) pixel (meet more than accreditation dpi350, meet stamp with the size > 26mm*32mm);

[driver's license] image image height wide > 378 (height) × 260 (wide) pixel (meet more than accreditation dpi300, meet stamp with the size > 32mm*22mm);

[entry and exit card] image image height wide > 640 pixel (height) × 480 pixel (wide) pixel (meets the wide > of 472 ~ 640 pixel < height > × 354 ~ 480 pixel <, meet the ratio of width to height 3:4, meet more than accreditation dpi300, meet stamp with the size > 33mm*48mm).

Background detection unit, for when background is not replaced, obtains standard mean square value P from background information, whether criterion mean square value P is greater than default mean square value (being generally set as 50), if not, then prospect and backcolor overlapping degree lower, meet the requirements; Or for when background replaces to normal background, judge whether the background information obtained meets the requirement of corresponding certificate photo, such as:

[I.D./residence permit/social security card] white background

[driver's license] white background

[entry and exit card] Guangdong blue background, other regional white background, white background tone value and intensity value are 0, and brightness value is 240; Light blue tone value is 135, and intensity value is 240, and brightness value is not less than 167;

Bounding box features detecting unit, for adopting edge detecting technology, detecting the whether well-regulated bounding box features of certificate photograph surrounding, if nothing, then meeting the requirements;

Brightness detection unit, for obtaining the abnormal COEFFICIENT K of brightness from monochrome information ₁with average luminance shift value DA, if the abnormal COEFFICIENT K of brightness ₁be less than 1, then brightness is normal, meets the requirements; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is greater than 0, then brightness is excessively bright; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is less than 0, then too dark brightness;

Colorimetric detection units, for obtaining colour cast COEFFICIENT K from chrominance information ₂, colourity mean value Da and Db, if colour cast COEFFICIENT K ₂be less than 1, then colourity is normal, and key diagram presents true tone as color balance, skin, meets the requirements; If colour cast COEFFICIENT K ₂be more than or equal to 1, when colourity mean value Da is greater than 0, represent partially red, when colourity mean value Da is less than 0, represent partially green, when colourity mean value Db is greater than 0, represent partially yellow, when colourity mean value Db is less than 0, represent partially blue;

Sharpness detecting unit, for obtaining sharpness factor DR from sharpness information, if sharpness factor DR is greater than default sharpness factor value (being generally set as 14), then it is clear to illustrate, focusing accurately, meets the requirements;

Face detecting unit, for obtaining eyes, nose, ear, face, the crown and chin characteristic from portrait characteristic information, by the eyes of acquisition, nose, ear, face, the crown and chin characteristic, compare with the standard feature preset, judge that whether face are complete, if so, then meet the requirements;

Facial contour detecting unit, for obtaining facial contour feature data from portrait characteristic information, detects the diversity factor between facial contour feature data and default facial contour feature, if diversity factor is greater than preset value, then distort, if diversity factor is less than or equal to preset value, then meet the requirements; Wherein, the diversity factor between described facial contour feature data and default facial contour feature is by the ratio calculation of non-overlapped area and overlapping area;

Portrait posture detecting unit, for obtaining eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic from portrait characteristic information, judges whether portrait attitude is rectified, and if so, then meets the requirements;

Head portrait size and location detecting unit, for obtaining eyes, ear, the crown and chin characteristic from portrait characteristic information, relevance calculating is carried out to eyes, ear, the crown and chin characteristic, obtaining the wide pixel of head, head high pixel, eyes distance pixel, eyes position distance up/down Edge Distance pixel and the crown gains fame and fortune apart from certificate photograph coboundary pixel, the parameter of these data and corresponding certificate photo is compared, judges whether to meet the requirements.

Wherein, various different certificate photo has different standards, as follows:

Wide 207 ± 14 pixels of [I.D./residence permit/social security card] head; Distance photograph lower edge, eyes position distance ≮ 207 pixels; Gain fame and fortune apart from photograph upper edge 7 ~ 21 pixel in the crown

Wide 165 ~ 189 pixels of [driver's license] head (meeting stamp with the size 14mm ~ 16mm); Height of head is (meet stamp with the size 19mm ~ 22mm) between 224 ~ 260 pixels; Gain fame and fortune apart from photograph upper edge 10 ~ 20 pixel in the crown

[entry and exit card] head width 189 ~ 283 pixel (meeting stamp with the size 16mm ~ 22mm, more than dpi300); Height of head is (meet stamp with the size 30mm ~ 34mm, more than dpi300) between 354 ~ 402 pixels; Interpupillary distance is 0.231 times of wide ~ 0.333 times wide pixel; Adult eye position is apart from high ~ 0.500,0.301 times, photo upper edge times high pixel (meeting stamp with the size 14mm ~ 22mm, more than dpi300); Children's eyes position is apart from high ~ 0.600,0.301 times, photo upper edge times high pixel (meeting stamp with the size 14mm ~ 26mm, more than dpi300); Gain fame and fortune apart from high ~ 0.074,0.025 times, photograph upper edge times high pixel in the crown.

