Target recognition method, target recognition system, device, equipment and storage medium

Abstract

The present application relates to a target recognition method, system, device, computer equipment and storage medium. The method includes: obtaining a depth image obtained by a preset binocular camera cruising and photographing a target scene; wherein, the target scene includes at least one target detection object; based on the position of the target detection object corresponding to the depth image, the target scene is divided into a preset number of recognition areas, and the preset zoom camera is configured with shooting parameters for each recognition area; the zoom camera is controlled to take close-up shots of each recognition area under the shooting parameters to obtain a corresponding close-up image; target recognition processing is performed on the target detection object in the close-up image to obtain recognition results for each target detection object in the target scene. The use of this method can effectively improve the efficiency and accuracy of identifying the detection objects in the target scene, which is conducive to reference and optimization of the subsequent activity behaviors of the target objects.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a target recognition method, a target recognition system, a target recognition device, a computer device, a storage medium and a computer program product. Background Art

[0002] Currently, in many scenarios related to on-site activities (for example, classroom scenarios where students come to the classroom to study, work scenarios where corporate employees come to the office to work, etc.), it is necessary to identify and record the target user objects in order to obtain the corresponding attendance or sign-in record results, so as to refer to and optimize the subsequent user object activity behaviors based on the record results.

[0003] In the current method of identifying the target object, the activity information of the target object is manually recorded on site or the audio and video of the active object is collected through smart terminals such as cameras and microphones, and then the corresponding attendance results or sign-in results are manually analyzed. However, the above method mainly relies on manual and detailed analysis operations, which leads to low efficiency and accuracy of object recognition, and is not conducive to reference and optimization of subsequent activity behaviors. Summary of the invention

[0004] Based on this, it is necessary to provide a target recognition method, target recognition system, target recognition device, computer equipment, storage medium and computer program product that can improve the efficiency and accuracy of object recognition in response to the above technical problems.

[0005] In a first aspect, the present application provides a target recognition method. The method comprises:

[0006] Acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object;

[0007] Based on the position of the target detection object corresponding to the depth image, the target scene is divided into a preset number of recognition areas, and a preset zoom camera is configured with shooting parameters for each of the recognition areas;

[0008] Controlling the zoom camera to take close-up photos of each of the identification areas under the shooting parameters to obtain corresponding close-up images;

[0009] Target recognition processing is performed on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0010] In one of the embodiments, the target detection object is a face object in the target scene;

[0011] The step of dividing the target scene into a preset number of recognition areas includes:

[0012] Performing coordinate detection on each face object in the depth image to determine the object coordinates of each face object;

[0013] Based on a preset number of clusters, coordinate clustering is performed on the object coordinates of each of the face objects to obtain a set of cluster coordinates corresponding to the number of clusters;

[0014] Based on the central coordinates in each of the cluster coordinate sets, the target scene is divided into recognition areas corresponding to the number of clusters.

[0015] In one embodiment, the configuration of the preset zoom camera shooting parameters for each of the identified areas includes:

[0016] Based on the preset pixel ratio, the preset camera coordinates and the center coordinates in each cluster coordinate set, configuring the shooting parameters of the zoom camera for each of the identification areas; the preset camera coordinates are the position coordinates of the initial shooting center of the zoom camera;

[0017] Wherein, the shooting parameters include the zoom factor and offset data of the zoom camera;

[0018] The zoom factor is used to limit the pixel ratio of the face object in the image captured by the zoom camera to meet the preset pixel ratio;

[0019] The offset data includes an offset angle and an offset distance of the zoom camera when the shooting center of the zoom camera is offset from the preset camera coordinates to the center coordinates of the cluster coordinate set.

[0020] In one embodiment, the performing target recognition processing on the target detection object in the close-up image includes:

[0021] Performing face detection, coordinate detection, quality detection and face matching on the face object in the close-up image in sequence;

[0022] Wherein, the face detection is used to detect at least one face object present in the close-up image;

[0023] The coordinate detection is used to detect the object coordinates of the face object;

[0024] The quality detection is used to evaluate the facial features of the facial objects, so as to eliminate the facial objects that do not meet the preset evaluation criteria and obtain the target facial objects among the facial objects;

[0025] The face matching is used to perform facial feature matching between the target face object and a reference face object in a preset face library to obtain a matching result for the target face object.

[0026] In one embodiment, the performing facial feature evaluation on the face objects to eliminate face objects that do not meet a preset evaluation standard and obtain a target face object from the face objects includes:

[0027] Performing feature recognition processing on the face object to determine the face feature clarity and face feature integrity of the face object;

[0028] Obtaining an evaluation score for the face object based on the face feature clarity and the face feature completeness;

[0029] Among the face objects, the face objects corresponding to the evaluation scores lower than the preset scores are eliminated to obtain the target face objects.

[0030] In one of the embodiments, the target detection object is a face object to be identified, and the recognition result represents the face information of the face object to be identified;

[0031] After obtaining the recognition result for each target detection object in the target scene, the method further includes:

[0032] Extracting a face information list matching the target scene at the current time stage from a preset database; wherein the target scene is bound to different face information lists at different time stages, and face information of multiple preset face objects is stored in the face information list;

[0033] The facial information of each facial object to be identified is compared with the facial information of the preset facial objects in the facial information list to obtain the corresponding target facial information in the facial information list that fails the comparison, and the target facial information is marked.

[0034] In one embodiment, the step of acquiring a depth image obtained by cruising and photographing a target scene with a preset binocular camera includes:

[0035] Acquire two monocular depth images taken by the binocular camera during cruising of the target scene;

[0036] The two monocular depth images are stitched together to obtain a panoramic stitching depth image for the target scene.

