CN113034764B

CN113034764B - Access control method, device, equipment and access control system

Info

Publication number: CN113034764B
Application number: CN201911345981.2A
Authority: CN
Inventors: 丁旭; 胡文泽
Original assignee: Shenzhen Intellifusion Technologies Co Ltd
Current assignee: Shenzhen Intellifusion Technologies Co Ltd
Priority date: 2019-12-24
Filing date: 2019-12-24
Publication date: 2023-03-03
Anticipated expiration: 2039-12-24
Also published as: CN113034764A

Abstract

The embodiment of the invention provides an access control method, an access control device, access control equipment and an access control system, wherein the method comprises the following steps: detecting the face of a visiting user in the access control area through a first face detection model, and judging whether the size of a first face to be identified is larger than a preset first size threshold value; if the size of the first face to be recognized is smaller than the preset first size threshold, recognizing the first face to be recognized through a second face recognition model to obtain a pre-recognition result; based on the pre-recognition result, recognizing a second face to be recognized after a preset time interval to obtain a second recognition result, wherein the size of the second face to be recognized is larger than that of a first face to be recognized, and the first face to be recognized and the second face to be recognized are faces of the same visiting user; and performing access control according to the second identification result. Through carrying out the pre-recognition to the people face, when visiting the user and arriving entrance guard's machine near, entrance guard's machine can quick response, improves user experience.

Description

Access control method, device, equipment and access control system

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to an access control method, an access control device, access control equipment and an access control system.

Background

Along with the development of artificial intelligence, the image recognition technology is widely applied to intelligent security, for example, an access control mechanism based on face recognition, the access control mechanism carries out face recognition on a visiting user at a door through a camera, the door is opened and released when the face recognition is passed, and the door is kept closed when the face recognition is not passed. Because entrance guard's deployment scene is mostly open-type scene, someone wants to pass through entrance guard, someone is only the entrance guard of passing by, for guaranteeing not the mistake and opening the door, current entrance guard machine need come the customer just to carry out face identification when walking near entrance guard machine (also the face size that extracts needs to be enough big), but because entrance guard machine's face identification includes face detection, extract facial features, processes such as feature matching, need certain arithmetic processing time, make visiting user need wait in front of entrance guard machine. Therefore, the response speed of the existing access control mechanism is low due to the fact that the face recognition speed of the existing access control mechanism is low, and therefore a visiting user needs to wait, and user experience is poor.

Disclosure of Invention

The embodiment of the invention provides an access control method which can improve the response speed of an access control mechanism and improve user experience.

In a first aspect, an embodiment of the present invention provides an access control method, including:

detecting the face of a visiting user in the access control area through a first face detection model, and judging whether the size of a first face to be identified is larger than a preset first size threshold value;

if the size of the first face to be recognized is smaller than the preset first size threshold, recognizing the first face to be recognized through a second face recognition model to obtain a pre-recognition result;

based on the pre-recognition result, recognizing a second face to be recognized after a preset time interval to obtain a second recognition result, wherein the size of the second face to be recognized is larger than that of a first face to be recognized, and the first face to be recognized and the second face to be recognized are faces of the same visiting user;

and performing access control according to the second recognition result.

Optionally, the method further includes:

if the size of the first face to be recognized is larger than the preset first size threshold, recognizing the first face to be recognized through a first face recognition model to obtain a first recognition result;

and performing access control according to the first recognition result.

Optionally, the first face recognition model includes: first face feature extraction engine, second face feature extraction engine, first face storehouse and second face storehouse, wherein, the precision of the first face feature that first face feature extraction engine extracted is less than the precision of the second face feature that second face feature extraction engine extracted, the face characteristic data precision of storage in the first face storehouse is less than the face characteristic data precision of storage in the second face storehouse, first face feature be used for match in the first face storehouse, second face feature be used for match in the second face storehouse, if the size of first face of awaiting discerning is greater than predetermined first size threshold value, then to through first face identification model first face of awaiting discerning is discerned, obtains first recognition result, includes:

if the size of the first face to be recognized is larger than the preset first size threshold, inputting the first face to be recognized into the first face feature extraction engine for first feature extraction, and extracting to obtain a first face feature to be recognized;

matching the first face features to be recognized in the first face library, and judging whether the matching is successful;

if the matching is successful, outputting the successful matching as a first recognition result;

if the matching fails, inputting the first face to be recognized into the second face feature extraction engine for second feature extraction, and extracting to obtain second face features to be recognized;

matching the second face features to be recognized in the second face library, and judging whether the matching is successful;

and if the matching fails, outputting the matching failure as a first recognition result.

Optionally, the recognizing the second face to be recognized after the interval of the preset time includes:

if the size of the first face to be recognized is smaller than the preset first size threshold, after the preset time interval, performing face detection on a visiting user in the entrance guard area through the second face detection model, and judging whether the size of the second face to be recognized is larger than a preset second size threshold or not;

and if the size of the second face to be recognized is larger than the preset second size threshold, recognizing the second face to be recognized through a second face recognition model.

Optionally, the second face recognition model includes: third face feature extraction engine, fourth face feature extraction engine, third face storehouse and fourth face storehouse, wherein, the precision that third face feature extracted by third face feature extraction engine is greater than the precision that fourth face feature extracted by fourth face feature extraction engine was extracted, the face characteristic data precision of storage in the third face storehouse is greater than the face characteristic data precision of storage in the fourth face storehouse, third face feature is used for matching in the third face storehouse, fourth face feature is used for matching in the fourth face storehouse, it is right to discern through second face identification model the first face that treats discernment includes:

respectively inputting the first face to be recognized into a third face feature extraction engine and a fourth face feature extraction engine for feature extraction, and respectively extracting to obtain a third face feature to be recognized and a fourth face feature to be recognized;

storing the third face feature to be recognized and the fourth face feature to be recognized as a pre-recognition result; or

And matching the third face features to be recognized in the third face library to obtain a third recognition result, and storing the third recognition result and the fourth face features to be recognized as pre-recognition results.