As shown in Figure 2, above-mentioned portrait posture detecting unit, specifically comprises:

Face state-detection subelement, for according to the eyes obtained, nose, face and facial contour feature data, judges whether face is rectified, and if so, then meets the requirements; Judge whether face is rectified, and specifically comprises:

According to eye feature data, generate two lines, calculate the angle of two lines and certificate photograph horizontal direction, judge whether this angle exceeds preset range, if not, then illustrate and face camera;

According to eyes, nose and face characteristic, obtain the center point coordinate of two lines, obtain the centre coordinate of nose, obtain the centre coordinate of face, fitting a straight line is done to three points and obtains straight line LC, calculate the angle of this straight line LC and certificate photograph vertical direction, judge whether this angle exceeds preset range, if not, then illustrate that head is vertical;

According to facial contour feature data, obtain the center line FLC of human face region vertical direction, calculate the center line FLC of human face region vertical direction and the distance DC of straight line LC, calculate the ratio of this distance DC and human face region width, judge whether this ratio exceeds preset range, if not, then face horizontal center is described;

When facing the vertical and face horizontal center of camera, head, judging that face is proper, meeting the requirements.

Shoulder state-detection subelement, for according to the trunk contour feature data obtained, judges whether shoulder flushes, and if so, then meets the requirements; Judge whether shoulder flushes, and specifically comprises:

To trunk profile coordinate array, fit to a polygon, calculate the coordinate (X1 that polygonal center of gravity obtains center of gravity, Y1), calculate the distance Dweight of X1 and certificate photograph vertical center of gravity line, calculate the ratio of distance Dweight and certificate photograph width W F, judge whether this ratio exceeds preset range;

To trunk profile coordinate array, adopt minimum enclosed rectangle algorithm, obtain a minimum rectangle, calculate the angle of rectangular vertical center of gravity line and certificate photograph vertical center of gravity line, judge whether this angle exceeds preset range;

If the angle of the ratio of distance Dweight and certificate photograph width W F, rectangular vertical center of gravity line and certificate photograph vertical center of gravity line does not all exceed preset range, then shoulder flushes, and meets the requirements.

Eyes open width detection subelement, for according to the eye feature data obtained, calculate the width opening part in the middle of eyes, if width is greater than preset value, then meet the requirements.

Face is closed opens detection sub-unit, for according to the face characteristic obtained, adopts edge detecting technology, judges on face whether be a curve in the middle of lower lip, if so, then illustrates that face closes, and meets the requirements, and if not, then illustrates that face opens.

Detection sub-unit without a hat on, for according to the portrait profile characteristic obtained, detects the diversity factor between portrait profile characteristic and default portrait profile feature, if diversity factor is less than preset value, and head color is without obvious segmentation, then illustrates without a hat on, meet the requirements.

Redeye detection subelement, for according to the eye feature data obtained, calculates the colour cast COEFFICIENT K of eyes ₃with colourity mean value Da1 (with reference to above-mentioned colour cast COEFFICIENT K ₂with the computation process of colourity mean value Da), if colour cast COEFFICIENT K ₃be less than 1, then eyes colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₃be more than or equal to 1, and when colourity mean value Da1 is greater than 0, then there is blood-shot eye illness.

The color detection sub-unit of adornment, for the facial contour according to acquisition, eyes and face characteristic, detect color error ratio value, the color error ratio value of eye face position, the color error ratio value at face position of face, if the color error ratio value of one of them is greater than default deviate, then illustrate there is heavy make-up; If the color error ratio value of three is all less than or equal to default deviate, then illustrates without heavy make-up, meet the requirements; The calculating of the color error ratio value of face, the color error ratio value of eye face position, the color error ratio value at face position, can refer to above-mentioned colour cast COEFFICIENT K ₂computation process.

Face detection subelement, for obtaining face mask characteristic, detect the diversity factor between face mask characteristic and default face mask feature, if diversity factor is less than preset value, then hair does not cover face, meets the requirements.