[0037] In a second aspect, the present application also provides an information processing system. The system comprises:

[0038] A binocular camera is used to cruise and shoot a target scene to obtain a corresponding depth image; wherein the target scene includes at least one target detection object;

[0039] A zoom camera, used to take close-up photos of each recognition area in the target scene based on corresponding shooting parameters to obtain corresponding close-up images;

[0040] A processor is used to divide the target scene into a preset number of recognition areas based on the positions of the target detection objects corresponding to the positions in the depth image; configure the shooting parameters of the zoom camera for each of the recognition areas; and perform target recognition processing on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0041] In a third aspect, the present application also provides a target recognition device. The device comprises:

[0042] An acquisition module, used to acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object;

[0043] A processing module, configured to divide the target scene into a preset number of identification areas based on the position of the target detection object corresponding to the position in the depth image, and configure a preset zoom camera shooting parameter for each of the identification areas;

[0044] A close-up module, used for controlling the zoom camera to take close-up photos of each of the identification areas under the shooting parameters to obtain corresponding close-up images;

[0045] The recognition module is used to perform target recognition processing on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0046] In a fourth aspect, the present application further provides a computer device, including:

[0047] processor;

[0048] a memory for storing executable instructions for the processor;

[0049] The processor is configured to execute the executable instructions to implement the target recognition method as described above.

[0050] In a fifth aspect, the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium includes program data, and is characterized in that when the program data is executed, the target recognition method as described above is implemented.

[0051] In a sixth aspect, the present application further provides a computer program product, wherein the computer program product includes program instructions, and when the program instructions are executed, the target recognition method described above is implemented.

[0052] The target recognition method, target recognition system, target recognition device, computer equipment, storage medium and computer program product described above, on the one hand, in a manner different from the prior art, first cruises and shoots the target scene through a binocular camera to obtain a depth image, and then uses a configured zoom camera to take close-up shots of each recognition area in the target scene, so as to perform target recognition processing on the obtained close-up images, and obtain recognition results for the target detection objects, thereby optimizing the target recognition process, effectively improving the efficiency of identifying the detection objects in the target scene, and reducing the consumption of manpower and material resources; on the other hand, this scheme divides the target scene into a preset number of recognition areas according to the depth image obtained by the binocular camera, and configures the shooting parameters of the zoom camera for each recognition area according to the corresponding position of the target detection object in the depth image; and then uses the configured zoom camera to take close-up shots of each recognition area, so as to perform target recognition processing on the obtained close-up images, thereby obtaining recognition results for the target detection objects in the target scene, effectively improving the accuracy of identifying the target objects, and facilitating reference and optimization of subsequent target object activity behaviors. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0054] Figure 1 The figure is a diagram showing an application environment of a target recognition method according to an exemplary embodiment.

[0055] Figure 2 The figure is a flow chart of a target recognition method according to an exemplary embodiment.

[0056] Figure 3 The figure is a flowchart showing a step of dividing an identification area according to an exemplary embodiment.

[0057] Figure 4 The present invention is a flow chart of a method for checking student attendance according to an exemplary embodiment.

[0058] Figure 5 The present invention is a module diagram of a student attendance method according to an exemplary embodiment.

[0059] Figure 6 The figure is a block diagram of a target recognition system according to an exemplary embodiment.

[0060] Figure 7 The figure is a block diagram of a target recognition device according to an exemplary embodiment.

[0061] Figure 8 The invention is a block diagram of a computer device for target recognition according to an exemplary embodiment. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0063] The terms "first", "second", etc. in this application are used to distinguish different objects, rather than to describe a specific order. In addition, although the terms "first", "second", etc. are used many times to describe various operations (or various thresholds or various applications or various instructions or various elements), etc., these operations (or thresholds or applications or instructions or elements) should not be limited by these terms. These terms are only used to distinguish one operation (or threshold or application or instruction or element) from another operation (or threshold or application or instruction or element).

[0064] The target recognition method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the terminal 102 communicates with the server 104 through a communication network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers.

[0065] In some embodiments, reference Figure 1 First, the server 104 obtains a depth image obtained by cruising and photographing the target scene with a preset binocular camera; wherein the target scene includes at least one target detection object; then, the server 104 divides the target scene into a preset number of recognition areas based on the position of the target detection object corresponding to the depth image, and configures the preset zoom camera with shooting parameters for each recognition area; the zoom camera is controlled to take close-up shots of each recognition area under the shooting parameters to obtain corresponding close-up images; finally, the server 104 performs target recognition processing on the target detection object in the close-up image to obtain recognition results for each target detection object in the target scene.

[0066] In some embodiments, the terminal 102 (such as a mobile terminal, a fixed terminal) can be implemented in various forms. The terminal 102 can be a mobile terminal including a mobile phone, a smart phone, a laptop, a portable handheld device, a personal digital assistant (PDA), a tablet computer (PAD), etc. The terminal 102 can also be a fixed terminal such as an automated teller machine (ATM), an automatic all-in-one machine, a digital TV, a desktop computer, a fixed computer, etc.

[0067] In the following, it is assumed that the terminal 102 is a fixed terminal. However, it will be understood by those skilled in the art that the configuration according to the embodiments disclosed in the present application can also be applied to a mobile type terminal 102 if there are operations or elements specifically for mobile purposes.

[0068] In some embodiments, the data processing component running on the server 104 may load any of a variety of additional server applications and / or middle-tier applications being executed, such as HTTP (Hypertext Transfer Protocol), FTP (File Transfer Protocol), CGI (Common Gateway Interface), RDBMS (Relational Database Management System), etc.

[0069] In some embodiments, the server 104 may be implemented as an independent server or a server cluster consisting of multiple servers. The server 104 may be suitable for running one or more application services or software components that provide the terminal 102 described in the above disclosure.