Optionally, the recognizing, based on the pre-recognition result, the second face to be recognized at intervals of a preset time to obtain a second recognition result, includes:

if the size of the second face to be recognized is larger than the preset second size threshold, performing feature extraction on the second face to be recognized through the fourth face feature extraction engine to extract fifth face features to be recognized;

comparing the similarity of the fifth to-be-recognized face feature with the similarity of the fourth to-be-recognized face feature, and judging whether the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature;

if the fifth to-be-recognized face feature is not similar to the fourth to-be-recognized face feature, matching the fifth to-be-recognized face feature in the fourth face library, and judging whether matching is successful;

if the matching is successful, outputting the successful matching as a second recognition result;

if the matching fails, inputting the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extracting to obtain sixth face features to be recognized;

matching the sixth face features to be recognized in the third face library, and judging whether the matching is successful;

and if the matching fails, outputting the matching failure as a second recognition result.

Optionally, the recognizing a second face to be recognized at preset time intervals based on the pre-recognition result to obtain a second recognition result, further includes:

if the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature, judging that the third recognition result is successful in matching or failed in matching;

if the third recognition result is successful in matching, outputting the successful matching as a second recognition result; or alternatively

If the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature, matching the third to-be-recognized face feature in the third face library, and judging whether matching is successful;

if the matching fails or the third recognition result is matching failure, inputting the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extracting to obtain sixth face features to be recognized;

In a second aspect, an embodiment of the present invention provides an access control apparatus, including:

the first detection module is used for detecting the face of a visiting user in the access control area through a first face detection model and judging whether the size of a first face to be identified is larger than a preset first size threshold value or not;

the first recognition module is used for recognizing the first face to be recognized through a second face recognition model to obtain a pre-recognition result if the size of the first face to be recognized is smaller than the preset first size threshold;

the second recognition module is used for recognizing a second face to be recognized after a preset time interval based on the pre-recognition result to obtain a second recognition result, wherein the size of the second face to be recognized is larger than that of a first face to be recognized, and the first face to be recognized and the second face to be recognized are faces of the same visiting user;

and the control module is used for carrying out access control according to the second identification result.

In a third aspect, an embodiment of the present invention provides an access control device, including: the invention relates to a door access control method, in particular to a door access control method, which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the steps in the door access control method provided by the embodiment of the invention are realized when the processor executes the computer program.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program implements steps in the access control method provided in the embodiment of the present invention.

In a fifth aspect, an embodiment of the present invention provides an access control system, including:

the entrance guard machine is used for releasing or blocking the visiting user; and

the image acquisition device is used for acquiring image information of an entrance guard area;

the access control device is used for executing the access control method according to the image information so as to control the access controller.

In the embodiment of the invention, a first face detection model is used for carrying out face detection on a visiting user in an entrance guard area, and whether the size of a first face to be recognized is larger than a preset first size threshold value is judged; if the size of the first face to be recognized is smaller than the preset first size threshold, recognizing the first face to be recognized through a second face recognition model to obtain a pre-recognition result; based on the pre-recognition result, recognizing a second face to be recognized after a preset time interval to obtain a second recognition result, wherein the size of the second face to be recognized is larger than that of a first face to be recognized, and the first face to be recognized and the second face to be recognized are faces of the same visiting user; and performing access control according to the second identification result. Through carrying out the pre-recognition to the people face that face size is less than first size threshold value, after the interval default time, carry out the second time discernment based on the result of pre-discernment again, make to discern the second time and can go on the basis of pre-discernment, the speed of second time discernment has been improved, be equivalent to once discerning in entrance guard's machine distantly to the visitor, discern again near the entrance guard's machine, because when discerning near the entrance guard's machine again, be based on the result of distantly discerning goes on, can improve the speed of second time discernment, thereby make the visitor to arrive when the entrance guard's machine is near, entrance guard's machine can quick response, and user experience is improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

Fig. 1 is a flowchart of an access control method according to an embodiment of the present invention;

fig. 2 is a flowchart of another access control method provided in the embodiment of the present invention;

fig. 3 is a flowchart of another access control method according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of an access control device according to an embodiment of the present invention;

fig. 5 is a schematic structural diagram of another access control device provided in the embodiment of the present invention;

fig. 6 is a schematic structural diagram of another access control device provided in an embodiment of the present invention;

fig. 7 is a schematic structural diagram of another access control device provided in the embodiment of the present invention;

fig. 8 is a schematic structural diagram of another access control device provided in the embodiment of the present invention;

fig. 9 is a schematic structural diagram of another access control device provided in the embodiment of the present invention;

fig. 10 is a schematic structural diagram of another access control device provided in an embodiment of the present invention;

fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.

Referring to fig. 1, fig. 1 is a flowchart of an access control method according to an embodiment of the present invention, as shown in fig. 1, including the following steps:

101. and performing face detection on the visiting user in the access control area through the first face detection model, and judging whether the size of the first face to be recognized is larger than a preset first size threshold value.

