CN101131727B - Image collecting method and its application - Google Patents

Abstract

The invention relates to a kind of method to collect images and its application. The method to collect the images includes: setting the camera and optical-reflect device in the pre-established place and with the pre-established angle; the camera collects the images of the shot objects in the optical-reflect device. The system to collect the images includes: the camera and the optical-reflect device. The method to recognize the faces of persons includes collecting the images and recognizing the persons' faces according the collected images, of which the progress to collect the images includes setting the camera and optical-reflect device in the pre-established place and with the pre-established angle; the camera collects the images that is reflected into the optical-reflect device by the check persons. The system to recognize the persons' faces includes camera, the optical-reflect device and the person-face recognizer which includes image-turning module. The invention enlarges the effective covering range of the image the camera can get with the optical-reflect device, and get the aim to improve the accuracy to recognize the faces of persons.

Description

A kind of method of image acquisition and application thereof

Technical field

The present invention relates to the Flame Image Process applied technical field, relate in particular to a kind of method and application of image acquisition.

Background technology

At present, acquisition technology is widely used, especially in the face recognition technology field.

Face recognition technology has directly, friendly, characteristics easily, is easier to be accepted by the user, so recognition of face becomes in the current feature identification one of technology of widespread use the most.

At present, the general process of recognition of face comprises the detection of people's face, eye location, feature extraction and four key steps of feature comparison/people face contrast as shown in Figure 1.

At present, the recognition of face device needing to be mainly used in the scene that the user cooperates, such as in systems such as gate inhibition, examination, because specific environment for use, the user can cooperate with the work of recognition of face device on one's own initiative, video camera can obtain whole people's face easily, and the recognition of face device is finished identifying according to the facial image that obtains then.

But under the ill-matched situation of user, human recognizer just is not easy to carry out smoothly work, because in this case, the recognition of face device can not or be difficult to obtain whole people's face, such as, on the airport or customs, the offender understands conscious seizure of hiding video camera, the facial image that makes video camera catch is imperfect, and the recognition of face device can't be discerned incomplete facial image, like this, just make the recognition of face device not play a role to greatest extent, make the recognition of face accuracy rate reduce greatly.

In addition, in the long and narrow spaces such as inspection passage of airport or customs, even if offender or examinee painstakingly do not hide the seizure of video camera, but because the limitation of video camera physical location, the effective coverage and the angle that can cause it to catch image are all smaller, make the accuracy rate of recognition of face reduce greatly equally.

In this case, can enlarge effective coverage and the angle that it catches image by setting up multiple cameras, but, output to a plurality of video cameras when this will seek common ground is handled, and cause the treatment capacity of background process server and the visit capacity of face database all to increase greatly, and existing equipment can't satisfy such requirement.Simultaneously, all compare costliness because the cost of video camera itself and background devices are changed the cost of upgrading, thereby this way can not be widely used.

Summary of the invention

In view of above-mentioned existing in prior technology problem, the purpose of this invention is to provide a kind of method and application thereof of image acquisition, the present invention has enlarged the effective coverage range of camera pickuping image by optical reflection unit, thereby reaches the purpose that improves the recognition of face accuracy rate under condition cheaply.

The objective of the invention is to be achieved through the following technical solutions:

The invention provides a kind of method of image acquisition, comprising:

A, video camera and optical reflection unit are laid by predefined position and angle, described video camera is relative with the reflecting surface of described optical reflection unit;

The image of B, camera acquisition subject and the subject image in optical reflection unit.

Wherein, predefined position and angle are meant described in the described steps A: the image of the subject in the known camera acquisition passage, described channel width is L, the maximum height of subject is Xa, minimum constructive height is Xm, and when the range of video camera coverage was 2 δ, the position that video camera is laid was at D place, the place ahead, optical reflection unit center, vertical optical reflection unit distance is W, overhead highly is H;

The angle that video camera is laid is: the elevation angle is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

The zone of setting camera pickuping image is first effective coverage, the zone of virtual video camera images acquired is second effective coverage, wherein virtual video camera is the picture that video camera becomes in optical reflection unit, and the 3rd effective coverage is the common factor of first effective coverage and second effective coverage

Video camera is gathered the image of subject in first effective coverage, and/or

In second effective coverage, gather the image of subject in optical reflection unit, and/or

Gather image and/or the image of subject in optical reflection unit that subject is absorbed by video camera in the 3rd effective coverage.

