Article tracing method, device and equipment
By obtaining and identifying item information in video images and establishing item archive files, the problem of difficult to trace the location of items in a messy environment is solved, and fast and accurate item traceability and location positioning are achieved.
Patent Information
- Application Number
- CN202310928470.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-26
- Publication Date
- 2025-02-07
AI Technical Summary
In daily life, especially in a messy environment, it is difficult to quickly and accurately trace the location of the item, especially the problem of not finding it due to the random placement of the item or the position of others.
By obtaining video images in the target area, extracting video frames, and using visual models to identify video frames, obtaining item information, and creating item archive files. Then, upon receiving the item traceability request, the archive file of the target item is obtained from the archive file and returned to the location used to trace the item.
It realizes rapid and accurate traceability of items, and can find the specific location of items through backtracking archive files when items are lost, improving the efficiency and accuracy of item tracking.
Smart Images
Figure BDA0004364616730000051 
Figure BDA0004364616730000061 
Figure HDA0004364616760000011
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of item tracing, and in particular to an item tracing method, device and equipment. Background Art
[0002] In daily life, it is often difficult to recall where an item is placed when needed, or it is difficult to find an item because someone else has moved it. It is especially difficult to find an item in a relatively messy environment. How to quickly and accurately trace items has become an urgent problem to be solved. Summary of the invention
[0003] In view of this, the present application provides an item tracing method, device and equipment for quickly and accurately tracing items.
[0004] Specifically, the present application is implemented through the following technical solutions:
[0005] The first aspect of the present application provides an item tracing method, the method comprising:
[0006] Acquire a video image within the target area, and extract at least one video frame from the video image;
[0007] Using the visual macro model to recognize each of the at least one video frame, to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item contained in each video frame;
[0008] According to the recognition result, establishing an archive file for each object in the at least one video frame; wherein the archive file includes the video frame containing the corresponding object;
[0009] When a tracing request for tracing a target item is received, a target archive file of the target item is obtained from the established archive files, and the target archive file is returned; wherein the target archive file is used to trace the target item.
[0010] The second aspect of the present application provides an item tracing device, the device comprising an acquisition module, an identification module, a file building module and a processing module; wherein:
[0011] The acquisition module is used to acquire a video image within a target area and extract at least one video frame from the video image;
[0012] The recognition module is used to recognize each of the at least one video frame using the visual macro model to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item contained in each video frame;
[0013] The archiving module is used to create an archive file for each object in the at least one video frame according to the recognition result; wherein the archive file includes the video frame containing the corresponding object;
[0014] The processing module is used to obtain a target archive file of the target item from the established archive files and return the target archive file when receiving a tracing request for tracing the target item; wherein the target archive file is used to trace the target item.
[0015] The third aspect of the present application provides an item tracing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any one of the methods provided in the first aspect of the present application are implemented.
[0016] The fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the methods provided in the first aspect of the present application.
[0017] The object tracing method, device and equipment provided by the present application obtains a video image in a target area, extracts at least one video frame from the video image, and then uses a visual macro model to identify each video frame in the at least one video frame to obtain the recognition result of the at least one video frame, wherein the recognition result includes the item information of each item contained in each video frame, thereby establishing an archive file of each item in the at least one video frame according to the recognition result; wherein the archive file includes a video frame containing the corresponding item. In this way, when a tracing request for tracing a target item is received, the target archive file of the target item can be obtained from the established archive file, and the target archive file can be returned. In this way, the item can be tracked and recorded based on the archive file of the item, and then when the item is lost, the specific location of the item can be found by tracing back the archive file of the item, so that the item can be traced quickly and accurately. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 A flowchart of the first embodiment of the method for tracing items provided by the present application;
[0019] Figure 2 A flowchart of Example 2 of the item tracing method provided in this application;
[0020] Figure 3 A flowchart of the third embodiment of the item tracing method provided in this application;
[0021] Figure 4A schematic diagram of a recognition result of a video frame shown as an exemplary embodiment of the present application;
[0022] Figure 5 A flowchart of a fourth embodiment of the item tracing method provided in this application;
[0023] Figure 6 A flowchart of Embodiment 5 of the item tracing method provided in this application;
[0024] Figure 7 A flowchart of Example 6 of the item tracing method provided in this application;
[0025] Figure 8 A hardware structure diagram of the item tracing device in which the item tracing device provided by this application is located;
[0026] Fig. 9 This is a schematic diagram of the structure of the first embodiment of the item tracing device provided by the present application;
[0027] Fig.10 This is a schematic diagram of the structure of the second embodiment of the item tracing device provided in this application. DETAILED DESCRIPTION
[0028] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.