The first face detection model is a pre-trained face detection model, and the first face detection model can extract a face in a face frame when the face is detected so as to obtain a first face to be recognized. The number of the first faces to be recognized may be one or more, when one face is detected by the first detection model, one face is output as the first face to be recognized, and when a plurality of faces are detected by the first detection model, a plurality of faces are output as the first face to be recognized.

The access control area refers to a field area of the image acquisition device, namely, an area that the image acquisition device can shoot. The image acquisition equipment can be arranged on the access control machine and is also independently arranged near the access control machine.

The visiting user refers to a person entering the access control area, and further refers to a person entering the field range of the image acquisition device and being shot by the image acquisition device. The person may be a person photographed at a distance or a person photographed at a near distance.

The first face to be recognized extracted through the first face detection model comprises size information of the face, the first face to be recognized extracted through the first face detection model is compared with a preset first size threshold, and whether the size of the first face to be recognized is larger than the preset first size threshold or not is judged. When the size of the first face to be recognized is larger than a preset first size threshold, it is indicated that the visiting user corresponding to the first face to be recognized is close to the access control machine, and when the size of the first face to be recognized is smaller than the preset first size threshold, it is indicated that the visiting user corresponding to the face to be recognized is far away from the access control machine.

102. And if the size of the first face to be recognized is smaller than a preset first size threshold, recognizing the first face to be recognized through the second face recognition model to obtain a pre-recognition result.

In this step, when the visiting user is far away from the access control machine, the size of the first face to be recognized extracted by the first face detection model is smaller than a preset first size threshold, and the first face to be recognized can be pre-recognized. The pre-recognition result can be used for assisting the face recognition of a visiting user when the visiting user is close to the access control machine.

The second face recognition model comprises a face feature extraction engine and a face library, the face feature extraction engine is used for extracting face features, the face features corresponding to the entrance guard white list personnel are stored in the face library, the face features extracted by the face feature extraction engine are matched with the face features in the face library, and if the matching is successful, the visiting user corresponding to the face features is the entrance guard white list personnel, and the visiting user can be released.

The pre-recognition result comprises a successful matching or a failed matching, the face feature extraction engine in the second face recognition model extracts the face features of the first face to be recognized and then matches the face features with the corresponding face library, if the matching is successful, the pre-recognition result is a successful matching, if the matching is failed, the pre-recognition result is a failed matching, and the pre-recognition result and the face features corresponding to the first face to be recognized are stored no matter the matching is successful or failed.

Optionally, if the size of the first face to be recognized is greater than a preset first size threshold, the first face to be recognized is recognized through the first face recognition model, and a first recognition result is obtained. If the size of the first face to be recognized is larger than the preset first size threshold, it indicates that the visiting user corresponding to the first face to be recognized is close to the access control machine and may not pass through the pre-recognition. The first face recognition model comprises a face feature extraction engine and a face library, and as the size of a face near the first face recognition model is large, excessive feature extraction steps are not needed in the recognition process, so that the accuracy of feature values extracted by the face feature extraction engine of the first face recognition model can be smaller than that of feature values extracted by the face feature extraction engine of the second face recognition model. Therefore, the algorithm complexity of the first face recognition model can be reduced, and the face recognition speed can be increased.

103. And based on the pre-recognition result, recognizing a second face to be recognized after a preset time interval to obtain a second recognition result, wherein the size of the second face to be recognized is larger than that of the first face to be recognized.

In this step, the above-mentioned pre-recognition result is a recognition result obtained by recognizing the first face to be recognized through the second face recognition model in step 102. The pre-recognition result is a pre-recognition result of a visiting user at a distance from the access control machine, and the pre-recognition result comprises successful matching or failed matching and face features corresponding to the first face to be recognized.

The second face to be recognized may be a face detected by the second face detection model, and specifically, when the size of the first face to be recognized is determined to be smaller than the preset first size threshold, the second face detection model is notified to perform face detection on a visiting user in the forbidden area after a preset time interval, the second face to be recognized is obtained through detection, and whether the size of the second face to be recognized is larger than the preset second size threshold is determined. If the size of the second face to be recognized is larger than a preset second size threshold, the fact that the visiting user is close to the access control machine is indicated, and if the size of the second face to be recognized is smaller than the preset second size threshold, the fact that the visiting user is far away from the access control machine is indicated. After the preset time interval, if a visiting user in a distance has the intention of entering the entrance guard, the visiting user can approach the entrance guard, and at the moment, the size of the detected face can be increased. It should be noted that the first size threshold and the second size threshold may be the same threshold or different thresholds, and the specific threshold may be selected according to a specific scenario.

The second face recognition model is used for recognizing on the basis of pre-recognition, and only visiting users who have access requirements from far to near are recognized, so when the second face detection model detects that the size of the second face to be recognized is smaller than a preset second size threshold value, the fact that the visiting users are far away from the access control machine and do not have access intentions or do not have corresponding pre-recognition results is shown, therefore, the second face to be recognized can be directly discarded, the second face to be recognized is not subjected to recognition processing, and therefore the fact that the visiting users who do not have access intentions are mistakenly recognized can be avoided. When the second face detection model detects that the size of the second face to be recognized is larger than a preset second size threshold, it is indicated that the visiting user has an entrance guard intention near the entrance guard. At this time, the face features of the second face to be recognized are extracted through the face feature extraction engine of the second face recognition model, the face features of the second face to be recognized extracted by the second face recognition model are compared with the face features of the first face to be recognized, whether the second face to be recognized is similar to the first face to be recognized is judged, if yes, the first face to be recognized and the second face to be recognized correspond to the same visiting user, and at this time, the visiting user is subjected to white list judgment through a pre-recognition result obtained by the first face to be recognized. Specifically, when the pre-recognition result is that the matching is successful, the matching is successful as a second recognition result, and in the second recognition result, the visiting user is a white list person that can be released. When the pre-recognition result is a matching failure, the matching failure can be used as a second recognition result, and in the second recognition result, the visiting user is a non-white list person. Certainly, when the pre-recognition result is that matching fails, matching of the face library can be performed again, the extracted face features of the second face to be recognized are compared with the face library, whether matching is successful or not is judged, if matching is successful, matching success is taken as a second recognition result, and if matching fails, matching failure is taken as a second recognition result.