The present invention also comprises a kind of image capturing system, and described system comprises: video camera and optical reflection unit, described video camera is relative with the reflecting surface of described optical reflection unit

Optical reflection unit:, manifest the image of subject by the optical reflection principle;

Video camera: image and the image of subject in optical reflection unit of being responsible for gathering subject.

And, the image of the subject in known camera acquisition passage, described channel width is L, the maximum height of subject is Xa, and minimum constructive height is Xm, when the range of video camera coverage is 2 δ, described position for video camera is in D place, the place ahead, described optical reflection unit center, apart from optical reflection unit vertical range W place, overhead highly be H, the elevation angle of video camera is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

Described optical reflection unit comprises level crossing or convex reflecting mirror.

The present invention also comprises a kind of method of recognition of face, comprises image acquisition, and carries out recognition of face according to the image of gathering, and described image acquisition process comprises:

A, video camera and optical reflection unit are laid by predefined position and angle, made video camera relative with the reflecting surface of described optical reflection unit;

B, the detected personnel's of camera acquisition image and the image of detected personnel in optical reflection unit.

Predefined position and angle are meant described in the described steps A:

The image of detected personnel in the known camera acquisition passage, described channel width is L, detected personnel's maximum height is Xa, minimum constructive height is Xm, when the range of video camera coverage is 2 δ, the position that video camera is laid is at D place, the place ahead, optical reflection unit center, and vertical optical reflection unit distance is W, overhead highly is H;

The angle that video camera is laid is: the elevation angle is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

The zone of setting camera pickuping image is first effective coverage, the zone of virtual video camera images acquired is second effective coverage, wherein virtual video camera is the picture that video camera becomes in optical reflection unit, and the 3rd effective coverage is the common factor of first effective coverage and second effective coverage

Video camera absorbs detected personnel's image in first effective coverage, and/or

In second effective coverage, gather the image of detected personnel in optical reflection unit, and/or

The image and/or the image of detected personnel in optical reflection unit of the detected personnel that collect in the 3rd effective coverage camera picked-up being shot.

The detected personnel's that normally absorb when video camera image or camera acquisition to image and the face template result that compare one of of detected personnel in optical reflection unit surpass when setting thresholding, then contrast successfully.

And, describedly carry out also comprising before the recognition of face carrying out according to the image of gathering: with camera acquisition to the image of detected personnel in optical reflection unit about overturn.

The present invention also comprises a kind of system of recognition of face, it is characterized in that, described system comprises: video camera, optical reflection unit and recognition of face device, described video camera is relative with the reflecting surface of described optical reflection unit

Optical reflection unit:, manifest detected personnel's image by the optical reflection principle;

Video camera: image and the image of detected personnel in optical reflection unit of being responsible for gathering detected personnel;

The recognition of face device: by camera acquisition to image carry out recognition of face.

And, the image of detected personnel in known camera acquisition passage, described channel width is L, detected personnel's maximum height is Xa, and minimum constructive height is Xm, when the range of video camera coverage is 2 δ, described position for video camera is in D place, the place ahead, described optical reflection unit center, apart from optical reflection unit vertical range W place, overhead highly be H, the elevation angle of video camera is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

Described recognition of face device also comprises the image flip module, be responsible for camera acquisition to optical reflection unit in image carry out about upset, the recognition of face device according to about image after the upset carry out recognition of face.

Described optical reflection unit comprises level crossing or convex reflecting mirror.

As seen from the above technical solution provided by the invention, among the present invention, video camera and optical reflection unit are laid by predefined relation, like this, video camera is except can normally collecting detected personnel's image, can also collect the image of examinee in optical reflection unit, therefore, the present invention is comparing with prior art, enlarge the effective range of camera pickuping image, thereby improved the recognition of face accuracy rate, adopted the present invention, can also reduce the quantity of video camera, reduce the performance requirement of identified server, thereby reach the purpose of saving cost.