[0029] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0030] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".
[0031] The present application provides an item tracing method, device and equipment for quickly and accurately tracing items.
[0032] The object tracing method, device and equipment provided by the present application obtains a video image in a target area, extracts at least one video frame from the video image, and then uses a visual macro model to identify each video frame in the at least one video frame to obtain the recognition result of the at least one video frame, wherein the recognition result includes the item information of each item contained in each video frame, thereby establishing an archive file of each item in the at least one video frame according to the recognition result; wherein the archive file includes a video frame containing the corresponding item. In this way, when a tracing request for tracing a target item is received, the target archive file of the target item can be obtained from the established archive file, and the target archive file can be returned. In this way, the item can be tracked and recorded based on the archive file of the item, and then when the item is lost, the specific location of the item can be found by tracing back the archive file of the item, so that the item can be traced quickly and accurately.
[0033] Specific embodiments are given below to introduce the technical solution of the present application in detail.
[0034] Figure 1 This is a flowchart of the first embodiment of the item tracing method provided by this application. Please refer to Figure 1 , the method provided in this embodiment may include:
[0035] S101: Acquire a video image within a target area, and extract at least one video frame from the video image.
[0036] The method provided in this embodiment may be applied to an item tracing device, which may be a server. The following description will be made by taking the item tracing device as a server as an example.
[0037] Specifically, the target area refers to an area to be identified, and the target area may be a bedroom, a living room, an entire house area, etc., which is not limited in this embodiment.
[0038] It should be noted that the monitoring device can collect video images of the target area and then transmit the collected video images to the server. In addition, the server can periodically obtain video images in the target area, or can obtain video images in the target area in real time, which is not limited in this embodiment. For example, in one embodiment, a video image can be obtained every 5 seconds.
[0039] Furthermore, when extracting at least one video frame from the acquired video image, in one embodiment, each video frame of the video image may be used as the extracted video frame.
[0040] In another embodiment, the video image may be subjected to frame extraction processing to obtain at least one video frame. It should be noted that the frame extraction processing refers to extracting a representative key frame from the video image.
[0041] S102. Use the visual big model to recognize each video frame in the at least one video frame to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item included in each video frame.
[0042] It should be noted that the visual big model is also called the visual foundation model, which refers to a powerful model that is pre-trained on large-scale data and can be applied to various downstream tasks. In the field of vision, it is very expensive to obtain a large amount of labeled data. The visual big model can be pre-trained by using unlabeled data, graphic data, or multimodal data, using contrastive learning, mask reconstruction and other learning methods for downstream tasks, such as object detection and semantic segmentation. For more detailed implementation principles and methods of the visual big model, please refer to the description in the relevant technology, which will not be repeated here.
[0043] In addition, the recognition result includes the item information of each item contained in each video frame. The item information may include category information and position information. It should be noted that the category information refers to the category to which each item belongs, and the position information refers to the position of the item in the video frame. Furthermore, the position information of the item can be represented by the three-dimensional coordinates of the item in the camera coordinate system, or the position information of the item can be represented by the two-dimensional coordinates of the item in the pixel coordinate system. The following is an example of "the position information of the item is represented by the three-dimensional coordinates of the item in the camera coordinate system". See Table 1, which is a recognition result of the first frame of video frame shown in an exemplary embodiment of the present application:
[0044] Table 1 Recognition results of the first video frame
[0045]
[0046]
[0047] S103. Create an archive file for each object in the at least one video frame according to the recognition result; wherein the archive file includes a video frame containing the corresponding object.
[0048] In a specific implementation, for each video frame in at least one video frame, according to the recognition result of the video frame, the video frame is stored in the archive file of each item included in the recognition result. For example, in combination with Table 1, for the first video frame, according to the recognition result of the first video frame, it is determined that the first video frame contains a mobile phone, a mouse and a charger. At this time, the first video frame is stored in the archive files of the mobile phone, the mouse and the charger respectively.