104. And performing access control according to the second identification result.

And the second identification result comprises successful matching or failed matching, when the second identification result is successful matching, the access control machine is controlled to release, and when the second identification result is failed matching, the closing state of the access control machine is kept.

In the embodiment of the invention, the face detection is carried out on the visiting user in the entrance guard area through the first face detection model, and whether the size of the first face to be recognized is larger than a preset first size threshold value is judged; if the size of the first face to be recognized is smaller than the preset first size threshold, recognizing the first face to be recognized through a second face recognition model to obtain a pre-recognition result; based on the pre-recognition result, recognizing a second face to be recognized after a preset time interval to obtain a second recognition result, wherein the size of the second face to be recognized is larger than that of a first face to be recognized, and the first face to be recognized and the second face to be recognized are faces of the same visiting user; and performing access control according to the second identification result. Through carrying out the pre-recognition to the people face that face size is less than first size threshold value, after the interval default time, carry out the second time discernment based on the result of pre-discernment again, make to discern the second time and can go on the basis of pre-discernment, the speed of second time discernment has been improved, be equivalent to once discerning in entrance guard's machine distantly to the visitor, discern again near the entrance guard's machine, because when discerning near the entrance guard's machine again, be based on the result of distantly discerning goes on, can improve the speed of second time discernment, thereby make the visitor to arrive when the entrance guard's machine is near, entrance guard's machine can quick response, and user experience is improved.

It should be noted that the access control method provided by the embodiment of the invention can be applied to access control devices, such as an access controller, a monitor, a computer, a server and the like.

Referring to fig. 2, fig. 2 is another access control method according to an embodiment of the present invention, which is different from the embodiment of fig. 1, in which the first face recognition model includes: the system comprises a first face feature extraction engine, a second face feature extraction engine, a first face library and a second face library, wherein the precision of first face features extracted by the first face feature extraction engine is smaller than that of second face features extracted by the second face feature extraction engine, the precision of face feature data stored in the first face library is smaller than that of face feature data stored in the second face library, the first face features are used for matching in the first face library, and the second face features are used for matching in the second face library. The second face recognition model comprises: the system comprises a third face feature extraction engine, a fourth face feature extraction engine, a third face library and a fourth face library, wherein the precision of third face features extracted by the third face feature extraction engine is higher than that of fourth face features extracted by the fourth face feature extraction engine, the precision of face feature data stored in the third face library is higher than that of face feature data stored in the fourth face library, the third face features are used for matching in the third face library, and the fourth face features are used for matching in the fourth face library.

It should be noted that, in a possible embodiment, the first face feature extraction engine and the fourth face feature extraction engine may be the same or use the same algorithm, the second face feature extraction engine and the third face feature extraction engine may be the same or use the same algorithm, the first face library and the fourth face library may be the same face library, and the second face library and the third face library may be the same face library.

As shown in fig. 2, the method comprises the steps of:

201. and carrying out face detection on the visiting user in the access control area through the first face detection model, and extracting to obtain a first face to be recognized.

202. And judging whether the size of the first face to be recognized is larger than a preset first size threshold value.

If the size of the first face to be recognized is greater than the preset first size threshold, go to step 203, and if the size of the first face to be recognized is less than the preset first size threshold, go to step 211.

203. And inputting a first face to be recognized into the first face feature extraction engine to perform first feature extraction, and extracting to obtain first face features to be recognized.

204. And matching the first face features to be recognized in a first face library.

205. And judging whether the first face features to be recognized are successfully matched in the first face library.

If the matching is successful, go to step 206, and if the matching is failed, go to step 207.

206. And outputting the successful matching as a first recognition result.

207. And inputting the first face to be recognized into a second face feature extraction engine to perform second feature extraction, and extracting to obtain second face features to be recognized.

It should be noted that the second to-be-recognized face feature is a face feature extracted from the first to-be-recognized face, and is not a face feature extracted from the second to-be-recognized face.

208. And matching the second face features to be recognized in a second face library.

209. And judging whether the second face features to be recognized are successfully matched in the second face library.

If the matching is successful, go to step 206, and if the matching is failed, go to step 210.

Through

steps

207, 208 and 209, the accuracy of face recognition can be increased through secondary recognition. The

steps

207, 208 and 209 are optional, and in a possible embodiment, after the step 205 determines that the matching fails, the process may directly proceed to the step 210.

210. And outputting the matching failure as a first recognition result.

211. And respectively inputting the first face to be recognized into a third face feature extraction engine and a fourth face feature extraction engine for feature extraction, and respectively extracting to obtain a third face feature to be recognized and a fourth face feature to be recognized.

212. And storing the third face feature to be recognized and the fourth face feature to be recognized as a pre-recognition result.

213. And after the preset time interval, carrying out face detection on the visiting user in the entrance guard area through the second face detection model, and extracting to obtain a second face to be recognized.