[0070] In some embodiments, the application service may include a service interface that provides the target recognition function to the user, and a corresponding program service, etc. Among them, the software component may include an application program (SDK) or a client (APP) that executes the target recognition function.

[0071] In some embodiments, the application or client with target recognition function provided by server 104 includes a portal port that provides one-to-one application services to users in the foreground and multiple business systems that perform target recognition in the background, so as to extend the application of target recognition function to APP or client, so that users can use and access the target recognition function at any time and any place.

[0072] In some embodiments, the user can input corresponding code data or control parameters to the APP or client through a preset input device or automatic control program to execute the application service of the computer program in the server 104 and display the application service in the user interface.

[0073] In some embodiments, the operating system running on the APP or client may include various versions of Microsoft Windows®, Apple Macintosh® and / or Linux operating systems, various commercial or UNIX®-like operating systems (including but not limited to various GNU / Linux operating systems, Google Chrome® OS, etc.) and / or mobile operating systems, such as iOS®, Windows® Phone, Android® OS, BlackBerry® OS, Palm® OS operating systems, and other online operating systems or offline operating systems, without specific limitations here.

[0074] In some embodiments, Figure 2 As shown, a target recognition method is provided, which is applied to Figure 1 Taking the server 104 in the example as an example, the method includes the following steps:

[0075] Step S11: obtaining a depth image obtained by cruising and photographing a target scene with a preset binocular camera.

[0076] In one embodiment, a binocular camera is pre-set in a target scene, and a preset control program is used to control the binocular camera to cruise and shoot the target scene, so as to upload the captured depth image to a database. Then, the server extracts the depth image of the target scene in the target time interval from the database.

[0077] In some embodiments, the target scene is an application scene in the embodiments of the present application, the target time interval is a time range corresponding to the application scene, and the target scene includes at least one target detection object, wherein the target object is a face object in the application scene.

[0078] As an example, the application scenario may be a classroom scenario where students come to the classroom to study, and the corresponding target detection object is the student face in the classroom, and the corresponding target time interval is the time range of the students' corresponding class time; or, the application scenario may be a work scenario where corporate employees come to the office to work, and the corresponding target detection object is the employee face, and the corresponding target time interval is the time range of the corporate employees working in the office.

[0079] In some embodiments, the binocular camera is composed of two monocular wide-angle cameras calibrated with external parameters, and the depth image captured by the binocular camera includes the monocular depth images captured by each of the two monocular wide-angle cameras.

[0080] In one embodiment, the process of a server acquiring a depth image obtained by a preset binocular camera cruising and photographing a target scene includes the following steps: acquiring two monocular depth images taken by the binocular camera cruising and photographing the target scene; and stitching the two monocular depth images to obtain a panoramic stitched depth image of the target scene.

[0081] Specifically, the two monocular wide-angle cameras of the binocular camera are calibrated with external parameters in advance to obtain a binocular camera with external parameters. When acquiring a depth image, the server first extracts two monocular depth images taken by the binocular camera from the database, and then stitches the two monocular depth images to obtain a panoramic stitching depth image, and uses the panoramic stitching depth image as the depth image for the target scene.

[0082] Step S12: based on the position of the target detection object corresponding to the depth image, the target scene is divided into a preset number of recognition areas, and a preset zoom camera is configured with shooting parameters for each recognition area.

[0083] In one embodiment, the server first identifies the object position of the depth image to obtain the corresponding regional position of each target detection object in the target scene; then, according to the regional position of each target detection object, the target scene is divided into a preset number of recognition areas; finally, according to the corresponding regional position of each recognition area, the shooting parameters of the zoom camera are configured when shooting each recognition area.

[0084] The preset number may be determined according to the scene characteristics of the target scene and the computing requirements of the server, and may be any natural number, for example, dividing the target scene into 3, 5 or 9 recognition areas, and the like.

[0085] In some embodiments, the zoom camera is a telephoto gimbal camera, which is used to take close-up photos of each identification area according to configured shooting parameters.

[0086] In some embodiments, the shooting parameters may include a zoom factor and offset data of a zoom camera, wherein the zoom factor is the zoom size when the zoom camera takes a close-up shot of the recognition area, and the offset data is a camera offset parameter when the camera center of the zoom camera is offset to be consistent with the area center of the recognition area.

[0087] Step S13: controlling the zoom camera to take close-up photos of each recognition area under the shooting parameters to obtain corresponding close-up images.

[0088] Specifically, the server controls the zoom camera to take close-up photos of each recognition area according to the zoom factor and offset data configured for each recognition area, so as to obtain a close-up image corresponding to each recognition area.

[0089] Step S14: performing target recognition processing on the target detection objects in the close-up image to obtain recognition results for each target detection object in the target scene.

[0090] In one embodiment, the server performs target recognition processing on the target detection objects in the close-up image, including: performing face detection, coordinate detection, quality detection and face matching on the facial objects in the close-up image in sequence to obtain recognition results for each target detection object in the target scene.

[0091] Wherein, face detection is used to detect at least one face object in the close-up image.

[0092] Optionally, the server detects all facial objects present in the close-up image using a pre-trained face detection algorithm.

[0093] Among them, coordinate detection is used to detect the object coordinates of the face object.

[0094] Optionally, the server uses a pre-trained coordinate detection algorithm to detect the coordinate position of each facial object present in the close-up image.

[0095] Among them, quality detection is used to evaluate the facial features of face objects, so as to eliminate face objects that do not meet the preset evaluation standards and obtain the target face objects among the face objects.

[0096] Optionally, the server first uses a pre-trained quality evaluation algorithm to perform facial feature evaluation on the facial images corresponding to each facial object to obtain an evaluation result for each facial object; then, it determines whether the evaluation result of each facial object meets the preset evaluation criteria, so as to eliminate the facial objects that do not meet the evaluation criteria and obtain the remaining target facial objects among each facial object.

