Target object identity recognition method, device and equipment

Abstract

The invention provides a target object identity recognition method, device and equipment. The method comprises the following steps: acquiring a first image of a target object, wherein the first image comprises a first surface of the target object; based on the first image, identifying a second image containing the first surface of the target object in a historical image set; obtaining a third image from a historical image set based on the second image; wherein the third image is associated with the second image, and the third image comprises the second surface of the target object; the first surface is one of the front surface and the back surface of the target object, and the second surface is the other one of the front surface and the back surface of the target object; and identifying the identity of the target object based on the third image. According to the method, the front (or back) image of the object can be obtained based on the back (or front) image of the object so as to complete object identity recognition.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device and equipment for identifying the identity of a target object. Background Art

[0002] With the development of image processing technology and the widespread deployment of cameras, the use of images to identify the identity of objects has been widely used in many aspects of social life, such as safe cities and public security investigations.

[0003] Typically, an image of the front (or back) of an object is required to identify the object. However, in some cases, only an image of the back (or front) of the object may be captured. In this case, it is difficult to identify the object. Summary of the Invention

[0004] The present application provides a method, apparatus and device for identifying the identity of a target object, which can obtain a front (or back) image of an object based on a back (or front) image of the object to complete the identification of the object.

[0005] In a first aspect, a method for identifying the identity of a target object is provided, the method comprising: acquiring a first image of the target object, the first image including a first side of the target object; based on the first image, identifying a second image including the first side of the target object in a historical image set; based on the second image, obtaining a third image from the historical image set; wherein the third image is associated with the second image, and the third image includes the second side of the target object; the first side is one of the front and back sides of the target object, and the second side is the other of the front and back sides of the target object; and based on the third image, identifying the identity of the target object.

[0006] When the front (or back) image of the target object is required to identify the identity of the target object, and the back (or front) image of the target object is currently available, this method can be used to obtain the front (or back) image of the target object based on the back (or front) image of the target object currently acquired, and then, based on the front (or back) image of the target object, the identity of the target object can be identified, thereby completing the identification of the identity of the target object.

[0007] In a possible implementation, the method further includes: confirming that a shooting position of the second image and a shooting position of the third image are located on a moving trajectory of the target object; and associating the second image with the third image.

[0008] The shooting positions of the second image and the third image are located on the moving trajectory of the target object, which means that the second image includes the first surface of the target object and the third image includes the second surface of the target object. In this way, the second image and the third image can be associated.

[0009] In one possible implementation, confirming that the shooting position of the second image and the shooting position of the third image are located on the movement trajectory of the target object includes: using a first camera to shoot the target object to obtain a target video; using an image of the first side of the target object in the target video as the second image, and using an image of the second side of the target object in the target video as the third image.

[0010] The second image and the third image are included in the same video. Based on the video, it can be confirmed that the shooting positions of the second image and the third image are located on the movement trajectory of the target object.

[0011] In one possible implementation, confirming that the shooting position of the second image and the shooting position of the third image are located on the moving trajectory of the target object includes: using a second camera to shoot the target object to obtain a first video; using a third camera to shoot the target object to obtain a second video; wherein the shooting range of the second camera and the shooting range of the third camera overlap; based on the first video and the second video, confirming that the shooting position of the second image and the shooting position of the third image are located on the moving trajectory of the target object.

[0012] The second camera's shooting range overlaps with the third camera's shooting range, so the first video captured by the second camera and the second video captured by the third camera also overlap. Based on this overlapping area, it can be confirmed that the shooting positions of the second image and the third image are located on the target object's movement trajectory.

[0013] In one possible implementation, the second image is captured by a fourth camera, the third image is captured by a fifth camera, and the distance between the shooting ranges of the fourth camera and the fifth camera is less than a first threshold; confirming that the shooting positions of the second image and the third image are located on the movement trajectory of the target object includes: based on feature information of the target object captured by the fourth camera and feature information of the target object captured by the fifth camera, confirming that the shooting positions of the second image and the third image are located on the same movement trajectory of the target object. The feature information of the target object captured by the fourth camera may include feature information of the target object identified from the second image, and the feature information of the target object captured by the fifth camera may include feature information of the target object identified from the third image.

[0014] When there is no overlapping area between the shooting ranges of the fourth camera that captures the second image and the fifth camera that captures the third image, and the distance between the shooting ranges of the fourth camera and the fifth camera is less than a first threshold, it is possible to confirm, based on the feature information of the target object captured by the fourth camera and the feature information of the target object captured by the fifth camera, that the second image and the third image are located on the same movement trajectory, and further confirm that the second image includes the first side of the target object and the third image includes the second side of the target object.