214. And judging whether the size of the second face to be recognized is larger than a preset second size threshold value.

If the size of the second face to be recognized is smaller than the preset second size threshold, the step 215 is performed, and if the size of the second face to be recognized is larger than the preset second size threshold, the step 216 is performed.

215. And discarding the second face to be recognized or not processing the second face to be recognized.

216. And performing feature extraction on the second face to be recognized through a fourth face feature extraction engine to obtain fifth face features to be recognized through extraction.

217. And carrying out similarity comparison on the fifth to-be-recognized face feature and the fourth to-be-recognized face feature.

218. And judging whether the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature.

If the fifth to-be-recognized face feature is not similar to the fourth to-be-recognized face feature, it is determined that the second to-be-recognized face and the first to-be-recognized face may not be the same visiting user, the step 219 is performed, and if the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature, it is determined that the second to-be-recognized face and the first to-be-recognized face are the same visiting user, the step 222 is performed.

219. And matching the fifth human face feature to be recognized in the fourth human face library.

220. And judging whether the fifth face feature to be recognized is successfully matched in the fourth face library.

If the matching is successful, the process proceeds to step 221, and if the matching is failed, the process proceeds to step 220.

221. And outputting the successful matching as a second recognition result to control the entrance guard to pass.

222. And inputting the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extracting to obtain sixth face features to be recognized.

223. And matching the sixth face feature to be recognized in a third face library.

224. And judging whether the sixth face feature to be recognized is successfully matched in the third face library.

If the matching is successful, the procedure goes to step 221, and if the matching is failed, the procedure goes to step 225.

Through

steps

222, 223, 224, the accuracy of face recognition can be increased through secondary recognition. The

steps

220, 221, and 222 are optional, and in a possible embodiment, after the step 220 determines that the matching fails, the process may directly proceed to the step 225.

225. And outputting the matching failure as a second recognition result, and keeping the entrance guard closed.

In the embodiment of the invention, the face with the face size smaller than the first size threshold is pre-identified, and after the preset time interval, the second identification is carried out based on the pre-identified result, so that the second identification can be carried out on the basis of the pre-identification, the second identification speed is improved, which is equivalent to that a visiting user is identified once at a far position of the access control machine and is identified once at a near position of the access control machine. In addition, through multiple recognition, the recognition accuracy is also increased.

Referring to fig. 3, fig. 3 is a schematic diagram illustrating another access control method according to an embodiment of the present invention, which is different from the embodiment of fig. 2 in that the pre-recognition result further includes a third recognition result obtained by matching a third to-be-recognized face feature in the third face library.

301. And carrying out face detection on the visiting user in the access control area through the first face detection model, and extracting to obtain a first face to be recognized.

302. And judging whether the size of the first face to be recognized is larger than a preset first size threshold value.

If the size of the first face to be recognized is larger than the preset first size threshold, the step 303 is performed, and if the size of the first face to be recognized is smaller than the preset first size threshold, the step 309 is performed.

303. And inputting a first face to be recognized into the first face feature extraction engine to perform first feature extraction, and extracting to obtain first face features to be recognized.

304. And matching the first face features to be recognized in a first face library.

305. And judging whether the first face features to be recognized are successfully matched in the first face library.

If the matching is successful, go to step 306, and if the matching is failed, go to step 311.

306. And outputting the successful matching as a first recognition result.

307. And inputting the first face to be recognized into a second face feature extraction engine for second feature extraction, and extracting to obtain second face features to be recognized.

308. And matching the second face features to be recognized in a second face library.

309. And judging whether the second face features to be recognized are successfully matched in the second face library.

If the matching is successful, go to step 306, and if the matching is failed, go to step 310.

Through

steps

307, 308, 309, the accuracy of face recognition can be increased through secondary recognition. The

steps

307, 308, 309 are optional, and in a possible embodiment, after the step 305 determines that the matching fails, the process may directly proceed to the step 310.

310. And outputting the matching failure as a first recognition result.

311. And respectively inputting the first face to be recognized into a third face feature extraction engine and a fourth face feature extraction engine for feature extraction, and respectively extracting to obtain a third face feature to be recognized and a fourth face feature to be recognized.

312. And matching the third face features to be recognized in a third face library to obtain a third recognition result.

313. And storing the third recognition result and the fourth face feature to be recognized as a pre-recognition result.

314. And after the preset time interval, carrying out face detection on the visiting user in the access control area through the second face detection model, and extracting to obtain a second face to be recognized.

315. And judging whether the size of the second face to be recognized is larger than a preset second size threshold value.

If the size of the second face to be recognized is smaller than the preset second size threshold, step 315 is performed, and if the size of the second face to be recognized is larger than the preset second size threshold, step 316 is performed.

316. And discarding the second face to be recognized or not processing the second face to be recognized.

317. And performing feature extraction on the second face to be recognized through a fourth face feature extraction engine to extract a fifth face feature to be recognized.

318. And carrying out similarity comparison on the fifth to-be-recognized face feature and the fourth to-be-recognized face feature.

319. And judging whether the fifth face feature to be recognized is similar to the fourth face feature to be recognized.

If the fifth to-be-recognized face feature is not similar to the fourth to-be-recognized face feature, it is determined that the second to-be-recognized face and the first to-be-recognized face may not be the same visiting user, the step 320 is performed, and if the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature, it is determined that the second to-be-recognized face and the first to-be-recognized face are the same visiting user, the step 327 is performed.