Description of drawings

Fig. 1 is the synoptic diagram of face recognition process in the prior art;

Fig. 2 obtains the synoptic diagram of facial image for the video camera of the embodiment of the invention;

Fig. 3 sets up synoptic diagram for the camera orientation of the embodiment of the invention;

The minute surface of Fig. 4 position embodiment of the invention and video camera schematic cross-section;

Fig. 5 sets up the orientation synoptic diagram for the virtual video camera of the embodiment of the invention;

Fig. 6 is the synoptic diagram of the face recognition process of the embodiment of the invention;

Fig. 7 obtains the synoptic diagram of facial image for the video camera of another embodiment of the present invention.

Embodiment

Core concept of the present invention is by video camera and optical reflection unit are laid according to predefined relation, pass through optical reflection unit, video camera has enlarged the effective range of its pickup image, has improved the recognition of face accuracy rate, thereby effectively overcomes the existing in prior technology problem.

To accompanying drawing 7, is example with mirror as optical reflection unit below in conjunction with accompanying drawing 2, and the method and system of image acquisition of the present invention, the method and system of recognition of face are elaborated.

At first the method to a kind of image acquisition of the present invention is elaborated.

The method of image acquisition of the present invention comprises:

At first, video camera and mirror are installed according to predefined angle and position;

As shown in Figure 2, the zone of camera pickuping image is the effective coverage 1 shown in the figure, according to optical imaging concept, the picture that video camera becomes in mirror, it is virtual video camera, the zone of its pickup image is the effective coverage 2 shown in the figure, and the common factor of note effective coverage 1 and effective coverage 2 is effective coverage 3.

Respectively the characteristic of effective coverage 1, effective coverage 2 and effective coverage 3 is described in detail below.

1, be in effective coverage 1 as the examinee, but when being not effective coverage 3, the shooting function is directly caught people's face, but the examinee painstakingly or painstakingly do not evade under the situation of video camera, the people's face that captures all might be imperfect.If this moment, the examinee was towards video camera, video camera will capture complete people's face, thereby human recognizer can carry out accurate recognition according to the people's face that captures.

2, be in effective coverage 2 as the examinee, but when being not effective coverage 3, according to image-forming principle, virtual video camera can capture people's face, but the examinee painstakingly or painstakingly do not evade under the situation of video camera, the people's face that captures all might be imperfect.If this moment, the examinee was towards virtual video camera, virtual video camera will capture complete people's face, and the recognition of face device can accurately be discerned according to the people's face that captures equally.

3, when the examinee is in effective coverage 3, examinee's image can appear in video camera and the virtual video camera simultaneously, that is, video camera can capture examinee and the imaging of examinee in mirror simultaneously.At this moment, no matter the position of examinee and video camera relation is how, whole people's face of examinee be certain to appear at video camera and virtual video camera the two one of in, perhaps appear in the two simultaneously.At this moment, even if the examinee painstakingly or not painstakingly evade video camera, when mutual occlusion issue took place, video camera can capture complete people's face equally.

When mutual occlusion issue takes place, be that examinee's periphery is when having other people to block video camera to the picked-up of examinee's facial image, at this moment shooting possibly can't capture people's face and maybe can't capture complete people's face, but according to the pinhole imaging system principle, if virtual video camera is regarded as the aperture of imaging, the scope that captures of virtual video camera is bigger than the scope that real camera captures so, so from intuitively describing, just being equivalent to photograph two row people from eminence is advancing, so people's face that virtual video camera captures can not blocked, that is,, still can capture complete people's face by virtual video camera.

Elaborate the relativeness of video camera and mirror among the present invention below.