[0015] In one possible implementation, a historical image set includes feature information of a first side of a target object in a second image and feature information of a second side of the target object in a third image; wherein the feature information of the first side of the target object in the second image is associated with the feature information of the second side of the target object in the third image; based on the first image, identifying a second image containing the first side of the target object in the historical image set includes: obtaining, from the historical image set, the feature information of the first side of the target object in the second image based on the feature information of the first side of the target object in the first image; obtaining, from the historical image set, a third image based on the second image, including: obtaining, from the historical image set, the feature information of the second side of the target object in the third image based on the feature information of the first side of the target object in the second image; and identifying, based on the third image, the identity of the target object, including: identifying the identity of the target object based on the feature information of the second side of the target object in the third image.

[0016] In this implementation, feature information of the target object in the image can be stored, and then, related matching, recognition operations, etc. can be performed based on the feature information. Compared with storing image data, storing feature information can save storage resources.

[0017] In a possible implementation, the target object is any one of a vehicle, a person, and an animal.

[0018] In a second aspect, a device for identifying the identity of a target object is provided, the device comprising: an acquisition unit for acquiring a first image of the target object, the first image including a first side of the target object; a first identification unit for identifying a second image including the first side of the target object in a historical image set based on the first image; an obtaining unit for obtaining a third image from the historical image set based on the second image; wherein the third image is associated with the second image, and the third image includes the second side of the target object; the first side is one of the front and back sides of the target object, and the second side is the other of the front and back sides of the target object; and a second identification unit for identifying the identity of the target object based on the third image.

[0019] In a possible implementation, the device further includes: a confirmation unit, configured to confirm that the shooting positions of the second image and the third image are located on a moving trajectory of the target object; and an association unit, configured to associate the second image with the third image.

[0020] In one possible implementation, the confirmation unit is used to: use a first camera to shoot the target object to obtain a target video; use the image of the first side of the target object in the target video as the second image, and use the image of the second side of the target object in the target video as the third image.

[0021] In one possible implementation, the confirmation unit is used to: use a second camera to shoot the target object to obtain a first video; use a third camera to shoot the target object to obtain a second video; wherein the shooting range of the second camera and the shooting range of the third camera overlap; based on the first video and the second video, confirm that the shooting position of the second image and the shooting position of the third image are located on the moving trajectory of the target object.

[0022] In one possible implementation, the second image is captured by a fourth camera, the third image is captured by a fifth camera, and the distance between the shooting range of the fourth camera and the shooting range of the fifth camera is less than a first threshold; the confirmation unit is used to: based on the feature information of the target object captured by the fourth camera and the feature information of the target object captured by the fifth camera, confirm that the shooting position of the second image and the shooting position of the third image are located on the same movement trajectory of the target object.

[0023] In one possible implementation, the historical image set includes feature information of the first side of the target object in the second image and feature information of the second side of the target object in the third image; wherein the feature information of the first side of the target object in the second image is associated with the feature information of the second side of the target object in the third image; the first recognition unit is used to: obtain the feature information of the first side of the target object in the second image from the historical image set based on the feature information of the first side of the target object in the first image; the obtaining unit is used to: obtain the feature information of the second side of the target object in the third image based on the feature information of the first side of the target object in the second image; the second recognition unit is used to: identify the identity of the target object based on the feature information of the second side of the target object in the third image.

[0024] In a possible implementation, the target object is any one of a vehicle, a person, and an animal.

[0025] In a third aspect, a computing device is provided, comprising: a memory for storing an executable program; and a processor for executing the method provided in the first aspect by running the executable program.

[0026] In a fourth aspect, a computer-readable storage medium is provided, comprising computer program instructions. When the computer program instructions are executed by a computing device, the computing device executes the method provided in the first aspect.

[0027] In a fifth aspect, a computer program product comprising instructions is provided, which, when executed by a computing device, causes the computing device to execute the method provided in the first aspect.

[0028] Among them, the beneficial effects of the second to fifth aspects can be found in the above description of the beneficial effects of the first aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic diagram of a system structure provided in an embodiment of the present application;

[0030] Figure 2 A flowchart of a target object identification method provided in an embodiment of the present application;

[0031] Figure 3 A schematic diagram of an image association solution provided in an embodiment of the present application;

[0032] Figure 4 A schematic diagram of an image association solution provided in an embodiment of the present application;

[0033] Figure 5 A schematic diagram of an image association solution provided in an embodiment of the present application;

[0034] Figure 6 A flowchart of an image association solution provided in an embodiment of the present application;

[0035] Figure 7 A flowchart of a target object identification method provided in an embodiment of the present application;

[0036] Figure 8 A schematic diagram of the structure of a target object identification device provided in an embodiment of the present application;

[0037] Figure 9 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] The following describes the solutions provided by the embodiments of the present application in conjunction with the accompanying drawings. In the embodiments of the present application, "plurality" refers to two or more, and "multiple" refers to two or more. Terms such as "first" and "second" are used only to distinguish similar objects and do not necessarily describe a specific order or quantity of objects.