320. And matching the fifth human face feature to be recognized in the fourth human face library.

321. And judging whether the fifth face feature to be recognized is successfully matched in the fourth face library.

If the matching is successful, the process proceeds to step 322, and if the matching is unsuccessful, the process proceeds to step 223.

322. And outputting the successful matching as a second recognition result to control the entrance guard to pass.

323. And inputting the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extracting to obtain sixth face features to be recognized.

324. And matching the sixth face feature to be recognized in a third face library.

325. And judging whether the sixth face feature to be recognized is successfully matched in the third face library.

If the matching is successful, go to step 322, and if the matching is failed, go to step 326.

326. And outputting the matching failure as a second recognition result, and keeping the entrance guard closed.

327. And judging whether the third identification result is matched successfully.

If the third recognition result is a successful match, the process proceeds to step 322, and if the third recognition result is a failed match, the process proceeds to step 323.

In the embodiment of the invention, the face with the face size smaller than the first size threshold is pre-identified, and after the interval of the preset time, the second identification is carried out based on the pre-identified result, so that the second identification can be carried out on the basis of the pre-identification, the speed of the second identification is improved, namely, the visiting user is identified once at the far position of the access control machine and is identified once again at the near position of the access control machine. In addition, through many identifications, the accuracy of identification is also increased.

It should be noted that the access control method provided by the embodiment of the present invention may be applied to an access controller, a monitor, a computer, a server, and other devices for controlling access.

Referring to fig. 4, fig. 4 is a schematic structural diagram of an access control apparatus according to an embodiment of the present invention, and as shown in fig. 4, the apparatus includes:

the first detection module 401 is configured to perform face detection on a visiting user in an access control area through a first face detection model, and determine whether a size of a first face to be recognized is greater than a preset first size threshold;

a first recognition module 402, configured to, if the size of the first face to be recognized is smaller than the preset first size threshold, recognize the first face to be recognized through a second face recognition model, and obtain a pre-recognition result;

a second recognition module 403, configured to, based on the pre-recognition result, recognize a second face to be recognized at preset time intervals, so as to obtain a second recognition result, where a size of the second face to be recognized is larger than a size of a first face to be recognized, and the first face to be recognized and the second face to be recognized are faces of the same visiting user;

and the control module 404 is configured to perform access control according to the second identification result.

Optionally, as shown in fig. 5, the apparatus further includes:

a third recognition module 405, configured to, if the size of the first face to be recognized is greater than the preset first size threshold, recognize the first face to be recognized through the first face recognition model to obtain a first recognition result;

the control module 404 is further configured to perform access control according to the first recognition result.

Optionally, the first face recognition model includes: first face characteristic extraction engine, second face characteristic extraction engine, first face storehouse and second face storehouse, wherein, the precision of first face characteristic that first face characteristic extraction engine extracted is less than the precision of second face characteristic that second face characteristic extraction engine extracted, the face characteristic data precision of storage in the first face storehouse is less than the face characteristic data precision of storage in the second face storehouse, first face characteristic is used for match in the first face storehouse, second face characteristic is used for match in the second face storehouse, as shown in fig. 6, third identification module 405 includes:

a first extraction unit 4051, configured to, if the size of the first face to be recognized is greater than the preset first size threshold, input the first face to be recognized into the first face feature extraction engine to perform first feature extraction, so as to extract a first face feature to be recognized;

a first matching unit 4052, configured to match the first to-be-recognized face feature in the first face library, and determine whether the matching is successful;

a first output unit 4053, configured to output, if matching is successful, that matching is successful as a first recognition result;

a second extraction unit 4054, configured to, if matching fails, input the first face to be recognized into the second face feature extraction engine to perform second feature extraction, so as to extract a second face feature to be recognized;

a second matching unit 4055, configured to match the second to-be-recognized face feature in the second face library, and determine whether the matching is successful;

the first output unit 4053 is further configured to, if the matching is successful, output the matching as a first recognition result;

the first output unit 4053 is further configured to output, if the matching fails, the matching failure as the first recognition result.

Optionally, as shown in fig. 7, the apparatus further includes:

a second detection module 406, configured to, if the size of the first face to be recognized is smaller than the preset first size threshold, perform face detection on a visiting user in the access control area through the second face detection model after the preset time interval, and determine whether the size of the second face to be recognized is larger than a preset second size threshold;

the first recognition module 402 is further configured to, if the size of the second face to be recognized is greater than the preset second size threshold, recognize the second face to be recognized through a second face recognition model.

Optionally, the second face recognition model includes: third face feature extraction engine, fourth face feature extraction engine, third face storehouse and fourth face storehouse, wherein, the precision of third face feature that third face feature extraction engine extracted is greater than the precision of fourth face feature that fourth face feature extraction engine extracted, the face feature data precision of storage in the third face storehouse is greater than the face feature data precision of storage in the fourth face storehouse, third face feature is used for matching in the third face storehouse, fourth face feature is used for matching in the fourth face storehouse, as shown in fig. 8, first identification module 402 includes:

the third extraction unit 4021 is configured to input the first face to be recognized into a third face feature extraction engine and a fourth face feature extraction engine, respectively, to perform feature extraction, and extract a third face feature to be recognized and a fourth face feature to be recognized, respectively;

the storage unit 4022 is configured to store the third to-be-recognized face feature and the fourth to-be-recognized face feature as a pre-recognition result; or alternatively

And the face recognition module is used for matching the third face features to be recognized in the third face library to obtain a third recognition result, and storing the third recognition result and the fourth face features to be recognized as a pre-recognition result.