[0097] Among them, face matching is used to match the target face object with the reference face object in a preset face library to obtain a matching result for the target face object.

[0098] Optionally, the server uses a pre-trained face matching algorithm to first extract the facial features of each target face object; then, the facial features of each target face object are matched with the facial features of a preset reference face object to determine which reference face object each target face object corresponds to, thereby obtaining a matching result for the target face object.

[0099] The above-mentioned target recognition method first obtains a depth image obtained by cruising and photographing the target scene by a preset binocular camera; wherein, the target scene includes at least one target detection object; based on the position of the target detection object corresponding to the depth image, the target scene is divided into a preset number of recognition areas, and the shooting parameters of the preset zoom camera for each recognition area are configured; the zoom camera is controlled to take close-up photos of each recognition area under the shooting parameters to obtain a corresponding close-up image; the target detection object in the close-up image is subjected to target recognition processing to obtain a recognition result for each target detection object in the target scene. In this way, on the one hand, in a manner different from the prior art, a binocular camera is first used to cruise and shoot the target scene to obtain a depth image, and then a configured zoom camera is used to take close-up shots of each recognition area in the target scene, so as to perform target recognition processing on the obtained close-up images and obtain recognition results for the target detection objects, thereby optimizing the target recognition process, effectively improving the efficiency of identifying the detection objects in the target scene, and reducing the consumption of manpower and material resources; on the other hand, this scheme divides the target scene into a preset number of recognition areas according to the depth image obtained by the binocular camera, and configures the shooting parameters of the zoom camera for each recognition area according to the corresponding position of the target detection object in the depth image; and then the configured zoom camera is used to take close-up shots of each recognition area, so as to perform target recognition processing on the obtained close-up images, so as to obtain recognition results for the target detection objects in the target scene, effectively improving the accuracy of identifying the target objects, and facilitating reference and optimization of subsequent target object activity behaviors.

[0100] Those skilled in the art can understand that in the above method of the specific implementation, the disclosed method can be implemented in a more specific manner. For example, the above described implementation in which the server performs target recognition processing on the target detection object in the close-up image to obtain the recognition result for each target detection object in the target scene is only illustrative.

[0101] Exemplarily, a binocular camera cruises and shoots a target scene to obtain a depth image; or a zoom camera takes close-up photos of each recognition area under shooting parameters to obtain corresponding close-up images, etc. This is only a collection method, and there may be other division methods in actual implementation. For example, the depth image taken by the binocular camera and the close-up image taken by the zoom camera can be combined or collected into another system, or some features can be ignored or not executed.

[0102] In an exemplary embodiment, see Figure 3 , Figure 3This is a flow chart of dividing the recognition area in this application. That is, in step S12, the server divides the target scene into a preset number of recognition areas, which can be specifically implemented by executing the following methods:

[0103] Step S121: performing coordinate detection on each face object in the depth image to determine the object coordinates of each face object.

[0104] Specifically, the server first locates the face frame of the depth image based on the face frame detection algorithm to obtain the pixel coordinates of each face object in the camera coordinate system; then, based on the pixel coordinates and the calibration parameters of the binocular camera, it determines the depth coordinates of each face object in the camera coordinate system, and uses the depth coordinates as the object coordinates of the face object.

[0105] The pixel coordinates are represented based on (x, y, w, h). x and y refer to the coordinate position of the upper left corner of the face frame for locating the face object in the camera coordinate system; w and h refer to the pixel width and pixel height of the face frame, which are used to locate the pixel position of the face frame in the depth image.

[0106] The depth coordinate is represented by (x, y, d), where d is the depth value of the face frame.

[0107] In some embodiments, in a binocular camera, the positions and postures of two independent monocular cameras have been obtained through external parameter calibration (i.e., the calibration parameters of the binocular cameras are known). When the binocular camera captures an image, the two monocular cameras simultaneously capture two perspective images of the same scene. The server can determine the pixel coordinates of the same human face object in the two corresponding perspective images through a feature point matching algorithm. Then, the disparity of the two pixel coordinates of the same human face object (the horizontal offset of the two pixel points) is calculated to obtain the depth value of the human face object.

[0108] Step S122: based on a preset number of clusters, coordinate clustering is performed on the object coordinates of each face object to obtain a set of cluster coordinates corresponding to the number of clusters.

[0109] In some embodiments, the server clusters the object coordinates of each face object into preset X cluster coordinate sets based on the K-Means clustering algorithm.

[0110] Among them, the K-Means clustering algorithm is a hard clustering algorithm, which uses the distance from the object coordinates to the cluster center coordinates as the optimization objective function, and uses the method of obtaining the extreme value of the objective function to obtain the adjustment rule of iterative operation. Optionally, the K-Means clustering algorithm can use the Euclidean distance as the similarity measure to obtain the optimal classification corresponding to a certain cluster center coordinate, so that the similarity evaluation index is minimized. Among them, the K-Means clustering algorithm can use the error square sum criterion function as the objective function of coordinate center clustering.

[0111] Step S123: based on the central coordinates in each cluster coordinate set, the target scene is divided into recognition areas corresponding to the number of clusters.

[0112] In some embodiments, the server uses the central coordinates in each cluster coordinate set as the corresponding division center, and divides the target scene into corresponding X recognition areas according to each division center.

[0113] In one embodiment, after the server divides the target scene into recognition areas corresponding to the number of clusters, the server configures the shooting parameters of the preset zoom camera for each recognition area according to the area position of each recognition area.

[0114] In one embodiment, configuring the shooting parameters of the zoom camera specifically includes: configuring the shooting parameters of the zoom camera for each recognition area based on the preset pixel ratio, the preset camera coordinates and the center coordinates in each cluster coordinate set; the preset camera coordinates are the position coordinates of the initial shooting center of the zoom camera. The shooting parameters include the zoom multiple and offset data of the zoom camera.