[0039] To facilitate understanding of the solutions provided by the embodiments of the present application, the technical terms that may be involved in the embodiments of the present application are first introduced.

[0040] Subject: This is what the camera captures. It can be any entity, such as a vehicle, a person, or an animal.

[0041] Structured information, also known as intelligent structured data (metadata), is information that, after analysis, can be broken down into multiple interconnected components, each with a clear hierarchical structure. It is generally used to represent further detailed information such as the attributes and characteristics of an object. Structured information about a person might include gender, age, height, hairstyle, whether they wear glasses, and clothing. Structured information about a vehicle might include color, length, and width.

[0042] Semi-structured information: is more abstract than structured information.

[0043] Human body feature value: a vector space used to represent human body features, used for human body retrieval.

[0044] Smart camera: A camera with data processing capabilities that can capture objects and extract features.

[0045] Monocular ball camera: a spherical camera with a rotating lens, one lens, and the ability to track targets.

[0046] Multi-camera: A spherical camera with a rotating lens, multiple lenses, which are independent of each other and have the ability to track targets.

[0047] Front: refers to one side of an object. The front of a person is the side with their face. The front of a vehicle is the front of the vehicle.

[0048] Back: The side opposite to the front. The back of a person is the back of the body. The back of a vehicle is the back of the vehicle.

[0049] Frontal image: refers to the image of an object captured by a camera when the front of the object faces the camera.

[0050] Back image: refers to the image of an object captured by a camera when the back of the object faces the camera.

[0051] The image's shooting position refers to the location of the target object in the image when the camera captured the image. In other words, the image is captured by the camera facing the object, and the image's shooting position is the location of the object when the camera captured the image.

[0052] Sometimes, a front-facing image of an object may be needed to confirm the identity of the currently captured object. However, in many cases, only the back-facing image of the object can be captured. The currently captured back-facing image itself cannot be used to confirm the identity of the object. Moreover, the related art can only realize back-searching by back. Specifically, only the back-facing image of the object can be searched in the historical image database using the back-facing image of the object, or the accuracy of the front-facing image searched in the historical image database using the back-facing image of the object is very low, and the search result is unusable. Therefore, in the related art, when only the back-facing image of the object is captured, the identity of the object cannot be confirmed.

[0053] Sometimes, the back image of the object may be needed to confirm the identity of the currently photographed object. However, in many cases, only the front image of the object can be photographed. The currently photographed front image itself cannot be used to confirm the identity of the object. Moreover, the related art can only implement front-to-front search. Specifically, the front image of the object can only be searched in the historical image database for the front image, or the accuracy of the back image searched in the historical image database for the front image of the object is very low, and the search result is unusable. Therefore, in the related art, when only the front image of the object is photographed, the identity of the object cannot be confirmed.

[0054] An embodiment of the present application provides a method for identifying a target object. This method can, upon capturing an image of surface A1 of a target object, obtain an image of surface A2 of the target object based on the image, and identify the target object based on the image of surface A2. Surface A1 is one of the front and back sides of the target object, and surface A2 is the other of the front and back sides. Thus, when only an image of surface A1 of the target object is captured, an image of surface A2 of the target object can be obtained, thereby identifying the target object.

[0055] Next, the target object identification method provided in the embodiment of the present application is described.

[0056] Figure 1 A system architecture that can be used to implement the target object identification method provided in the embodiment of the present application is shown. Figure 1 As shown, the system architecture includes a platform 100, a camera 200, and at least one camera 300. Camera 200 can be located remotely from platform 100, and the camera 200 and platform 100 can be connected via a network. Camera 300 can also be located remotely from platform 100, and the camera 300 and platform 100 can be connected via a network. Camera 300 and camera 200 can be located in different geographical locations, and can be a considerable distance apart. For example, camera 200 and camera 300 can be located on different streets or in different cities.

[0057] The platform 100 can be any device, equipment, platform, or cluster with data processing and data storage capabilities. The platform 100 can receive and store data such as images sent by the camera 300. The platform 100 can receive an image B1 containing the surface A1 of the target object sent by the camera 200, and based on this image B1, search the images stored on the platform 100 for an image B3 containing the surface A2 of the target object, and identify the target object based on image B3.