As shown in Figure 3, suppose that known channel width is L, camera pedestal is located at D place, the place ahead, minute surface center, and its perpendicular mirrors identity distance is from being W, and overhead high H, the range of video camera coverage are 2 δ.From statistical figure, the maximum height that can learn detected personnel is Xa simultaneously, and minimum height is Xm.By above-mentioned known parameters, can calculate following unknown parameter, concrete solution procedure is as follows:

As shown in Figure 4, for convenience of description, now intercept the plane at video camera and minute surface axis place, because W and D are orthogonal, if promptly with W and D two right-angle sides as right-angle triangle, the ratio of establishing hypotenuse and W then can obtain for going into $\mathrm{\λ}=\frac{\sqrt{D^2+W^2}}{W}$ (formula 1), and then draw the distance of minute surface center to H $\mathrm{W\λ}=\sqrt{D^2+W^2}.$

According to above data, can extrapolate the elevation angle formula of video camera on the cross section:

$\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}}$ Simultaneously $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}}$ (formula 2)

When $\mathrm{\α}=\arctan \[\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}}\]$ The time, this moment, α was the minimum rate of accumulation elevation angle of video camera on the cross section.

Usually, the span of D is $D\≤\sqrt{W^2+H^2}$ (formula 3), but, should make the optical axis meeting of video camera and virtual video camera vertical mutually, promptly in order to obtain complete face as far as possible on a large scale $D=\sqrt{W^2+H^2}$ Because this moment effective coverage 3 the scope maximum.

As shown in Figure 5, for virtual video camera, can draw following relation:

$\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}$ (formula 4)

By five restrictive conditions in formula 1, formula 2, formula 3 and the formula 4, thereby can calculate unknown parameter D, W, H, λ and angle α.

After video camera and mirror are laid according to position recited above and angle, the image of video camera picked-up has comprised examinee's image and/or the image of examinee in optical reflection unit, and, upset about the image of examinee in optical reflection unit that collects carried out.

The system of image acquisition of the present invention comprises: optical reflection unit and video camera.

Optical reflection unit: utilize the optical reflection principle, detected personnel's image is apparent in the optical reflection unit, this moment, the image in optical reflection unit was opposite with its true picture.

Video camera: in described system, be used for images acquired, described image comprises detected person person's image and the image of detected personnel in optical reflection unit.Described video camera and described optical reflection unit are laid according to predefined relation, and like this, video camera can also absorb the image of detected personnel in optical reflection unit except can normally absorbing detected personnel's the image.

Method and system to a kind of recognition of face of the present invention are elaborated then.

The method of recognition of face of the present invention comprises:

At first, video camera and mirror are installed according to predefined angle and position;

As shown in Figure 2, the zone of camera pickuping image is the effective coverage 1 shown in the figure, according to optical imaging concept, the picture that video camera becomes in mirror, it is virtual video camera, the zone of its pickup image is the effective coverage 2 shown in the figure, and the common factor of note effective coverage 1 and effective coverage 2 is effective coverage 3.

Respectively the characteristic of effective coverage 1, effective coverage 2 and effective coverage 3 is described in detail below.

1, be in effective coverage 1 as the examinee, but when being not effective coverage 3, the shooting function is directly caught people's face, but the examinee painstakingly or painstakingly do not evade under the situation of video camera, the people's face that captures all might be imperfect.If this moment, the examinee was towards video camera, video camera will capture complete people's face, thereby human recognizer can carry out accurate recognition according to the people's face that captures.

2, be in effective coverage 2 as the examinee, but when being not effective coverage 3, according to image-forming principle, virtual video camera can capture people's face, but the examinee painstakingly or painstakingly do not evade under the situation of video camera, the people's face that captures all might be imperfect.If this moment, the examinee was towards virtual video camera, virtual video camera will capture complete people's face, and the recognition of face device can accurately be discerned according to the people's face that captures equally.

3, when the examinee is in effective coverage 3, examinee's image can appear in video camera and the virtual video camera simultaneously, that is, video camera can capture examinee and the imaging of examinee in mirror simultaneously.At this moment, no matter the position of examinee and video camera relation is how, whole people's face of examinee be certain to appear at video camera and virtual video camera the two one of in, perhaps appear in the two simultaneously.At this moment, even if the examinee painstakingly or not painstakingly evade video camera, when mutual occlusion issue took place, video camera can capture complete people's face equally.

When mutual occlusion issue takes place, be that examinee's periphery is when having other people to block video camera to the picked-up of examinee's facial image, at this moment shooting possibly can't capture people's face and maybe can't capture complete people's face, but according to the pinhole imaging system principle, if virtual video camera is regarded as the aperture of imaging, the scope that captures of virtual video camera is bigger than the scope that real camera captures so, so from intuitively describing, just being equivalent to photograph two row people from eminence is advancing, so people's face that virtual video camera captures can not blocked, that is,, still can capture complete people's face by virtual video camera.