[0115] The zoom factor is used to limit the pixel ratio of the face object in the image captured by the zoom camera to meet a preset pixel ratio.

[0116] The offset data includes an offset angle and an offset distance of the zoom camera when the shooting center of the zoom camera is offset from the preset camera coordinates to the center coordinates of the cluster coordinate set.

[0117] In one embodiment, when the server evaluates the facial features of each facial object photographed by the zoom camera, it is necessary to eliminate the facial objects that do not meet the preset evaluation criteria, so as to obtain the target facial object among the facial objects.

[0118] In some embodiments, the server obtains the target face object from each face object, specifically including the following steps:

[0119] Step 1: Perform feature recognition processing on the face object to determine the face feature clarity and face feature completeness of the face object.

[0120] Step 2: Based on the clarity and completeness of facial features, obtain an evaluation score for the facial object.

[0121] Step 3: Among the face objects, eliminate the face objects whose corresponding evaluation scores are lower than the preset scores to obtain the target face objects.

[0122] As an example, the server first uses a preset feature recognition algorithm to calculate the facial feature clarity and facial feature completeness of each facial object; then, it performs weighted summation of the facial feature clarity and facial feature completeness of each facial object to obtain an evaluation score for each facial object; finally, among each facial object, the facial objects whose corresponding evaluation scores are lower than the preset scores are eliminated, thereby obtaining the target facial objects photographed by the zoom camera.

[0123] In one embodiment, when the server performs face matching on a target face object, it needs to match the face features of the target face object with a reference face object in a preset face library to obtain a matching result for the target face object, and use the matching result as the recognition result for each target detection object in the target scene.

[0124] The target detection object (i.e., the target face object) is the face object to be identified, and the recognition result represents the face information of the face object to be identified. The face information includes the user information and face feature information of the user corresponding to the face object.

[0125] In one embodiment, after obtaining the recognition result for each target detection object in the target scene, the server may further include the following steps:

[0126] Step 1: Extract a list of face information matching the target scene at the current time stage from a preset database.

[0127] The face information list stores face information of multiple preset face objects, and the target scene is bound to different face information lists at different time stages.

[0128] As an example, for a scenario where students are taking classes in a classroom, the face information list in classroom A from 9:00 to 11:00 in the morning is face information list S1, and the face information list in classroom A from 14:00 to 16:00 in the afternoon is face information list S2. Among them, face information list S1 is different from face information list S2.

[0129] Step 2: Compare the facial information of each face object to be identified with the facial information of the preset face objects in the face information list, obtain the target face information in the face information list that fails the comparison, and mark the target face information.

[0130] As an example, the facial feature information of each face object to be identified is compared with the facial feature information of the preset face objects in the face information list to calculate the feature similarity; then, the face objects whose object similarity is higher than or equal to the preset value are determined to be successfully compared, and the face objects whose object similarity is lower than the preset value are determined to have failed to be compared, thereby obtaining the target face information in the face information list that corresponds to the failed comparison.

[0131] In some embodiments, since the user's facial appearance will change over time, if the facial feature information of the preset face object is relied upon as a reference for a long time to compare the facial information of the face object to be identified, the accuracy of the comparison may become lower over time. Therefore, it is necessary to automatically update the face library, thereby updating the facial feature information of each preset face object in the face information list to improve the accuracy of face comparison.

[0132] Specifically, the server can extract the target facial feature information of the target facial object whose corresponding evaluation score is greater than the preset score from the results of the quality detection of the facial object, and after obtaining the recognition result for the facial object, update the facial feature information of the preset facial object based on the target facial feature information of the target facial object, thereby updating the face library to improve the accuracy of face comparison.

[0133] In order to more clearly illustrate the target recognition method provided by the embodiment of the present disclosure, the target recognition method is specifically described below with reference to a specific embodiment. Figure 4 and Figure 5 , Figure 4 is a flow chart of a student attendance method according to an exemplary embodiment. Figure 5 The module diagram of a student attendance method according to an exemplary embodiment is shown. The student attendance method is used in a server and specifically includes the following contents:

[0134] Step S21: Importing initial face images of all students.

[0135] Specifically, the server imports the initial facial images, student names and student numbers of all students in the school's management background according to the dimensions of college, major, grade and class, and binds the students' initial facial images with their student names and student numbers.

[0136] Step S22: extracting facial features from each initial facial image using a facial feature extraction algorithm.

[0137] Specifically, the server extracts facial key point features from the student's initial facial image through a facial feature extraction algorithm.

[0138] Step S23: Building a face database for each user based on the face features of each initial face image.

[0139] Specifically, the server establishes a face library for each user based on the facial key point features and initial face pictures of the students, and stores the data in the face library in a database.

[0140] Step S24: obtaining a set of images captured by a binocular camera calibrated with external parameters while cruising in the classroom.

[0141] The image set includes two depth images taken by a binocular camera.

[0142] Step S25: stitching the two depth images in the image set to obtain a panoramic stitching image of the classroom.

[0143] Step S26: Based on the face target detection algorithm, face frame detection is performed on the panoramic stitching image to obtain the pixel coordinates of each target face frame in the classroom in the camera coordinate system.

[0144] The pixel coordinates are represented based on (x, y, w, h). x and y refer to the coordinate position of the upper left corner of the target face frame in the camera coordinate system; w and h refer to the pixel width and pixel height of the target face frame, which are used to locate the pixel position of the target face frame in the panoramic stitching image.

[0145] Step S27: Determine the depth coordinates of the target face frame in the camera coordinate system according to the pixel coordinates and the calibration parameters of the binocular camera.

[0146] The depth coordinate is represented by (x, y, d), where d is the depth value of the target face frame.