Institute's scoring module, for the default score value according to every dimensions, marks to the quality measurements of every dimensions, obtains the score value of every dimensions; The default score value of every dimensions of the present embodiment is as follows:

1) whether be recent photochrome, when meeting, individual event, individual event score=(-40) point when not meeting if being 5 points;

2) whether be not less than 24 RGB true color, when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

3) whether be JPEG compress technique, and whether compression quality meets the requirements, when meeting, individual event, individual event score=(-40) point when not meeting if being 5 points;

4) whether file size meets the requirements, image file size >=type of credential minimum, and individual event is 5 points, individual event score=(-40) point during image file size < type of credential minimum;

5) whether the wide pixel of image image height meets the requirements, image image height >=type of credential minimum pixel, wide >=type of credential minimum pixel, individual event is 10 points, image image height < type of credential minimum pixel, during wide < type of credential minimum pixel, individual event score=(-40) point;

6) whether background color meets the requirements, and when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

7) photo surrounding whether Rimless, when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

8) whether focusing is accurate, and namely whether image is clear, and when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

9) whether clear, the levels are rich of face, when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

10) whether facial contour is without obviously distortion, and when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

11) whether brightness is evenly sufficient, face's whether shadow-free, speck, and when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

12) whether image color is balanced, and whether skin presents true tone, and when meeting, individual event is 5 points, individual event score=(-40) point when not meeting;

13) whether attitude is rectified, and all detections meet to obtain 15 points, individual event score=(-40) point when not meeting;

14) whether head portrait size and location meet the requirements, and obtain 20 points when all detections meet, individual event score=(-40) point when not meeting;

Described quality overall score computing module, for the score value according to every dimensions, and scoring weight, calculate the quality overall score of certificate photograph; When general comment score value is lower than 0 timesharing, overall score all assigns to record and display by 0.

Described shooting suggestion module, for the quality overall score according to certificate photograph, provides the shooting suggestion of specialty, such as:

1) 90-100 divides: senior photographer: your shooting effect is very good, can enter next step!

2) 80--89 divides: intermediate photographer: your shooting effect is fine, can enter next step!

3) 60-79 divides: elementary photographer: your shooting effect is general, then claps one and have a try.

4) 0-59 divides: photography green hand: your shooting effect is not good enough, and suggestion is taken again.

Embodiment 2:

As shown in Figure 3, present embodiments provide a kind of shooting quality detection method, the method for be shooting imaging, but do not replace the certificate photograph of background shooting quality detect, it is applied in license camera, comprises the following steps:

S1, according to shooting imaging certificate photograph, obtain file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, specifically comprise:

The file attribute information of S101, acquisition certificate photograph; Wherein, described file attribute information comprises shooting time, colored figure place, file suffixes, file size and image pixel;

The portrait characteristic information of S102, acquisition certificate photograph; Wherein, described portrait characteristic comprises eyes, nose, ear, face, facial contour, trunk profile, portrait profile, the crown and chin characteristic;

S103, certificate photograph is converted to Lab color space image, to Lab color space image, reads each pixel of foreground area, set up the mixed Gauss model of prospect colourity; To each pixel of background area, the histogram of background extraction colourity, calculates the overlapping possibility of the histogram of background colourity and the mixed Gauss model of prospect colourity, to the accurate mean square value P of overlapping possibility label taking;

The brightness histogram of S104, acquisition certificate photograph, carries out data analysis to brightness histogram, calculates the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA;

S105, certificate photograph is converted to Lab color space image, adds up a axle of each pixel at Lab color space image and colourity mean value Da and Db of b axle, according to colourity mean value Da and Db, calculate colour cast COEFFICIENT K ₂;

S106, certificate photograph is converted to gray level image, calculates sharpness factor DR;

S2, the file attribute information according to obtaining, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carry out quality testing item by item to every dimensions, specific as follows:

S201, from file attribute information, obtain shooting time, judge whether to take in the recent times preset, if so, then determine whether photochrome, if so, then meet the requirements; If not take in the recent times preset, or when not being photochrome, then undesirable;

S202, from file attribute information, obtain colored figure place, judge whether to be not less than 24, if so, then meet the requirements, if not, then undesirable;

S203, from file attribute information, obtain file suffixes, judge whether file is compressed file, when file is compressed file, determines whether that JPEG compresses, if so, judge to compress the requirement whether quality meets corresponding certificate photo;

S204, from file attribute information, obtain file size, judge whether file size is greater than the preset value of corresponding certificate photo, if so, then meets the requirements, if not, then undesirable;

S205, from file attribute information, obtain image pixel, judge whether image pixel meets the requirement of corresponding certificate photo;