Elaborate the relativeness of video camera and mirror among the present invention below.

As shown in Figure 3, suppose that known channel width is L, camera pedestal is located at D place, the place ahead, minute surface center, and its perpendicular mirrors identity distance is from being W, and overhead high H, the range of video camera coverage are 2 δ.From statistical figure, the maximum height that can learn detected personnel is Xa simultaneously, and minimum height is Xm.By above-mentioned known parameters, can calculate following unknown parameter, concrete solution procedure is as follows:

As shown in Figure 4, describe for the aspect, now intercept the plane at video camera and minute surface axis place, because W and D are orthogonal, if promptly with W and D two right-angle sides as right-angle triangle, the ratio of establishing hypotenuse and W then can obtain for going into $\mathrm{\λ}=\frac{\sqrt{D^2+W^2}}{W}$ (formula 1), and then draw the distance of minute surface center to H $\mathrm{W\λ}=\sqrt{D^2+W^2}.$

According to above data, can extrapolate the elevation angle formula of video camera on the cross section:

$\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}}$ Simultaneously $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}}$ (formula 2)

When $\mathrm{\α}=\arctan \[\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}}\]$ The time, this moment, α was the minimum rate of accumulation elevation angle of video camera on the cross section.

Usually, the span of D is $D\≤\sqrt{W^2+H^2}$ (formula 3), but, should make the optical axis meeting of video camera and virtual video camera vertical mutually, promptly in order to obtain complete face as far as possible on a large scale $D=\sqrt{W^2+H^2},$ Because this moment effective coverage 3 the scope maximum.

As shown in Figure 5, for virtual video camera, can draw following relation:

$\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}$ (formula 4)

By five restrictive conditions in formula 1, formula 2, formula 3 and the formula 4, thereby can calculate unknown parameter D, W, H, λ and angle α.

After video camera and mirror are laid according to position recited above and angle, the image of video camera picked-up has comprised examinee's image and/or the image of examinee in optical reflection unit, then, human recognizer carries out recognition of face by the image of video camera picked-up, specifies below.

The process of present recognition of face generally comprises the detection of people's face, eye location, feature extraction, four key steps of feature comparison/people face contrast as shown in Figure 1.In face recognition process of the present invention, except people's face to normally obtaining, the people's face that is the detected personnel that normally collect of video camera is done beyond the feature extraction, also will with camera acquisition to optical reflection unit in people's face carry out about the upset, obtain another feature (this is because because tested personnel's image is a image with tested personnel is opposite) in mirror, and after feature that will normally obtain or the upset feature that obtains respectively with database in face template compare, when one of two features surpass the setting thresholding, compare successfully, illustrate that the examinee is exactly the recognition objective of the other recognizer of people, otherwise, the comparison failure illustrates that the examinee is not the recognition objective of recognition of face device.

Certainly, when the face characteristic that will extract from optical reflection unit and face template compare, also can the face characteristic of obtaining not processed, but face template overturns, then the face template after described face characteristic and the upset is compared.

In addition, upset can also be arranged in after the camera acquisition image except being included in the feature extraction step about above-mentioned people's face, and people's face detects before the step.

The system of recognition of face of the present invention comprises: video camera, optical reflection unit and recognition of face device.

Optical reflection unit: utilize the optical reflection principle, detected personnel's image is apparent in the optical reflection unit, this moment, the image in optical reflection unit was opposite with its true picture.

Video camera: in described system, be used for images acquired, described image comprises detected person person's image and the image of detected personnel in optical reflection unit.Described video camera and described optical reflection unit are laid according to predefined relation, and like this, video camera can also absorb the image of detected personnel in optical reflection unit except can normally absorbing detected personnel's the image.

The recognition of face device: be used for receiving camera acquisition to detected personnel's image and detected personnel at the image of optical reflection unit, and carry out recognition of face according to described image.