[0049] It should be noted that, in a specific implementation, when storing the video frame in the archive file of each item included in the recognition result, for each item, it is possible to first search whether the established archive file has an archive file of the item, and then, if it exists, directly store the video frame in the archive file of the item; if it does not exist, create a new archive file using the category information of the item as an identifier, and then store the video frame in the newly created archive file.
[0050] It should be noted that, in addition to the video frame containing the object, the archive file of each object may also include the recognition result of the video frame containing the object. In specific implementation, the recognition result of the video frame may be visually added to the video frame.
[0051] S104. When receiving a tracing request for tracing a target item, obtain a target archive file of the target item from the established archive files, and return the target archive file; wherein the target archive file is used to trace the target item.
[0052] Specifically, when the user cannot find the target item (for example, in one embodiment, the target item is a mobile phone), a tracing request for indicating tracing the target item may be triggered. For example, in one embodiment, the tracing request may be triggered by a client (for example, a web client, an application client, etc.). For example, the client provides a tracing button, and the user may trigger a tracing request for indicating tracing the target item by clicking the tracing button. In specific implementation, when the client detects that the tracing button is clicked, it displays a target item input interface, and the user may input the target item to be traced based on the interface. Further, after detecting that the user inputs the target item based on the interface, the client generates a tracing request for indicating tracing the target item, and sends the tracing request to the server.
[0053] Specifically, the traceability request may carry identification information of the target item and time information of the request.
[0054] Correspondingly, after receiving the tracing request, the server can search for the target archive file of the target item in the established archive file according to the identification information of the target item carried in the tracing request, and return the target archive file to the client, so that the user on the client side can trace the target item based on the target archive file.
[0055] For example, in combination with the above example, after receiving the tracing request triggered by the user to instruct tracing the mobile phone, the server searches for the target archive file corresponding to the mobile phone from the established archive file and returns the target archive file to the client. In this way, the user on the client side can trace the specific location of the mobile phone based on the multiple video frames containing the mobile phone in the target archive file.
[0056] The item tracing method provided in this embodiment obtains a video image in a target area, extracts at least one video frame from the video image, and then uses a visual macro model to identify each video frame in the at least one video frame to obtain the identification result of the at least one video frame, wherein the identification result includes the item information of each item contained in each video frame, thereby establishing an archive file of each item in the at least one video frame according to the identification result; wherein the archive file includes a video frame containing the corresponding item. In this way, when a tracing request for tracing a target item is received, the target archive file of the target item can be obtained from the established archive file, and the target archive file can be returned. In this way, the item can be tracked and recorded based on the archive file of the item, and then when the item is lost, the specific location of the item can be found by tracing back the archive file of the item, so that the item can be traced quickly and accurately.
[0057] Figure 2 This is a flow chart of Example 2 of the item tracing method provided by this application. Please refer to Figure 2 In the method provided in this embodiment, based on the above embodiment, the object information includes category information and location information; after obtaining the recognition result of the at least one video frame, the method further includes:
[0058] S201. According to the object information of each object contained in each video frame, obtain a labeled object image with labeling information corresponding to each object.
[0059] For example, in one embodiment, for a video frame, based on the recognition information of the video frame, for each object included in the recognition result, the video frame can be labeled with the category information of the object to obtain a labeled object image with the labeled information.
[0060] S202: Perform intensive training on the visual large model using the labeled object image.
[0061] Specifically, when a labeled object image is obtained, the labeled object image can be used as a sample to perform intensive training on the visual large model to further improve the recognition accuracy of the visual large model.
[0062] The method provided in this embodiment obtains the labeled object images with labeled information corresponding to each object according to the object information of each object contained in each video frame, and uses the labeled object images to perform intensive training on the visual macro model. In this way, the visual macro model can be further trained to improve the accuracy of the visual macro model, thereby quickly and accurately tracing objects.
[0063] Figure 3 This is a flowchart of Example 3 of the item tracing method provided by this application. Please refer to Figure 3 Based on the above embodiment, the method provided in this embodiment is mainly used to explain how to obtain the labeled object images with labeled information corresponding to each object, and use the labeled object images to perform enhanced training on the visual large model. Figure 3 The method provided in this embodiment, wherein the step of obtaining the labeled object images with the labeled information corresponding to the respective objects and using the labeled object images to perform intensive training on the visual macro model comprises:
[0064] S301. According to the position information of each object contained in each video frame, capture the object image of each object from each video frame.