[0058] The camera 300 can capture an image B2 of the surface A1 of the target object and an image B3 of the surface A2 of the target object. In some embodiments, the camera 300 can transmit the image B2 of the surface A1 of the target object and the image B3 of the surface A2 of the target object to the platform 100. The platform 100 associates the image B2 of the surface A1 of the target object with the image B3 of the surface A2 of the target object to obtain and store a front-to-back image combination of the target object. In some embodiments, the camera 300 can associate the image B2 of the surface A1 of the target object with the image B3 of the surface A2 of the target object to obtain a front-to-back image combination of the target object, and then transmit the front-to-back image combination of the target object to the platform 100.

[0059] Camera 200 can capture image B1 of surface A1 of the target object. This image of surface A1 cannot be directly used to identify the target object; an image of surface A2 is required. In this case, camera 200 can send image B1 to platform 100. Platform 100 can then match image B1 with image B2, thereby obtaining image B3 from the front-back image combination containing image B2. The target object can then be identified based on image B3.

[0060] In the embodiment of the present application, the camera 200 or the camera 300 can be any device or equipment with an image capture function, such as a smart camera, a monocular camera, a multi-camera camera, a network camera, etc.

[0061] Furthermore, the phrase "an image containing an object" refers to a shadow of the image containing the object, and the phrase "a surface containing an object in an image ..." refers to a shadow of the surface of the image containing the object.

[0062] The above example introduces a system architecture provided by an embodiment of the present application. Next, in conjunction with this system architecture, the target object identification method provided by an embodiment of the present application is introduced.

[0063] like Figure 2 As shown, the method includes the following steps.

[0064] Step 201: Acquire an image B1 of a target object, where the image B1 includes the surface A1 of the target object. Step 201 may be performed by the platform 100. In one example, the target object may be a vehicle. In another example, the target object may be a person. In another example, the target object may be an animal.

[0065] When the target object enters the shooting range of the camera 200 and the face A1 of the target object faces the camera 200, the camera 200 shoots the obtained image B1. The camera 200 can send the image B1 to the platform 100. Thus, the platform 100 can obtain the image B1.

[0066] Step 202 : Based on the image B1 , identify the image B2 including the surface A1 of the target object in the historical image collection. Step 202 may be executed by the platform 100 .

[0067] The images in the historical image set are images captured historically by cameras such as camera 300. The historical image set includes at least one front-back image combination of an object. Each front-back image combination of an object includes an image of the object's A1 side and an image of the object's A2 side. In other words, in the historical image set, the A1 and A2 images of the same object are associated.

[0068] Next, taking associating image B2 containing surface A1 of the target object and image B3 containing surface A2 of the target object as an example, a scheme for associating images of surface A1 and surface A2 of the same object will be described.

[0069] Camera 300 can capture images B2 and B3. If it is confirmed that surface A1 in image B2 and surface A2 in image B3 belong to the same object, images B2 and B3 are associated to obtain a front-back image combination of the target object. Determining whether surface A1 in image B2 and surface A2 in image B3 belong to the same object and associating images B2 and B3 can be performed by platform 100 or camera 300.

[0070] In some embodiments, platform 100 or camera 300 may determine that the shooting locations of image B2 and image B3 are located on the movement trajectory of the same object. There is a one-to-one correspondence between movement trajectories and objects; that is, the movement of an object generates a movement trajectory, and a movement trajectory is generated by the movement of an object. If the shooting locations of image B2 and image B3 are located on the movement trajectory of the same object, platform 100 or camera 300 may confirm that surface A1 in image B2 and surface A2 in image B3 belong to the same object, and then associate image B2 with image B3. In other words, if it is confirmed that the shooting locations of image B2 and image B3 are located on the movement trajectory of the target object, image B2 and image B3 are associated to obtain a front-back image combination of the target object. When the association of image B2 and image B3 is performed by camera 300, camera 300 may send the front-back image combination of the target object to platform 100. Platform 100 may include the front-back image combination of the target object in the historical image collection.