Optionally, as shown in fig. 9, the second identifying module 403 includes:

a fourth extraction unit 4031, configured to, if the size of the second face to be recognized is greater than the preset second size threshold, perform feature extraction on the second face to be recognized by using the fourth face feature extraction engine, and extract a fifth face feature to be recognized;

a comparison unit 4032, configured to compare the similarity between the fifth to-be-recognized face feature and the fourth to-be-recognized face feature, and determine whether the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature;

a third matching unit 4033, configured to match the fifth to-be-recognized face feature in the fourth face library if the fifth to-be-recognized face feature is not similar to the fourth to-be-recognized face feature, and determine whether the matching is successful;

a second output unit 4034, configured to output, if matching is successful, the matching success as a second recognition result;

a fifth extraction unit 4035, configured to, if matching fails, input the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extract a sixth face feature to be recognized;

a fourth matching unit 4036, configured to match the sixth to-be-recognized face feature in the third face library, and determine whether the matching is successful;

the second output unit 4034 is further configured to output, if the matching is successful, the matching is successful as a second recognition result;

the second output unit 4034 is further configured to output, if the matching fails, the matching failure as a second recognition result.

Optionally, as shown in fig. 10, the second identifying module 403 further includes:

a determining unit 4037, configured to determine that the third recognition result is a successful matching or a failed matching if the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature;

a third output unit 4038, configured to, if the third recognition result is a successful matching, output a successful matching as a second recognition result; or alternatively

A fifth matching unit 4039, configured to match the third to-be-recognized face feature in the third face library if the fifth to-be-recognized face feature is similar to the fourth to-be-recognized face feature, and determine whether the matching is successful;

the third output unit 4038 is further configured to output, if the matching is successful, the matching is successful as a second recognition result;

a sixth extraction unit 40310, configured to, if matching fails or the third recognition result is matching failure, input the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extract a sixth face feature to be recognized;

a sixth matching unit 40311, configured to match the sixth to-be-recognized face feature in the third face library, and determine whether the matching is successful;

the third output unit 4038 is further configured to output, if the matching fails, the matching failure as a second recognition result.

It should be noted that the access control device provided by the embodiment of the invention can be applied to access control devices, such as an access controller, a monitor, a computer, a server and the like.

The access control device provided by the embodiment of the invention can realize each process realized by the access control method in the method embodiment, and can achieve the same beneficial effects. To avoid repetition, further description is omitted here.

Referring to fig. 11, fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 11, including: a memory 1102, a processor 1101, and a computer program stored on the memory 1102 and executable on the processor 1101, wherein:

the processor 1101 is configured to call the computer program stored in the memory 1102, and perform the following steps:

detecting the face of a visiting user in the access control area through a first face detection model, and judging whether the size of a first face to be recognized is larger than a preset first size threshold value or not;

and performing access control according to the second recognition result.

Optionally, the processor 1101 further performs the steps including:

and performing access control according to the first identification result.

Optionally, the first face recognition model includes: first face feature extraction engine, second face feature extraction engine, first face storehouse and second face storehouse, wherein, the precision of the first face feature that first face feature extraction engine extracted is less than the precision of the second face feature that second face feature extraction engine extracted, the face feature data precision that stores in the first face storehouse is less than the face feature data precision of storing in the second face storehouse, first face feature be used for matching in the first face storehouse, second face feature be used for matching in the second face storehouse, the treater 1101 is carried out if the size of first face of awaiting discerning is greater than preset first size threshold value, then to discerning first face of awaiting discerning through first face identification model obtains first recognition result, includes:

if the size of the first face to be recognized is larger than the preset first size threshold, inputting the first face to be recognized into the first face feature extraction engine to perform first feature extraction, and extracting to obtain a first face feature to be recognized;

Optionally, the recognizing, performed by the processor 1101, the second face to be recognized after the interval of the preset time includes:

Optionally, the second face recognition model includes: third face feature extraction engine, fourth face feature extraction engine, third face storehouse and fourth face storehouse, wherein, the precision of third face feature that third face feature extraction engine extracted is greater than the precision of the fourth face feature that fourth face feature extraction engine extracted, the facial feature data precision of storage in the third face storehouse is greater than the facial feature data precision of storage in the fourth face storehouse, third face feature is used for matching in the third face storehouse, fourth face feature is used for matching in the fourth face storehouse, and what treater 1101 was carried out it is right to discern the first face of waiting to discern through second face identification model, include:

Optionally, the recognizing, performed by the processor 1101, the second face to be recognized after a preset time interval based on the pre-recognition result to obtain a second recognition result, where the recognizing includes:

matching the sixth face feature to be recognized in the third face library, and judging whether the matching is successful;

if the matching is successful, outputting the successful matching as a second identification result;

Optionally, the recognizing, performed by the processor 1101, the second face to be recognized after a preset time interval based on the pre-recognition result to obtain a second recognition result, further includes:

if the third recognition result is successful in matching, outputting the successful matching as a second recognition result; or

if the matching fails or the third recognition result is the matching failure, inputting the second face to be recognized into a third face feature extraction engine to perform feature extraction on the second face to be recognized, and extracting to obtain sixth face features to be recognized;

It should be noted that the electronic device may be a door access device, a monitor, a computer, a server, and the like, which can be applied to door access control.

The electronic equipment provided by the embodiment of the invention can realize each process realized by the access control method in the method embodiment, can achieve the same beneficial effects, and is not repeated here for avoiding repetition.

The embodiment of the invention also provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when being executed by a processor, the computer program realizes each process of the access control method provided by the embodiment of the invention, can achieve the same technical effect, and is not repeated here to avoid repetition.