S206, from background information, obtain standard mean square value P, whether criterion mean square value P is greater than default mean square value, if not, then meets the requirements; If so, then undesirable;

S207, employing edge detecting technology, detect the whether well-regulated bounding box features of certificate photograph surrounding, if nothing, then meet the requirements; If have, then undesirable;

S208, from monochrome information, obtain the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA, if the abnormal COEFFICIENT K of brightness ₁be less than 1, then brightness is normal, meets the requirements; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is greater than 0, then brightness is excessively bright; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is less than 0, then too dark brightness; When brightness is crossed bright or excessively dark, all undesirable;

S209, from chrominance information, obtain colour cast COEFFICIENT K ₂, colourity mean value Da and Db, if colour cast COEFFICIENT K ₂be less than 1, then colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₂be more than or equal to 1, when colourity mean value Da is greater than 0, represent partially red, when colourity mean value Da is less than 0, represent partially green, when colourity mean value Db is greater than 0, represent partially yellow, when colourity mean value Db is less than 0, represent partially blue; When partially red, partially green, partially yellow or partially blue, all undesirable;

S210, from sharpness information, obtain sharpness factor DR, if sharpness factor DR is greater than default sharpness factor value, then it is clear to illustrate, meets the requirements; If sharpness factor DR is less than or equal to default sharpness factor value, then undesirable;

S211, from portrait characteristic information, obtain eyes, nose, ear, face, the crown and chin characteristic, by the eyes of acquisition, nose, ear, face, the crown and chin characteristic, compare with the standard feature preset, judge that whether face are complete, if, then meet the requirements, if not, then undesirable;

S212, from portrait characteristic information, obtain facial contour feature data, detect the diversity factor between facial contour feature data and default facial contour feature, if diversity factor is greater than preset value, then distort, undesirable; If diversity factor is less than or equal to preset value, then meet the requirements;

S213, from portrait characteristic information, obtain eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic, judge whether portrait attitude is rectified, and if so, then meets the requirements, if not, then undesirable;

S214, from portrait characteristic information, obtain eyes, ear, the crown and chin characteristic, relevance calculating is carried out to eyes, ear, the crown and chin characteristic, obtaining the wide pixel of head, head high pixel, eyes distance pixel, eyes position distance up/down Edge Distance pixel and the crown gains fame and fortune apart from certificate photograph coboundary pixel, the parameter of these data and corresponding certificate photo is compared, judges whether to meet the requirements;

S3, default score value according to every dimensions, mark to the quality measurements of dimensions every in above-mentioned steps S201 ~ S214, obtain the score value of every dimensions;

S4, score value according to every dimensions, and scoring weight, calculate the quality overall score of certificate photograph;

S5, quality overall score according to certificate photograph, provide the shooting suggestion of specialty.

Embodiment 3:

The shooting quality detection method of the present embodiment for be shooting imaging, and the shooting quality replacing to the certificate photograph of normal background detects, and the difference part of itself and embodiment 2 is:

S103, adopt edge detecting technology to detect portrait profile to certificate photograph, obtain the background information beyond certificate photograph portrait profile;

S206, judge whether the background information obtained meets the requirement of corresponding certificate photo.

In sum, license camera of the present invention and method be every dimensions according to government's public security industry standard formulation, and set up the standard database of dissimilar certificate photo, by obtaining the information of the rear certificate photograph of shooting, quality testing item by item being carried out to every dimensions, guaranteeing that the certificate photograph by detecting can meet the requirement of certificate photograph relevant criterion.

The above; be only patent preferred embodiment of the present invention; but the protection domain of patent of the present invention is not limited thereto; anyly be familiar with those skilled in the art in the scope disclosed in patent of the present invention; be equal to according to the technical scheme of patent of the present invention and inventive concept thereof and replace or change, all belonged to the protection domain of patent of the present invention.

Claims (12)

1. can detect a license camera for shooting quality, it is characterized in that: described license camera comprises:

Data obtaining module, for the certificate photograph according to shooting imaging, obtains file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information;

Quality detection module, for according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carries out quality testing item by item to every dimensions; Wherein, every dimensions is the examination criteria according to government's public security industry standard formulation;

Grading module, for the default score value according to every dimensions, marks to the quality measurements of every dimensions, obtains the score value of every dimensions;

Quality overall score computing module, for the score value according to every dimensions, and scoring weight, calculate the quality overall score of certificate photograph;

Shooting suggestion module, for the quality overall score according to certificate photograph, provides the shooting suggestion of specialty.