[0071] In one example of this embodiment, Figure 3As shown, camera 300 can be a single camera. Exemplarily, the camera 300 has a sufficiently large shooting range, for example, the camera 300 has a sufficiently large field of view (FOV) or a wide-angle camera. Exemplarily, the lens of camera 300 can move with the movement of the target object, for example, camera 300 is a monocular camera or a multi-lens camera. The target object moves within the shooting range of camera 300. When the target object is at point C1 within the shooting range of camera 300, surface A1 of the target object faces camera 300, thereby allowing camera 300 to capture image B2. When the target object is at point C2 within the shooting range of camera 300, surface A2 of the target object faces camera 300, thereby allowing camera 300 to capture image B3. In other words, camera 300 can capture a moving target object to obtain a target video. This target video includes an image of surface A1 of the target object and an image of surface A2 of the target object. The image containing the target object A1 surface in the target video can be used as image B2, and the image containing the target object A2 surface in the target video can be used as image B3. The target video is shot by the same camera (i.e., camera 300). Therefore, the target video can be used to determine whether the shooting position of the image containing the target object A1 surface (i.e., point C1) and the shooting position of the image containing the target object A2 surface (i.e., point C2) in the target video are located on the movement trajectory of the same object. When it is confirmed that the shooting positions of the image containing the target object A1 surface and the image containing the target object A2 surface in the target video are located on the movement trajectory of the same object, the image containing the target object A1 surface in the target video is used as image B2, and the image containing the target object A2 surface in the target video is used as image B3.

[0072] In this embodiment of the present application, the shooting position of an image refers to the position of the object in the image when the image is shot.

[0073] In another example of this embodiment, Figure 4As shown, camera 300 may include multiple cameras, such as camera 300A and camera 300B. The shooting range of camera 300A overlaps with the shooting range of camera 300B. When the target object moves within the shooting range of camera 300A, camera 300A can capture video D1. That is, camera 300A captures the moving target object, generating video D1. When the target object is at point C1 within the shooting range of camera 300A, surface A1 of the target object faces camera 300A. Thus, camera 300A can capture an image containing surface A1 of the target object, i.e., video D1 includes an image containing surface A1 of the target object. The target object moves to the overlapping area between the shooting range of camera 300A and the shooting range of camera 300B, thereby entering the shooting range of camera 300B. Camera 300B can capture the moving target object, generating video D2. When the target object is at point C2 within the shooting range of camera 300B, surface A2 of the target object faces camera 300B. Thus, camera 300B can capture an image containing surface A2 of the target object, that is, video D2 contains an image containing surface A2 of the target object.

[0074] The image containing the target object's surface A2 in video D2 can be used as image B3, and the image containing the target object's surface A1 in video D1 can be used as image B2. Furthermore, through videos D1 and D2, it can be confirmed that the shooting locations of images B3 and B2 are located on the target object's movement trajectory. This is because the target object moves to the overlapping area between the shooting ranges of camera 300A and camera 300B, thus entering the shooting range of camera 300B. Therefore, there is also an overlap between videos D1 and D2. This overlap can be used to confirm the target object's continuous movement trajectory, and thus it can be confirmed that the shooting locations of images B3 and B2 are located on the target object's movement trajectory.

[0075] In another example of this embodiment, Figure 4 As shown, camera 300 may include multiple cameras, such as camera 300C and camera 300D. The shooting ranges of camera 300C and camera 300D do not overlap, but the distance between the shooting ranges of camera 300C and camera 300D is less than a threshold value Y1. Threshold value Y1 may be a preset value, which may be determined based on experience or experimentation. For example, threshold value Y1 may be 10 meters, 50 meters, or 100 meters.

[0076] When the target object is at position C1, camera 300C can capture image B2 containing surface A1 of the target object, meaning that image B2 was captured at position C1. When the target object is at position C2, camera 300D can capture image B3 containing surface A2 of the target object, meaning that image B3 was captured at position C2. When the time difference between the time camera 300C captured image B2 and the time camera 300D captured image B2 is less than threshold Y2, it can be determined whether positions C1 and C2 are located on the same object's movement trajectory. Threshold Y2 can be set based on threshold Y1. For example, based on the target object's movement pattern when camera 300C captured image B2 or when camera 300D captured image B3 and threshold Y1, the maximum duration required for the target object to move by threshold Y1 according to that pattern can be determined. This maximum duration can then be used as threshold Y2.

[0077] Based on the characteristic information of the target object captured by camera 300C and the characteristic information of the target object captured by camera 300D, it can be determined whether position C1 and position C2 are located on the movement trajectory of the same object. If the characteristic information of the target object captured by camera 300C and the characteristic information of the target object captured by camera 300D are consistent, it can be determined that position C1 and position C2 are located on the movement trajectory of the same object. In other words, position C1 and position C2 are located on the movement trajectory of the target object. If the characteristic information of the target object captured by camera 300C and the characteristic information of the target object captured by camera 300D are inconsistent, it can be determined that position C1 and position C2 are not located on the movement trajectory of the same object.