An embodiment of the present invention further provides an access control system, including:

the access control machine is used for releasing or blocking the visiting user; and

the image acquisition device is used for acquiring image information of the access control area;

the access control equipment is used for executing the access control method according to the image information so as to control the access controller.

It should be noted that the access control device may be a device such as an access controller, a monitor, a computer, and a server that can be applied to access control.

The access control system provided by the embodiment of the invention can realize each process realized by the access control method in the embodiment of the method, can achieve the same beneficial effects, and is not repeated herein for avoiding repetition.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by a computer program, which may be stored in a computer readable storage medium and executed by a computer to implement the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.

The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.

Claims

1. An access control method, comprising the steps of:

if the size of the first face to be recognized is larger than the preset first size threshold, recognizing the first face to be recognized through a first face recognition model to obtain a first recognition result, wherein the first face recognition model comprises: the system comprises a first face feature extraction engine, a second face feature extraction engine, a first face library and a second face library, wherein the precision of first face features extracted by the first face feature extraction engine is smaller than that of second face features extracted by the second face feature extraction engine, the precision of face feature data stored in the first face library is smaller than that of face feature data stored in the second face library, the first face features are used for matching in the first face library, and the second face features are used for matching in the second face library; if the size of the first face to be recognized is larger than the preset first size threshold, inputting the first face to be recognized into the first face feature extraction engine to perform first feature extraction, and extracting to obtain a first face feature to be recognized; matching the first face features to be recognized in the first face library, and judging whether the matching is successful; if the matching is successful, outputting the successful matching as a first recognition result; if the matching fails, inputting the first face to be recognized into the second face feature extraction engine for second feature extraction, and extracting to obtain second face features to be recognized; matching the second face features to be recognized in the second face library, and judging whether the matching is successful; if the matching is successful, outputting the successful matching as a first recognition result; if the matching fails, outputting the matching failure as a first recognition result;

performing access control according to the first identification result;

and performing access control according to the second identification result.

2. The method of claim 1, wherein the recognizing the second face to be recognized after the interval of the preset time comprises:

if the size of the first face to be recognized is smaller than the preset first size threshold, after the preset time interval, performing face detection on a visiting user in the entrance guard area through a second face detection model, and judging whether the size of a second face to be recognized is larger than a preset second size threshold or not;

3. The method of claim 2, wherein the second face recognition model comprises: third face feature extraction engine, fourth face feature extraction engine, third face storehouse and fourth face storehouse, wherein, the precision of third face feature that third face feature extraction engine extracted is greater than the precision of fourth face feature that fourth face feature extraction engine extracted, the face feature data precision of storage in the third face storehouse is greater than the face feature data precision of storage in the fourth face storehouse, third face feature is used for matching in the third face storehouse, fourth face feature is used for matching in the fourth face storehouse match, it is right through second face identification model the first face that waits to discern is discerned, include:

storing the third face feature to be recognized and the fourth face feature to be recognized as a pre-recognition result; or alternatively

4. The method as claimed in claim 3, wherein the recognizing the second face to be recognized after a preset time interval based on the pre-recognition result to obtain a second recognition result comprises:

5. The method as claimed in claim 4, wherein the recognizing a second face to be recognized after a preset time interval based on the pre-recognition result to obtain a second recognition result further comprises:

when the third face features to be recognized are matched in the third face library to obtain a third recognition result, and the third recognition result and the fourth face features to be recognized are stored as pre-recognition results, if the fifth face features to be recognized are similar to the fourth face features to be recognized, the third recognition result is judged to be successful in matching or failed in matching;

When the third to-be-recognized face features and the fourth to-be-recognized face features are stored as pre-recognition results, if the fifth to-be-recognized face features are similar to the fourth to-be-recognized face features, matching the third to-be-recognized face features in the third face library, and judging whether matching is successful or not;

6. An access control apparatus, comprising:

a third recognition module, configured to, if the size of the first face to be recognized is greater than the preset first size threshold, recognize the first face to be recognized through a first face recognition model to obtain a first recognition result, where the first face recognition model includes: the system comprises a first face feature extraction engine, a second face feature extraction engine, a first face library and a second face library, wherein the precision of a first face feature extracted by the first face feature extraction engine is smaller than that of a second face feature extracted by the second face feature extraction engine, the precision of face feature data stored in the first face library is smaller than that of face feature data stored in the second face library, the first face feature is used for matching in the first face library, and the second face feature is used for matching in the second face library; if the size of the first face to be recognized is larger than the preset first size threshold, inputting the first face to be recognized into the first face feature extraction engine for first feature extraction, and extracting to obtain a first face feature to be recognized; matching the first face features to be recognized in the first face library, and judging whether the matching is successful; if the matching is successful, outputting the successful matching as a first recognition result; if the matching fails, inputting the first face to be recognized into the second face feature extraction engine for second feature extraction, and extracting to obtain second face features to be recognized; matching the second face features to be recognized in the second face library, and judging whether the matching is successful; if the matching is successful, outputting the successful matching as a first recognition result; if the matching fails, outputting the matching failure as a first recognition result;

and the control module is used for performing access control according to the second recognition result and is also used for performing access control according to the first recognition result.

7. An access control apparatus, comprising: a memory, a processor and a computer program stored on the memory and operable on the processor, the processor implementing the steps of the access control method according to any one of claims 1 to 5 when executing the computer program.

8. An access control system, comprising:

the entrance guard control equipment is used for executing the entrance guard control method according to any one of claims 1 to 5 according to the image information so as to control the entrance guard.