2. a kind of license camera detecting shooting quality according to claim 1, is characterized in that: described data obtaining module comprises:

File attribute information acquiring unit, for obtaining the file attribute information of certificate photograph; Wherein, described file attribute information comprises shooting time, colored figure place, file suffixes, file size and image pixel;

Portrait characteristic acquisition unit, for obtaining the portrait characteristic information of certificate photograph; Wherein, described portrait characteristic comprises eyes, nose, ear, face, facial contour, trunk profile, portrait profile, the crown and chin characteristic;

Background information acquiring unit, for when background is not replaced, is converted to Lab color space image by certificate photograph, to Lab color space image, read each pixel of foreground area, set up the mixed Gauss model of prospect colourity; To each pixel of background area, the histogram of background extraction colourity, calculates the overlapping possibility of the histogram of background colourity and the mixed Gauss model of prospect colourity, to the accurate mean square value P of overlapping possibility label taking; Or for when background replaces to normal background, adopt edge detecting technology to detect portrait profile to certificate photograph, obtain the background information beyond certificate photograph portrait profile;

Monochrome information acquiring unit, for obtaining the brightness histogram of certificate photograph, carries out data analysis to brightness histogram, calculates the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA;

Chrominance information acquiring unit, for certificate photograph being converted to Lab color space image, adds up a axle of each pixel at Lab color space image and colourity mean value Da and Db of b axle, according to colourity mean value Da and Db, calculates colour cast COEFFICIENT K ₂;

Sharpness information acquiring unit, for certificate photograph is converted to gray level image, calculates sharpness factor DR.

3. a kind of license camera detecting shooting quality according to claim 2, is characterized in that: described quality detection module comprises:

Shooting time detecting unit, for obtaining shooting time from file attribute information, judging whether to take in the recent times preset, if so, then determining whether photochrome, if so, then meet the requirements;

Colored figure place detecting unit, for obtaining colored figure place from file attribute information, judging whether to be not less than 24, if so, then meeting the requirements;

Compressed article quality detection unit, for obtaining file suffixes from file attribute information, judges whether file is compressed file, when file is compressed file, determines whether that JPEG compresses, and if so, judges whether compression quality meets the requirement of corresponding certificate photo;

File size detecting unit, for obtaining file size from file attribute information, judges whether file size is greater than the preset value of corresponding certificate photo, if so, then meets the requirements;

Image pixel detecting unit, for obtaining image pixel from file attribute information, judges whether image pixel meets the requirement of corresponding certificate photo;

Background detection unit, for when background is not replaced, obtain standard mean square value P from background information, whether criterion mean square value P is greater than default mean square value, if not, then meets the requirements; Or for when background replaces to normal background, judge whether the background information obtained meets the requirement of corresponding certificate photo;

Bounding box features detecting unit, for adopting edge detecting technology, detecting the whether well-regulated bounding box features of certificate photograph surrounding, if nothing, then meeting the requirements;

Brightness detection unit, for obtaining the abnormal COEFFICIENT K of brightness from monochrome information ₁with average luminance shift value DA, if the abnormal COEFFICIENT K of brightness ₁be less than 1, then brightness is normal, meets the requirements; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is greater than 0, then brightness is excessively bright; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is less than 0, then too dark brightness;

Colorimetric detection units, for obtaining colour cast COEFFICIENT K from chrominance information ₂, colourity mean value Da and Db, if colour cast COEFFICIENT K ₂be less than 1, then colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₂be more than or equal to 1, when colourity mean value Da is greater than 0, represent partially red, when colourity mean value Da is less than 0, represent partially green, when colourity mean value Db is greater than 0, represent partially yellow, when colourity mean value Db is less than 0, represent partially blue;

Sharpness detecting unit, for obtaining sharpness factor DR from sharpness information, if sharpness factor DR is greater than default sharpness factor value, then it is clear to illustrate, meets the requirements;

Face detecting unit, for obtaining eyes, nose, ear, face, the crown and chin characteristic from portrait characteristic information, by the eyes of acquisition, nose, ear, face, the crown and chin characteristic, compare with the standard feature preset, judge that whether face are complete, if so, then meet the requirements;

Facial contour detecting unit, for obtaining facial contour feature data from portrait characteristic information, detects the diversity factor between facial contour feature data and default facial contour feature, if diversity factor is greater than preset value, then distort, if diversity factor is less than or equal to preset value, then meet the requirements; Wherein, the diversity factor between described facial contour feature data and default facial contour feature is by the ratio calculation of non-overlapped area and overlapping area;

Portrait posture detecting unit, for obtaining eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic from portrait characteristic information, judges whether portrait attitude is rectified, and if so, then meets the requirements;

Head portrait size and location detecting unit, for obtaining eyes, ear, the crown and chin characteristic from portrait characteristic information, relevance calculating is carried out to eyes, ear, the crown and chin characteristic, obtaining the wide pixel of head, head high pixel, eyes distance pixel, eyes position distance up/down Edge Distance pixel and the crown gains fame and fortune apart from certificate photograph coboundary pixel, the parameter of these data and corresponding certificate photo is compared, judges whether to meet the requirements.