[0147] Step S28: Based on the K-Means clustering algorithm, the depth coordinates of each target face frame in the panoramic stitching image are clustered to obtain 9 preset clustering coordinate sets.

[0148] Step S29: Divide the current scene into corresponding 9 recognition areas according to the cluster center coordinates in each cluster coordinate set.

[0149] Step S30: According to the preset effective pixel ratio of the screen and the cluster center coordinates of each cluster coordinate set, the zoom factor and rotation displacement of the telephoto gimbal camera for each recognition area are determined respectively.

[0150] Among them, the zoom factor determines the effective proportion of the pixels of the target face occupying the image screen; the rotational displacement represents the offset angle of the camera in the horizontal and vertical directions when the shooting center of the camera shifts from the center of the panoramic stitching image to the corresponding cluster center coordinates in the recognition area.

[0151] Among them, the telephoto gimbal camera is used to enlarge the faces in different recognition areas, thereby increasing the effective pixel ratio of the face image.

[0152] Step S31: Based on the zoom factor and the rotation displacement, the telephoto gimbal camera is controlled to zoom in and shoot each recognition area to obtain a close-up image of each recognition area.

[0153] Step S32: performing face recognition on the close-up images of each recognition area respectively to obtain corresponding face recognition results.

[0154] Among them, the process of face recognition includes face frame detection, face detection, face quality detection and face library matching on the close-up image in sequence.

[0155] Specifically, the server first performs face frame detection on the close-up image to determine the coordinate position of each student face in the close-up image; then performs face detection on the close-up image to determine the facial features of each student face in the close-up image; then performs face quality detection on the close-up image to determine whether the facial features of each student face in the close-up image meet the preset quality requirements, and retains the student faces that meet the quality requirements; finally, the facial features of each retained student face are feature matched with the facial features of each initial student face in the face library to determine the student information corresponding to the retained student faces, thereby obtaining a face recognition result.

[0156] Among them, face quality detection is specifically used to calculate the quality score of each student face from dimensions such as image clarity, face orientation (i.e. front face / side face / back head), and facial feature clarity of the face picture, and determine student faces with quality scores lower than a preset value as student faces that do not meet the quality requirements, and determine student faces with quality scores greater than or equal to the preset value as student faces that meet the quality requirements.

[0157] Among them, face library matching is specifically used for: based on the school timetable, obtaining the facial features of each initial student face related to this class from the pre-built face library; then comparing the facial features of each retained student face with the facial features of the initial student face to calculate the feature similarity; finally, determining that the student face match is successful if the similarity is higher than the preset value (that is, it is considered that the student in the close-up picture is the same person as the student entered in the face library) to determine the student information corresponding to the retained student face, thereby obtaining the face recognition result.

[0158] Among them, since the appearance of students may change over time, if only the initial face picture of the student is compared, the recognition accuracy may decrease over time, so it is necessary to automatically update the face library to improve the accuracy of face comparison. Specifically, the server extracts the face pictures of the target students whose corresponding quality scores are greater than the preset value from the results of the face quality detection, and automatically updates the student face library based on the face pictures of the target students after the student attendance is completed.

[0159] Step S33: Determine the final student attendance result based on the face recognition results of each recognition area.

[0160] Specifically, the server obtains an attendance list based on the face recognition results corresponding to each close-up image; then, based on the school timetable, it obtains the student list of each student who should be present in this class from the database; finally, the attendance list is compared with the student list to determine the final student attendance result, that is, to determine which students have attended and which students have not.

[0161] In this way, on the one hand, in a manner different from the prior art, a binocular camera is first used to cruise and shoot the target scene to obtain a depth image, and then a configured zoom camera is used to take close-up shots of each recognition area in the target scene, so as to perform target recognition processing on the obtained close-up images and obtain recognition results for the target detection objects, thereby optimizing the target recognition process, effectively improving the efficiency of identifying the detection objects in the target scene, and reducing the consumption of manpower and material resources; on the other hand, this scheme divides the target scene into a preset number of recognition areas according to the depth image obtained by the binocular camera, and configures the shooting parameters of the zoom camera for each recognition area according to the corresponding position of the target detection object in the depth image; and then the configured zoom camera is used to take close-up shots of each recognition area, so as to perform target recognition processing on the obtained close-up images, so as to obtain recognition results for the target detection objects in the target scene, effectively improving the accuracy of identifying the target objects, and facilitating reference and optimization of subsequent target object activity behaviors.

[0162] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0163] Based on the same inventive concept, the embodiment of the present application also provides a target recognition system and a target recognition device for implementing the target recognition method involved above. The implementation scheme for solving the problem provided by the target recognition system and the target recognition device is similar to the implementation scheme recorded in the above method, so the specific limitations of one or more unit embodiments provided below can refer to the limitations of the information processing method above, and will not be repeated here.

[0164] In one embodiment, Figure 6 As shown, a target recognition system 10 is provided, comprising: a binocular camera 11, a zoom camera 12, and a processor 13, wherein:

[0165] The binocular camera 11 is used to cruise and shoot the target scene to obtain a corresponding depth image; wherein the target scene includes at least one target detection object.

[0166] The binocular camera 11 is composed of two independent monocular cameras.

[0167] The zoom camera 12 is used to take close-up photos of each recognition area in the target scene based on corresponding shooting parameters to obtain corresponding close-up images.

[0168] The processor 13 is used to divide the target scene into a preset number of recognition areas based on the positions of the target detection objects corresponding to the positions in the depth image; configure the shooting parameters of the zoom camera for each of the recognition areas; and perform target recognition processing on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0169] In another embodiment, Figure 7 As shown, a target recognition device 20 is provided, comprising: an acquisition module 21, a processing module 22, a close-up module 23 and a recognition module 24.