[0078] The feature information of the target object may be extracted from the captured image of the target object. For example, the feature information of the target object may be structured information or semi-structured information of the target object.

[0079] In the above manner, the image B2 and the image B3 can be associated to obtain a front and back image combination of the target object, wherein the front and back image combination of the target object is composed of the image B2 and the image B3.

[0080] After obtaining the front and back image combination of the target object, the platform 100 can add the front and back image combination of the target object to the historical image set, that is, add image B2 and image B3 to the historical image set, wherein image B2 and image B3 are associated. In step 202, when the platform 100 receives image B1 sent by the camera 200, it can match image B2 in the historical image set based on image B1, thereby identifying image B2 in the historical image set. Specifically, image B1 and image B2 both contain the A1 side of the target object, therefore, image B1 and image B2 have a high similarity, or in other words, the feature information of the object in the image and the feature information of the object have a high similarity. Therefore, it can be confirmed that image B1 and image B2 match based on the similarity of the images or the similarity of the feature information of the objects in the images.

[0081] Next, in step 203, based on image B2, image B3 is obtained from the historical image set; wherein, image B3 is associated with image B2, and image B3 includes surface A2 of the target object; surface A1 is one of the front and back sides of the target object, and surface A2 is the other of the front and back sides of the target object.

[0082] Image B2 and image B3 belong to the same front-back image combination. By obtaining image B2, the front-back image combination in which image B2 is located can be obtained, and image B3 can be obtained from the front-back image combination.

[0083] In this way, the image B1 of the surface A1 including the target object taken by the camera 200 can be used to obtain the image B3 of the surface A2 including the target object taken by the camera 300, thereby realizing the search for the back with the front or the search for the front with the back.

[0084] Step 204 : Identify the identity of the target object based on image B3 .

[0085] The target object's identity is associated with face A2 of the target object, and the target object's identity can be obtained using image B3. For example, in an identity information database, the target object's identity is associated with face A2 of the target object. Based on image B3, the target object's identity can be matched from the identity database.

[0086] In some embodiments, the historical image set includes feature information of the target object surface A1 in image B2 and feature information of the target object surface A2 in image B3. The feature information may be structured information or semi-structured information.

[0087] In one example of this embodiment, Figure 6As shown, in step 601, the camera 300 performs target object detection, i.e., detects an object of interest and sets it as the target object. Next, in step 602, target object tracking is performed to track the target object's movement trajectory. Next, in step 603, target object optimization is performed, i.e., a suitable shooting angle is selected. Then, in step 604, an image of surface A1 of the target object is captured and features extracted. Image B2 is obtained by capturing surface A1 of the target object, and feature extraction is performed to obtain feature information of surface A1 of the target object in image B2. Then, in step 605, target object tracking is performed again, and in step 606, target object optimization is performed again. In step 607, an image of surface A2 of the target object is captured and features extracted. Image B3 is obtained by capturing surface A2 of the target object, and feature extraction is performed to obtain feature information of surface A2 of the target object in image B3. In step 608, data upload is performed. Specifically, images B2 and B3 are associated, and the associated images B2 and B3 are uploaded to the platform 110. The feature information of the target object's surface A1 in the image B2 and the feature information of the target object's surface A2 in the image B3 are associated, and the associated feature information of the target object's surface A1 and the feature information of the target object's surface A2 in the image B3 are uploaded to the platform 110 .

[0088] In this embodiment, see Figure 7 Step 202 includes step 701, based on image B1, extracting feature information of surface A1 of the target object in image B1. Image B1 includes surface A1 of the target object. Feature information of surface A1 of the target object can be extracted from image B1.

[0089] Step 202 also includes step 702, feature information retrieval, which is to obtain feature information of face A1 of the target object in image B2 from the historical image collection based on the feature information of face A1 of the target object in image B1. Step 202 may also include step 703, retrieving the feature information of face A1 of the target object in image B2.

[0090] Step 203 includes step 702 , obtaining feature information of surface A2 of the target object in image B3 based on feature information of surface A1 of the target object in image B2 .

[0091] Step 204 includes identifying the identity of the target object based on the feature information of the surface A2 of the target object in the image B3.

[0092] In summary, the method provided in the embodiments of the present application, when the front (or back) image of the target object is required to identify the identity of the target object, and the back (or front) image of the target object is currently available, the front (or back) image of the target object can be obtained based on the back (or front) image of the target object currently acquired, thereby identifying the identity of the target object.

[0093] This embodiment of the application also provides a target object identification device 800. Figure 8 As shown, the apparatus 800 includes:

[0094] An acquiring unit 810 is configured to acquire a first image of a target object, where the first image includes a first surface of the target object.