And, at known channel width is L, tested personnel's maximum height is Xa, minimum constructive height is Xm, when the range of video camera coverage is 2 δ, described video camera is placed in D place, the place ahead, described optical reflection unit center, vertical optical reflection unit distance is W, overhead highly be H, and adjust the angle of video camera, making its elevation angle is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

The concrete derivation of the relative position relation of video camera and mirror and formula had been carried out detailed description when discussing the method for the invention, just repeat no more at this.

Recognition of face device of the present invention also comprises the image flip module, described image flip module be used for camera acquisition to the facial image of optical reflection unit carry out about upset, the recognition of face device according to about image after the upset carry out recognition of face.

Promptly, in face recognition process of the present invention, except people's face of normally obtaining being done feature extraction, also to overturn about with the people's face in the optical reflection unit by the image flip module, obtain another feature (this is because because tested personnel's image is a image with tested personnel is opposite) in mirror, and the feature that obtains and the face template of the database in the recognition of face device are compared after feature that will normally obtain or the upset, when one of two features surpass a setting thresholding, compare successfully, illustrate that the examinee is exactly the recognition objective of the other recognizer of people, otherwise, the comparison failure illustrates that the examinee is not the recognition objective of recognition of face device.

Certainly, when the face characteristic that will extract from optical reflection unit and face template compare, can the face characteristic of obtaining not processed yet, but utilize the image flip module that face template is overturn, then the face template after described face characteristic and the upset is compared.

As shown in Figure 7,, also can obtain bigger picked-up scope here with convex reflecting mirror.And in this case, the relation between camera and the mirror is just freer.Yet because convex reflecting mirror can bring certain pattern distortion, the accuracy rate of recognition of face can have decline slightly, but from whole real-time recognition process, convex reflecting mirror still can promote the accuracy rate of recognition of face.

And the optical reflection unit among the present invention can also comprise a plurality of mirrors, and according to the optical reflection principle, the picture that detected personnel become in a mirror can also repeatedly imaging in other mirror, and like this, video camera just can obtain bigger picked-up scope.

Above-mentioned to explanation of the present invention in, be example as optical reflection unit only with mirror, method and system of the present invention are described, be not limited to mirror but those of ordinary skill in the art should be understood that described optical reflection unit.In addition, the present invention is that the passage with bounded is that example describes, but in fact, the present invention is equally applicable to open passage or zone.

The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.

Claims (11)

1. the method for an image acquisition is characterized in that, comprising:

A, video camera and optical reflection unit are laid by predefined position and angle, made video camera relative with the reflecting surface of described optical reflection unit;

The image of B, camera acquisition subject and the subject image in optical reflection unit;

Wherein, predefined position and angle are meant described in the steps A:

Described subject is on the passage, described channel width is L, the maximum height of subject is Xa, minimum constructive height is Xm, when the range of video camera coverage was 2 δ, the position that video camera is laid was at D place, the place ahead, optical reflection unit center, and vertical optical reflection unit distance is W, overhead highly be H, λ is as the hypotenuse of the right-angle triangle of two right-angle sides and the ratio of W with W and D;

The angle that video camera is laid is: the elevation angle is α; Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

2. the method for a kind of image acquisition according to claim 1, it is characterized in that, the zone of setting camera pickuping image is first effective coverage, the zone of virtual video camera images acquired is second effective coverage, wherein virtual video camera is the picture that video camera becomes in optical reflection unit, the 3rd effective coverage is the common factor of first effective coverage and second effective coverage

Described step B specifically comprises:

In second effective coverage, gather the image of subject in optical reflection unit,

And/or

Gather image and/or the image of subject in optical reflection unit that the body of subject is absorbed by video camera in the 3rd effective coverage.

3. the system of an image acquisition, it is characterized in that, described system comprises: video camera and optical reflection unit, described video camera and described optical reflection unit are laid by following position and angle: subject is on the passage, described channel width is L, the maximum height of subject is Xa, minimum constructive height is Xm, when the range of video camera coverage is 2 δ, the position that video camera is laid is at D place, the place ahead, optical reflection unit center, vertical optical reflection unit distance is W, overhead highly is H, and λ is as the hypotenuse of the right-angle triangle of two right-angle sides and the ratio of W with W and D;

The angle that video camera is laid is: the elevation angle is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L};$

Optical reflection unit:, manifest the image of subject by the optical reflection principle;

Video camera: image and the image of subject in optical reflection unit of gathering subject.