4. a kind of license camera detecting shooting quality according to claim 3, is characterized in that: described portrait posture detecting unit comprises:

Face state-detection subelement, for according to the eyes obtained, nose, face and facial contour feature data, judges whether face is rectified, and if so, then meets the requirements;

Shoulder state-detection subelement, for according to the trunk contour feature data obtained, judges whether shoulder flushes, and if so, then meets the requirements;

Eyes open width detection subelement, for according to the eye feature data obtained, calculate the width opening part in the middle of eyes, if width is greater than preset value, then meet the requirements;

Face is closed opens detection sub-unit, for according to the face characteristic obtained, adopts edge detecting technology, judges on face whether be a curve in the middle of lower lip, if so, then illustrates that face closes, and meets the requirements, and if not, then illustrates that face opens;

Detection sub-unit without a hat on, for according to the portrait profile characteristic obtained, detects the diversity factor between portrait profile characteristic and default portrait profile feature, if diversity factor is less than preset value, and head color is without obvious segmentation, then illustrates without a hat on, meet the requirements;

Redeye detection subelement, for according to the eye feature data obtained, calculates the colour cast COEFFICIENT K of eyes ₃with colourity mean value Da1, if colour cast COEFFICIENT K ₃be less than 1, then eyes colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₃be more than or equal to 1, and when colourity mean value Da1 is greater than 0, then there is blood-shot eye illness;

The color detection sub-unit of adornment, for the facial contour according to acquisition, eyes and face characteristic, detect color error ratio value, the color error ratio value of eye face position, the color error ratio value at face position of face, if the color error ratio value of one of them is greater than default deviate, then illustrate there is heavy make-up; If the color error ratio value of three is all less than or equal to default deviate, then illustrates without heavy make-up, meet the requirements;

Face detection subelement, for obtaining face mask characteristic, detect the diversity factor between face mask characteristic and default face mask feature, if diversity factor is less than preset value, then hair does not cover face, meets the requirements.

5. a kind of license camera detecting shooting quality according to claim 4, is characterized in that: described face state-detection subelement judges whether face is rectified, and specifically comprises:

According to eye feature data, generate two lines, calculate the angle of two lines and certificate photograph horizontal direction, judge whether this angle exceeds preset range, if not, then illustrate and face camera;

According to eyes, nose and face characteristic, obtain the center point coordinate of two lines, obtain the centre coordinate of nose, obtain the centre coordinate of face, fitting a straight line is done to three points and obtains straight line LC, calculate the angle of this straight line LC and certificate photograph vertical direction, judge whether this angle exceeds preset range, if not, then illustrate that head is vertical;

According to facial contour feature data, obtain the center line FLC of human face region vertical direction, calculate the center line FLC of human face region vertical direction and the distance DC of straight line LC, calculate the ratio of this distance DC and human face region width, judge whether this ratio exceeds preset range, if not, then face horizontal center is described;

When facing the vertical and face horizontal center of camera, head, judging that face is proper, meeting the requirements.

6. a kind of license camera detecting shooting quality according to claim 4, is characterized in that: described shoulder state-detection subelement judges whether shoulder flushes, and specifically comprises:

To trunk profile coordinate array, fit to a polygon, calculate the coordinate (X1 that polygonal center of gravity obtains center of gravity, Y1), calculate the distance Dweight of X1 and certificate photograph vertical center of gravity line, calculate the ratio of distance Dweight and certificate photograph width W F, judge whether this ratio exceeds preset range;

To trunk profile coordinate array, adopt minimum enclosed rectangle algorithm, obtain a minimum rectangle, calculate the angle of rectangular vertical center of gravity line and certificate photograph vertical center of gravity line, judge whether this angle exceeds preset range;

If the angle of the ratio of distance Dweight and certificate photograph width W F, rectangular vertical center of gravity line and certificate photograph vertical center of gravity line does not all exceed preset range, then shoulder flushes, and meets the requirements.