[0065] Specifically, Figure 4 This is a schematic diagram of a video frame recognition result shown in an exemplary embodiment of the present application. Figure 4 ,exist Figure 4 In the example shown, the recognition result of a video frame is visually displayed on the video frame, and the location information of each object contained in the video frame is represented by a detection box.
[0066] In a specific implementation, for each object included in the video frame, when capturing the object image of the object from the video frame, the capture can be performed directly according to the detection frame, and the area where the detection frame is located can be captured as the object image of the object.
[0067] For example, combined with Figure 4 For the convenience of explanation, the detection frame corresponding to the mobile phone is recorded as detection frame 1. At this time, for the mobile phone, the area where the detection frame 1 is located can be directly captured as the object image of the mobile phone.
[0068] S302: for the object image of each of the objects, label the object image according to the category information of each object to obtain a labeled object image with labeling information.
[0069] It should be noted that labeling an object image refers to the process of adding a label to the object image. In this step, for each object image of an object, the category information of the object is used as the label of the object image.
[0070] S303: Save the annotated object image into a corresponding classification component; wherein the annotated object image included in each classification component has the same annotated information.
[0071] Specifically, after the object images are annotated, the annotated object images with the same annotation information are saved in the category file corresponding to the annotation information. It should be noted that the category file corresponding to the annotation information refers to a file that stores a number of annotated object images with category information corresponding to the annotation information. For example, in one embodiment, the category file corresponding to the annotation information may be a file with the annotation information as the file name.
[0072] For example, in one embodiment, an annotated object image with the annotation information of a mobile phone is saved in a classification file corresponding to the mobile phone, and the classification file stores annotated object images of mobile phones.
[0073] S304: When there is a target classification component in the classification component, use the labeled object images in the target classification component to perform intensive training on the visual large model; wherein the number of labeled object images stored in the target classification component is greater than a preset threshold.
[0074] Specifically, the target classification file refers to a classification file in which the number of labeled object images stored is greater than a preset threshold.
[0075] It should be noted that the preset threshold is set according to actual needs, and in this embodiment, the specific value of the preset threshold is not limited. For example, in one embodiment, the preset threshold may be 10.
[0076] Further, in combination with the above example, for example, when the number of labeled object images saved in the classification file corresponding to the mobile phone reaches 10, at this time, the labeled object images in the classification file corresponding to the mobile phone are used to perform intensive training on the visual large model. After the visual large model is intensively trained using the labeled object images in the classification file corresponding to the mobile phone, the visual large model will have a better recognition accuracy for the mobile phone.
[0077] The object tracing method provided in this embodiment intercepts the object image of each object from each video frame according to the location information of each object contained in each video frame, and then annotates the object image of each object in the object according to the category information of each object to obtain an annotated object image with annotated information, thereby saving the annotated object image into the corresponding classification component; wherein the annotated object images included in each classification component have the same annotated information, so that when there is a target classification component in the classification component, the annotated object images in the target classification component can be used to perform intensive training on the visual large model; wherein the number of annotated object images saved in the target classification component is greater than a preset threshold. In this way, the visual large model can have a higher recognition accuracy rate for objects in the target area, and the accuracy of object tracing can be further improved.
[0078] Figure 5 This is a flowchart of the fourth embodiment of the item tracing method provided by this application. Please refer to Figure 5 The method provided in this embodiment, based on the above embodiment, wherein returning the target archive file comprises:
[0079] S501. Select a specified number of target video frames from the target archive file in order of shooting time from the back to the front.
[0080] Specifically, the designated number is set according to actual needs, and in this embodiment, the specific value of the designated number is not limited. For example, in one embodiment, the designated number is 5.
[0081] Combined with the above example, when the user on the client side traces the mobile phone, the server selects 5 target video frames from the archive file of the mobile phone in the order of shooting time from the back to the front. For example, in one embodiment, the archive file of the mobile phone includes 10 video frames, and the shooting times of these 10 video frames are: video frame 1: 8:00, video frame 2: 8:02, video frame 3: 8:05, video frame 4: 8:10, video frame 5: 8:15, video frame 6: 8:18, video frame 7: 8.20, video frame 8: 8:25, video frame 9: 8:30, video frame 10: 8.33. In this step, the selected target video frames are video frame 10, video frame 9, video frame 8, video frame 7, and video frame 6.