4. the system of a kind of image acquisition according to claim 3 is characterized in that, described optical reflection unit comprises level crossing or convex reflecting mirror.

5. the method for a recognition of face is characterized in that, described method specifically comprises:

A, video camera laid by predefined position and angle with optical reflection unit make video camera relative with the reflecting surface of described optical reflection unit;

B, the detected personnel's of camera acquisition image and the image of detected personnel in optical reflection unit;

C, carry out recognition of face according to the described image that collects;

Wherein, predefined position and angle are meant described in the steps A:

Described detected personnel are on the passage, described channel width is L, detected personnel's maximum height is Xa, minimum constructive height is Xm, when the range of video camera coverage was 2 δ, the position that video camera is laid was at D place, the place ahead, optical reflection unit center, and vertical optical reflection unit distance is W, overhead highly be H, λ is as the hypotenuse of the right-angle triangle of two right-angle sides and the ratio of W with W and D;

The angle that video camera is laid is: the elevation angle is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L}.$

6. the method for a kind of recognition of face according to claim 5, it is characterized in that, the zone that the setting video camera is taken the diagram is first effective coverage, the zone that virtual video camera is gathered figure is second effective coverage, wherein virtual video camera is the picture that video camera becomes in optical reflection unit, the 3rd effective coverage is the common factor of first effective coverage and second effective coverage

Described step B specifically comprises:

In second effective coverage, gather the image of detected personnel in optical reflection unit,

And/or

Gather image and/or the image of detected personnel in optical reflection unit that detected personnel's body is absorbed by video camera in the 3rd effective coverage.

7. the method for a kind of recognition of face according to claim 6 is characterized in that, described step C specifically comprises:

The detected personnel's that normally absorb when video camera image or camera acquisition to image and the face template result that compare one of of detected personnel in optical reflection unit surpass when setting thresholding, then contrast successfully.

8. a kind of face identification method according to claim 5, it is characterized in that described step C is carrying out also comprising before the recognition of face according to the described image that collects: with camera acquisition to the image of detected personnel in optical reflection unit about overturn.

9. the system of a recognition of face, it is characterized in that, described system comprises: video camera, optical reflection unit and recognition of face device, described video camera and described optical reflection unit are laid by following position and angle: detected personnel are on the passage, described channel width is L, detected personnel's maximum height is Xa, minimum constructive height is Xm, when the range of video camera coverage is 2 δ, the position that video camera is laid is at D place, the place ahead, optical reflection unit center, vertical optical reflection unit distance is W, overhead highly is H, and λ is as the hypotenuse of the right-angle triangle of two right-angle sides and the ratio of W with W and D;

The angle that video camera is laid is: the elevation angle is α;

Above-mentioned parameter D, W, H, λ, α calculate by following formula:

$\mathrm{\λ}=\frac{\sqrt[2]{D^2+W^2}}{W},$ $\tan (\mathrm{\α}+\mathrm{\δ})\≤\frac{\mathrm{W\λ}}{H-\mathrm{Xa}},$ $\tan \left(\mathrm{\α}\right)\≤\frac{(L-W)\mathrm{\λ}}{H-\mathrm{Xm}},$ $D\≤\sqrt{W^2+H^2},$ $\frac{H-\mathrm{Xa}}{\mathrm{Xa}-\mathrm{Xm}}=\frac{W}{L};$

Optical reflection unit:, manifest detected personnel's image by the optical reflection principle;

Video camera: image and the image of detected personnel in optical reflection unit of gathering detected personnel;

The recognition of face device: according to camera acquisition to image carry out recognition of face.

10. the system of a kind of recognition of face according to claim 9, it is characterized in that, described recognition of face device also comprises the image flip module, be responsible for camera acquisition to optical reflection unit in image carry out about upset, the recognition of face device according to about image after the upset carry out recognition of face.

11. the system according to claim 9 or 10 described a kind of recognitions of face is characterized in that described optical reflection unit comprises level crossing or convex reflecting mirror.

Images