7. a shooting quality detection method, is applied in license camera, it is characterized in that: described method comprises:

Described license camera, according to the certificate photograph of shooting imaging, obtains file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information;

Described license camera, according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, carries out quality testing item by item to every dimensions; Wherein, every dimensions is the examination criteria according to government's public security industry standard formulation;

Described license camera, according to the default score value of every dimensions, is marked to the quality measurements of every dimensions, obtains the score value of every dimensions;

Described license camera is according to the score value of every dimensions, and scoring weight, calculates the quality overall score of certificate photograph;

Described license camera, according to the quality overall score of certificate photograph, provides the shooting suggestion of specialty.

8. a kind of shooting quality detection method according to claim 7, it is characterized in that: described license camera is according to the certificate photograph of shooting imaging, obtain file attribute information, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, specifically comprise:

Obtain the file attribute information of certificate photograph; Wherein, described file attribute information comprises shooting time, colored figure place, file suffixes, file size and image pixel;

Obtain the portrait characteristic information of certificate photograph; Wherein, described portrait characteristic comprises eyes, nose, ear, face, facial contour, trunk profile, portrait profile, the crown and chin characteristic;

When background is not replaced, certificate photograph is converted to Lab color space image, to Lab color space image, reads each pixel of foreground area, set up the mixed Gauss model of prospect colourity; To each pixel of background area, the histogram of background extraction colourity, calculates the overlapping possibility of the histogram of background colourity and the mixed Gauss model of prospect colourity, to the accurate mean square value P of overlapping possibility label taking; Or when background replaces to normal background, adopt edge detecting technology to detect portrait profile to certificate photograph, obtain the background information beyond certificate photograph portrait profile;

Obtain the brightness histogram of certificate photograph, data analysis is carried out to brightness histogram, calculate the abnormal COEFFICIENT K of brightness ₁with average luminance shift value DA;

Certificate photograph is converted to Lab color space image, adds up a axle of each pixel at Lab color space image and colourity mean value Da and Db of b axle, according to colourity mean value Da and Db, calculate colour cast COEFFICIENT K ₂;

Certificate photograph is converted to gray level image, calculates sharpness factor DR.

9. a kind of shooting quality detection method according to claim 8, it is characterized in that: described license camera is according to the file attribute information obtained, portrait characteristic information, background information, monochrome information, chrominance information and sharpness information, quality testing is item by item carried out to every dimensions, specifically comprises:

From file attribute information, obtain shooting time, judge whether to take in the recent times preset, if so, then determine whether photochrome, if so, then meet the requirements;

From file attribute information, obtain colored figure place, judge whether to be not less than 24, if so, then meet the requirements;

From file attribute information, obtain file suffixes, judge whether file is compressed file, when file is compressed file, determines whether that JPEG compresses, if so, judge whether compression quality meets the requirement of corresponding certificate photo;

From file attribute information, obtain file size, judge whether file size is greater than the preset value of corresponding certificate photo, if so, then meets the requirements;

From file attribute information, obtain image pixel, judge whether image pixel meets the requirement of corresponding certificate photo;

When background is not replaced, obtain standard mean square value P from background information, whether criterion mean square value P is greater than default mean square value, if not, then meets the requirements; Or for when background replaces to normal background, judge whether the background information obtained meets the requirement of corresponding certificate photo;

Adopt edge detecting technology, detect the whether well-regulated bounding box features of certificate photograph surrounding, if nothing, then meet the requirements;

The abnormal COEFFICIENT K of brightness is obtained from monochrome information ₁with average luminance shift value DA, if the abnormal COEFFICIENT K of brightness ₁be less than 1, then brightness is normal, meets the requirements; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is greater than 0, then brightness is excessively bright; If the abnormal COEFFICIENT K of brightness ₁be more than or equal to 1, and average luminance shift value DA is less than 0, then too dark brightness;

Colour cast COEFFICIENT K is obtained from chrominance information ₂, colourity mean value Da and Db, if colour cast COEFFICIENT K ₂be less than 1, then colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₂be more than or equal to 1, when colourity mean value Da is greater than 0, represent partially red, when colourity mean value Da is less than 0, represent partially green, when colourity mean value Db is greater than 0, represent partially yellow, when colourity mean value Db is less than 0, represent partially blue;

From sharpness information, obtain sharpness factor DR, if sharpness factor DR is greater than default sharpness factor value, then it is clear to illustrate, meets the requirements;

Eyes, nose, ear, face, the crown and chin characteristic is obtained from portrait characteristic information, by the eyes of acquisition, nose, ear, face, the crown and chin characteristic, compare with the standard feature preset, judge that whether face are complete, if so, then meet the requirements;