[0095] A first recognition unit 820 is configured to recognize, based on the first image, a second image containing the first surface of the target object in a historical image set;

[0096] an obtaining unit 830 configured to obtain a third image from the historical image set based on the second image; wherein the third image is associated with the second image and includes the second side of the target object; the first side is one of the front side and the back side of the target object, and the second side is the other of the front side and the back side of the target object;

[0097] The second recognition unit 840 is configured to recognize the identity of the target object based on the third image.

[0098] In some embodiments, the device 800 further includes: a confirmation unit (not shown) for confirming that the shooting position of the second image and the shooting position of the third image are located on the moving trajectory of the target object; and an association unit (not shown) for associating the second image and the third image.

[0099] In an example of this embodiment, the confirmation unit is used to: use a first camera to shoot the target object to obtain a target video; use the image of the first side of the target object in the target video as the second image, and use the image of the second side of the target object in the target video as the third image.

[0100] In another example of this embodiment, the confirmation unit is used to: use a second camera to shoot the target object to obtain a first video; use a third camera to shoot the target object to obtain a second video; wherein the shooting range of the second camera and the shooting range of the third camera overlap; based on the first video and the second video, confirm that the shooting position of the second image and the shooting position of the third image are located on the moving trajectory of the target object.

[0101] In another example of this embodiment, the second image is captured by a fourth camera, the third image is captured by a fifth camera, and the distance between the shooting range of the fourth camera and the shooting range of the fifth camera is less than a first threshold; the confirmation unit is used to: based on the feature information of the target object captured by the fourth camera and the feature information of the target object captured by the fifth camera, confirm that the shooting position of the second image and the shooting position of the third image are located on the same moving trajectory of the target object.

[0102] In some embodiments, the historical image set includes feature information of the first side of the target object in the second image and feature information of the second side of the target object in the third image; wherein, the feature information of the first side of the target object in the second image is associated with the feature information of the second side of the target object in the third image; the first recognition unit 820 is used to: obtain the feature information of the first side of the target object in the second image from the historical image set based on the feature information of the first side of the target object in the first image; the obtaining unit 830 is used to: obtain the feature information of the second side of the target object in the third image based on the feature information of the first side of the target object in the second image; the second recognition unit 840 is used to: identify the identity of the target object based on the feature information of the second side of the target object in the third image.

[0103] In some embodiments, the target object is any one of a vehicle, a person, and an animal.

[0104] The functions of the functional units of the device 800 can also refer to the above description of Figure 2 The introduction and implementation of the various method embodiments shown will not be repeated here.

[0105] The embodiment of the present application provides a computing device 900. Figure 9 As shown, the computing device 900 includes a processor 910 and a memory 920. The memory 920 is used to store executable programs. The processor 910 is used to execute the executable programs stored in the memory 920, so that the computing device 900 can execute the above Figure 2 The method shown.

[0106] It is understood that the processor in the embodiments of the present application may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0107] The present application also provides a computer program product containing instructions. The computer program product may be a software or program product containing instructions that can be run on a computing device or stored in any available medium. When the computer program product is run on a computing device, the computing device is caused to execute Figure 2 The method shown.

[0108] The present application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that can be stored by a computing device or a data storage device such as a data center that contains one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to execute Figure 2 The method shown.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the protection scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying a target object, characterized in that: The method comprises: Acquire a first image of a target object, where the first image includes a first surface of the target object; Based on the first image, identifying a second image containing the first side of the target object in a collection of historical images; Based on the second image, obtaining a third image from the historical image collection; wherein the third image is associated with the second image and includes a second side of the target object; the first side is one of the front side and the back side of the target object, and the second side is the other of the front side and the back side of the target object; Based on the third image, the identity of the target object is identified.

2. The method according to claim 1, characterized in that The method further comprises: confirming that a shooting position of the second image and a shooting position of the third image are located on a moving trajectory of the target object; The second image and the third image are associated.

3. The method according to claim 2, characterized in that The confirming that the shooting position of the second image and the shooting position of the third image are located on the moving track of the target object includes: Using a first camera to shoot the target object to obtain a target video; An image including the first surface of the target object in the target video is used as the second image, and an image including the second surface of the target object in the target video is used as the third image.

4. The method according to claim 2, characterized in that The confirming that the shooting position of the second image and the shooting position of the third image are located on the moving track of the target object includes: Using a second camera to shoot the target object to obtain a first video; Using a third camera to shoot the target object to obtain a second video; wherein the shooting range of the second camera overlaps with the shooting range of the third camera; Based on the first video and the second video, it is confirmed that the shooting position of the second image and the shooting position of the third image are located on the movement trajectory of the target object.