[0170] An acquisition module 21 is used to acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object;

[0171] A processing module 22, configured to divide the target scene into a preset number of identification areas based on the position of the target detection object corresponding to the position in the depth image, and configure a preset zoom camera shooting parameter for each of the identification areas;

[0172] A close-up module 23, used for controlling the zoom camera to take close-up photos of each of the identified areas under the shooting parameters to obtain corresponding close-up images;

[0173] The recognition module 24 is used to perform target recognition processing on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0174] In some embodiments, the target detection object is a face object in the target scene;

[0175] The step of dividing the target scene into a preset number of recognition areas includes:

[0176] Performing coordinate detection on each face object in the depth image to determine the object coordinates of each face object;

[0177] Based on a preset number of clusters, coordinate clustering is performed on the object coordinates of each of the face objects to obtain a set of cluster coordinates corresponding to the number of clusters;

[0178] Based on the central coordinates in each of the cluster coordinate sets, the target scene is divided into recognition areas corresponding to the number of clusters.

[0179] In some embodiments, the configuration presets the shooting parameters of the zoom camera for each of the identified areas, including:

[0180] Based on the preset pixel ratio, the preset camera coordinates and the center coordinates in each cluster coordinate set, configuring the shooting parameters of the zoom camera for each of the identification areas; the preset camera coordinates are the position coordinates of the initial shooting center of the zoom camera;

[0181] Wherein, the shooting parameters include the zoom factor and offset data of the zoom camera;

[0182] The zoom factor is used to limit the pixel ratio of the face object in the image captured by the zoom camera to meet the preset pixel ratio;

[0183] The offset data includes an offset angle and an offset distance of the zoom camera when the shooting center of the zoom camera is offset from the preset camera coordinates to the center coordinates of the cluster coordinate set.

[0184] In some embodiments, the performing target recognition processing on the target detection object in the close-up image includes:

[0185] Performing face detection, coordinate detection, quality detection and face matching on the face object in the close-up image in sequence;

[0186] Wherein, the face detection is used to detect at least one face object present in the close-up image;

[0187] The coordinate detection is used to detect the object coordinates of the face object;

[0188] The quality detection is used to evaluate the facial features of the facial objects, so as to eliminate the facial objects that do not meet the preset evaluation criteria and obtain the target facial objects among the facial objects;

[0189] The face matching is used to perform facial feature matching between the target face object and a reference face object in a preset face library to obtain a matching result for the target face object.

[0190] In some embodiments, the step of evaluating facial features of the facial objects to eliminate facial objects that do not meet a preset evaluation standard and obtain a target facial object from the facial objects includes:

[0191] Performing feature recognition processing on the face object to determine the face feature clarity and face feature integrity of the face object;

[0192] Obtaining an evaluation score for the face object based on the face feature clarity and the face feature completeness;

[0193] Among the face objects, the face objects corresponding to the evaluation scores lower than the preset scores are eliminated to obtain the target face objects.

[0194] In some embodiments, the target detection object is a face object to be identified, and the recognition result represents the face information of the face object to be identified;

[0195] After obtaining the recognition result for each target detection object in the target scene, the method further includes:

[0196] Extracting a face information list matching the target scene at the current time stage from a preset database; wherein the target scene is bound to different face information lists at different time stages, and face information of multiple preset face objects is stored in the face information list;

[0197] The facial information of each facial object to be identified is compared with the facial information of the preset facial objects in the facial information list to obtain the corresponding target facial information in the facial information list that fails the comparison, and the target facial information is marked.

[0198] In some embodiments, the step of acquiring a depth image obtained by cruising and photographing a target scene with a preset binocular camera includes:

[0199] Acquire two monocular depth images taken by the binocular camera during cruising of the target scene;

[0200] The two monocular depth images are stitched together to obtain a panoramic stitching depth image for the target scene.

[0201] In one embodiment, a computer device is provided. The computer device may be an electronic device or a server according to the program logic to be applied. The internal structure diagram of the computer device may be as follows: Figure 8 The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus.

[0202] The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies.

[0203] The computer program is executed to implement a target recognition method. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0204] Those skilled in the art will understand that Figure 8The structure shown in the figure is only a block diagram of a part of the structure related to the present application scheme, and does not constitute a limitation on the computer equipment (including servers and electronic devices) to which the present application scheme is applied. The specific computer equipment may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0205] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed, the following steps are implemented:

[0206] Acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object;

[0207] Based on the position of the target detection object corresponding to the depth image, the target scene is divided into a preset number of recognition areas, and a preset zoom camera is configured with shooting parameters for each of the recognition areas;

[0208] Controlling the zoom camera to take close-up photos of each of the identification areas under the shooting parameters to obtain corresponding close-up images;

[0209] Target recognition processing is performed on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0210] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0211] Acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object;

[0212] Based on the position of the target detection object corresponding to the depth image, the target scene is divided into a preset number of recognition areas, and a preset zoom camera is configured with shooting parameters for each of the recognition areas;

[0213] Controlling the zoom camera to take close-up photos of each of the identification areas under the shooting parameters to obtain corresponding close-up images;

[0214] Target recognition processing is performed on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

[0215] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0216] Those skilled in the art will appreciate that the embodiments of the present application may provide a target identification method, a target identification system, a target identification device, a server, a computer device, a computer-readable storage medium, or a computer program product. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer program instructions (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0217] The present application is described with reference to the flowcharts and / or block diagrams of the target identification method, target identification system, target identification apparatus, server, computer equipment, computer-readable storage medium or computer program product according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by a computer program product. These computer program products can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the program instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0218] These computer program products can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the program instructions stored in the computer program product produce a product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0219] These program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the program instructions executed on the computer or other programmable device for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0220] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0221] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0222] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0223] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A target recognition method, characterized in that: The method comprises: Acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object; the target detection object includes a human face object in the target scene; Determine the depth coordinates of the target detection object based on the position of the target detection object corresponding to the depth image, divide the target scene into a preset number of recognition areas according to the clustering results of the depth coordinates of each of the target detection objects, and configure a preset zoom camera shooting parameter for each of the recognition areas; Controlling the zoom camera to take close-up photos of each of the identification areas under the shooting parameters to obtain corresponding close-up images; Target recognition processing is performed on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