From portrait characteristic information, obtain facial contour feature data, detect the diversity factor between facial contour feature data and default facial contour feature, if diversity factor is greater than preset value, then distort, if diversity factor is less than or equal to preset value, then meet the requirements; Wherein, the diversity factor between described facial contour feature data and default facial contour feature is by the ratio calculation of non-overlapped area and overlapping area;

From portrait characteristic information, obtain eyes, nose, ear, face, facial contour, trunk profile, portrait profile characteristic, judge whether portrait attitude is rectified, and if so, then meets the requirements;

Eyes, ear, the crown and chin characteristic is obtained from portrait characteristic information, relevance calculating is carried out to eyes, ear, the crown and chin characteristic, obtaining the wide pixel of head, head high pixel, eyes distance pixel, eyes position distance up/down Edge Distance pixel and the crown gains fame and fortune apart from certificate photograph coboundary pixel, the parameter of these data and corresponding certificate photo is compared, judges whether to meet the requirements.

10. a kind of shooting quality detection method according to claim 9, is characterized in that: describedly judge whether portrait attitude is rectified, and specifically comprises:

According to the eyes obtained, nose, face and facial contour feature data, judge whether face is rectified, and if so, then meets the requirements;

According to the trunk contour feature data obtained, judge whether shoulder flushes, and if so, then meets the requirements;

According to the eye feature data obtained, calculate the width opening part in the middle of eyes, if width is greater than preset value, then meet the requirements;

According to the face characteristic obtained, adopt edge detecting technology, judge on face whether be a curve in the middle of lower lip, if so, then illustrate that face closes, and meets the requirements, if not, then illustrate that face opens;

According to the portrait profile characteristic obtained, detect the diversity factor between portrait profile characteristic and default portrait profile feature, if diversity factor is less than preset value, and head color is without obvious segmentation, then illustrate without a hat on, meet the requirements;

According to the eye feature data obtained, calculate the colour cast COEFFICIENT K of eyes ₃with colourity mean value Da1, if colour cast COEFFICIENT K ₃be less than 1, then eyes colourity is normal, meets the requirements; If colour cast COEFFICIENT K ₁be more than or equal to 1, and when colourity mean value Da1 is greater than 0, then there is blood-shot eye illness;

According to the facial contour obtained, eyes and face characteristic, detect color error ratio value, the color error ratio value of eye face position, the color error ratio value at face position of face, if the color error ratio value of one of them is greater than default deviate, then illustrate there is heavy make-up; If the color error ratio value of three is all less than or equal to default deviate, then illustrates without heavy make-up, meet the requirements;

Obtain face mask characteristic, detect the diversity factor between face mask characteristic and default face mask feature, if diversity factor is less than preset value, then hair does not cover face, meets the requirements.

11. a kind of shooting quality detection methods according to claim 10, is characterized in that: describedly judge whether face is rectified, and specifically comprises:

According to eye feature data, generate two lines, calculate the angle of two lines and certificate photograph horizontal direction, judge whether this angle exceeds preset range, if not, then illustrate and face camera;

According to eyes, nose and face characteristic, obtain the center point coordinate of two lines, obtain the centre coordinate of nose, obtain the centre coordinate of face, fitting a straight line is done to three points and obtains straight line LC, calculate the angle of this straight line LC and certificate photograph vertical direction, judge whether this angle exceeds preset range, if not, then illustrate that head is vertical;

According to facial contour feature data, obtain the center line FLC of human face region vertical direction, calculate the center line FLC of human face region vertical direction and the distance DC of straight line LC, calculate the ratio of this distance DC and human face region width, judge whether this ratio exceeds preset range, if not, then face horizontal center is described;

When facing the vertical and face horizontal center of camera, head, judging that face is proper, meeting the requirements.

12. a kind of shooting quality detection methods according to claim 10, is characterized in that: describedly judge whether shoulder flushes, and specifically comprises:

To trunk profile coordinate array, fit to a polygon, calculate the coordinate (X1 that polygonal center of gravity obtains center of gravity, Y1), calculate the distance Dweight of X1 and certificate photograph vertical center of gravity line, calculate the ratio of distance Dweight and certificate photograph width W F, judge whether this ratio exceeds preset range;

To trunk profile coordinate array, adopt minimum enclosed rectangle algorithm, obtain a minimum rectangle, calculate the angle of rectangular vertical center of gravity line and certificate photograph vertical center of gravity line, judge whether this angle exceeds preset range;

If the angle of the ratio of distance Dweight and certificate photograph width W F, rectangular vertical center of gravity line and certificate photograph vertical center of gravity line does not all exceed preset range, then shoulder flushes, and meets the requirements.