5. The method according to claim 2, characterized in that The second image is captured by a fourth camera, the third image is captured by a fifth camera, and a distance between a shooting range of the fourth camera and a shooting range of the fifth camera is less than a first threshold; The confirming that the shooting position of the second image and the shooting position of the third image are located on the moving track of the target object includes: Based on the feature information of the target object captured by the fourth camera and the feature information of the target object captured by the fifth camera, it is confirmed that the shooting position of the second image and the shooting position of the third image are located on the same movement trajectory of the target object.

6. The method according to any one of claims 1 to 5, characterized in that The historical image set includes feature information of a first surface of the target object in the second image and feature information of a second surface of the target object in the third image; wherein the feature information of the first surface of the target object in the second image and the feature information of the second surface of the target object in the third image are associated; The step of identifying, based on the first image, a second image containing the first surface of the target object in a historical image collection includes: Based on the feature information of the first surface of the target object in the first image, obtaining the feature information of the first surface of the target object in the second image from the historical image set; The obtaining, based on the second image, a third image from the historical image set includes: obtaining feature information of a second surface of the target object in the third image based on feature information of the first surface of the target object in the second image; The identifying the identity of the target object based on the third image includes: The identity of the target object is identified based on the feature information of the second surface of the target object in the third image.

7. The method according to any one of claims 1 to 6, characterized in that The target object is any one of a vehicle, a person, and an animal.

8. A target object identification device, characterized in that: The device comprises: an acquiring unit, configured to acquire a first image of a target object, wherein the first image includes a first surface of the target object; a first recognition unit, configured to recognize, based on the first image, a second image containing the first surface of the target object in a historical image set; an obtaining unit, configured to obtain a third image from the historical image set based on the second image; wherein the third image is associated with the second image and includes the second side of the target object; the first side is one of the front side and the back side of the target object, and the second side is the other of the front side and the back side of the target object; The second recognition unit is configured to recognize the identity of the target object based on the third image.

9. The device according to claim 8, characterized in that The device further comprises: a confirmation unit, configured to confirm that a shooting position of the second image and a shooting position of the third image are located on a moving track of the target object; An associating unit is configured to associate the second image with the third image.

10. The device according to claim 9, characterized in that The confirmation unit is used for: Using a first camera to shoot the target object to obtain a target video; An image including the first surface of the target object in the target video is used as the second image, and an image including the second surface of the target object in the target video is used as the third image.

11. The device according to claim 9, characterized in that The confirmation unit is used for: Using a second camera to shoot the target object to obtain a first video; Using a third camera to shoot the target object to obtain a second video; wherein the shooting range of the second camera overlaps with the shooting range of the third camera; Based on the first video and the second video, it is confirmed that the shooting position of the second image and the shooting position of the third image are located on the movement trajectory of the target object.

12. The device according to claim 9, characterized in that The second image is captured by a fourth camera, the third image is captured by a fifth camera, and a distance between a shooting range of the fourth camera and a shooting range of the fifth camera is less than a first threshold; The confirmation unit is used for: Based on the feature information of the target object captured by the fourth camera and the feature information of the target object captured by the fifth camera, it is confirmed that the shooting position of the second image and the shooting position of the third image are located on the same movement trajectory of the target object.

13. The device according to any one of claims 8 to 12, characterized in that The historical image set includes feature information of a first surface of the target object in the second image and feature information of a second surface of the target object in the third image; wherein the feature information of the first surface of the target object in the second image and the feature information of the second surface of the target object in the third image are associated; The first recognition unit is configured to obtain, from the historical image set, feature information of the first surface of the target object in the second image based on feature information of the first surface of the target object in the first image; The obtaining unit is configured to obtain feature information of a second surface of the target object in the third image based on feature information of the first surface of the target object in the second image; The second recognition unit is configured to recognize the identity of the target object based on feature information of the second surface of the target object in the third image.

14. The device according to any one of claims 8 to 13, characterized in that The target object is any one of a vehicle, a person, and an animal.

15. A computing device, characterized in that include: a memory for storing executable programs; A processor, configured to execute the method according to any one of claims 1 to 7 by running the executable program.

16. A computer-readable storage medium, characterized in that The method comprises computer program instructions, which, when executed by a computing device, causes the computing device to perform the method according to any one of claims 1 to 7.

17. A computer program product comprising instructions, characterized in that When the instructions are executed by a computing device, the computing device is caused to perform the method according to any one of claims 1 to 7.

Images