2. The method according to claim 1, characterized in that The method of determining the depth coordinates of the target detection object based on the position of the target detection object corresponding to the position in the depth image, and dividing the target scene into a preset number of recognition areas according to the clustering result of the depth coordinates of each target detection object, includes: Performing coordinate detection on each face object in the depth image, and determining the depth coordinates of each face object as object coordinates; Based on a preset number of clusters, coordinate clustering is performed on the object coordinates of each of the face objects to obtain a set of cluster coordinates corresponding to the number of clusters; Based on the central coordinates in each of the cluster coordinate sets, the target scene is divided into recognition areas corresponding to the number of clusters.

3. The method according to claim 2, characterized in that The configuration presets the shooting parameters of the zoom camera for each of the identification areas, including: Based on the preset pixel ratio, the preset camera coordinates and the center coordinates in each cluster coordinate set, configuring the shooting parameters of the zoom camera for each of the identification areas; the preset camera coordinates are the position coordinates of the initial shooting center of the zoom camera; Wherein, the shooting parameters include the zoom factor and offset data of the zoom camera; The zoom factor is used to limit the pixel ratio of the face object in the image captured by the zoom camera to meet the preset pixel ratio; The offset data includes an offset angle and an offset distance of the zoom camera when the shooting center of the zoom camera is offset from the preset camera coordinates to the center coordinates of the cluster coordinate set.

4. The method according to claim 2, characterized in that: The performing target recognition processing on the target detection object in the close-up image includes: Performing face detection, coordinate detection, quality detection and face matching on the face object in the close-up image in sequence; Wherein, the face detection is used to detect at least one face object present in the close-up image; The coordinate detection is used to detect the object coordinates of the face object; The quality detection is used to evaluate the facial features of the facial objects, so as to eliminate the facial objects that do not meet the preset evaluation criteria and obtain the target facial objects among the facial objects; The face matching is used to perform facial feature matching between the target face object and a reference face object in a preset face library to obtain a matching result for the target face object.

5. The method according to claim 4, characterized in that The step of evaluating the facial features of the facial objects to eliminate the facial objects that do not meet the preset evaluation criteria and obtain the target facial objects among the facial objects includes: Performing feature recognition processing on the face object to determine the face feature clarity and face feature integrity of the face object; Based on the clarity of the facial features and the completeness of the facial features, obtaining an evaluation score for the facial object; Among the face objects, the face objects corresponding to the evaluation scores lower than the preset scores are eliminated to obtain the target face objects.

6. The method according to claim 1, characterized in that The target detection object is a face object to be identified, and the recognition result represents the face information of the face object to be identified; After obtaining the recognition result for each target detection object in the target scene, the method further includes: Extracting a face information list matching the target scene at the current time stage from a preset database; wherein the target scene is bound to different face information lists at different time stages, and face information of multiple preset face objects is stored in the face information list; The facial information of each facial object to be identified is compared with the facial information of the preset facial objects in the facial information list to obtain the corresponding target facial information in the facial information list that fails the comparison, and the target facial information is marked.

7. The method according to claim 1, characterized in that The step of obtaining a depth image obtained by cruising and photographing a target scene with a preset binocular camera includes: Acquire two monocular depth images taken by the binocular camera during cruising of the target scene; The two monocular depth images are stitched together to obtain a panoramic stitching depth image for the target scene.

8. A target recognition system, characterized in that: The system comprises: A binocular camera is used to cruise and shoot a target scene to obtain a corresponding depth image; wherein the target scene includes at least one target detection object; the target detection object includes a human face object in the target scene; A zoom camera, used to take close-up photos of each recognition area in the target scene based on corresponding shooting parameters to obtain corresponding close-up images; A processor is used to determine the depth coordinates of the target detection object based on the position of the target detection object corresponding to the position in the depth image, and divide the target scene into a preset number of recognition areas according to the clustering results of the depth coordinates of each of the target detection objects; configure the shooting parameters of the zoom camera for each of the recognition areas; and perform target recognition processing on the target detection object in the close-up image to obtain a recognition result for each of the target detection objects in the target scene.

9. A target recognition device, characterized in that: The device comprises: An acquisition module, used to acquire a depth image obtained by cruising and photographing a target scene with a preset binocular camera; wherein the target scene includes at least one target detection object; and the target detection object includes a human face object in the target scene; a processing module, configured to determine the depth coordinates of the target detection object based on the position of the target detection object corresponding to the position in the depth image, divide the target scene into a preset number of recognition areas according to the clustering result of the depth coordinates of each of the target detection objects, and configure a preset zoom camera for shooting parameters of each of the recognition areas; A close-up module, used for controlling the zoom camera to take close-up photos of each of the identification areas under the shooting parameters to obtain corresponding close-up images; The recognition module is used to perform target recognition processing on the target detection objects in the close-up image to obtain recognition results for each of the target detection objects in the target scene.

10. A computer device, characterized in that: include: processor; a memory for storing executable instructions for the processor; The processor is configured to execute the executable instructions to implement the target recognition method as described in any one of claims 1 to 7.

11. A computer-readable storage medium, the computer-readable storage medium comprising program data, characterized in that: When the program data is executed by a processor of a computer device, the computer device is enabled to execute the target recognition method according to any one of claims 1 to 7.

Images