[0082] S502: Return the target video frame.
[0083] Specifically, in one embodiment, the recognition result of the target video frame can also be returned to the client. In specific implementation, for example, the recognition result of the target video frame can be added to the target video frame, and the target video frame and the recognition result of the target video frame can be returned to the client together.
[0084] The object tracing method provided in this embodiment selects a specified number of target video frames from the target archive file in the order of shooting time from the last to the first, and then returns the target video frames to the client. In this way, the target object can be traced with a small number of video frames closest to the disappearance time of the target object, thereby improving the efficiency of object tracing.
[0085] Figure 6 This is a flowchart of Example 5 of the item tracing method provided by this application. Please refer to Figure 6 The method provided in this embodiment, based on the above embodiment, saves the labeled object image to the corresponding classification component, and the method further includes:
[0086] S601: In response to a display operation on a designated classification file, each labeled object image stored in the designated classification file is displayed to a user.
[0087] Specifically, when a user on the server side (for example, an administrator) wants to view the labeled object images in a specified classification file, a display operation for the specified classification file may be triggered.
[0088] Furthermore, in response to the user's display operation on the designated classification file, the server displays all the labeled object images stored in the designated classification file to the user.
[0089] S602. In response to a deletion operation on a target labeled object image in each displayed labeled object image, the target labeled object image is deleted; wherein the target labeled object image is a labeled object image whose real category in the specified classification file is not the labeling information corresponding to the specified classification file.
[0090] Specifically, after seeing all the labeled object images displayed, the user on the server side will check whether the real category of these labeled object images is the labeling information corresponding to the specified classification file, and then trigger the deletion operation for the target labeled object image when it is found that the real category of the target labeled object image does not match the labeling information corresponding to the classification file where the target labeled object image is located.
[0091] Furthermore, after detecting a deletion operation on the target labeled object image, the server deletes the target labeled object image from the designated classification file in response to the deletion operation.
[0092] For example, a user on the server side wants to view the classification file corresponding to the remote control, triggering a corresponding display operation. In response to the display operation, the server displays all the labeled object images in the classification file corresponding to the remote control to the user. For example, if 10 pictures are saved in the classification file, these 10 pictures are displayed to the user.
[0093] Furthermore, after viewing these 10 pictures, the user discovers that the true category of the 5th picture is not the remote control. At this time, the user triggers a deletion operation for the 5th picture. Accordingly, the server responds to the deletion operation and deletes the 5th picture from the classification file corresponding to the remote control.
[0094] The item tracing method provided in this embodiment displays each annotated item image stored in the specified classification file to the user in response to a display operation for the specified classification file, and then deletes the target annotated item image in response to a deletion operation for each displayed annotated item image; wherein the target annotated item image is an annotated item image whose real category in the specified classification file is not the annotated information corresponding to the specified classification file. In this way, the annotated item images that do not match the actual situation can be artificially eliminated, thereby improving the efficiency of model reinforcement training and the accuracy of model recognition.
[0095] Figure 7 This is a flowchart of Example 6 of the item tracing method provided by this application. Please refer to Figure 7 , the method provided in this embodiment includes:
[0096] S701: Acquire a video image within a target area, and extract at least one video frame from the video image.
[0097] S702. Use the visual big model to identify each video frame in the at least one video frame to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item contained in each video frame.
[0098] S703: According to the position information of each object contained in each video frame, capture the object image of each object from each video frame.
[0099] S704: for the object image of each of the objects, label the object image according to the category information of each object to obtain a labeled object image with labeling information.
[0100] S705: Save the annotated object image into a corresponding classification component; wherein the annotated object image included in each classification component has the same annotated information.
[0101] S706. When there is a target classification component in the classification component, use the labeled object images in the target classification component to perform intensive training on the visual large model; wherein the number of labeled object images stored in the target classification component is greater than a preset threshold.
[0102] S707. Create an archive file for each object in the at least one video frame according to the recognition result; wherein the archive file includes the video frame containing the corresponding object.
[0103] Specifically, in this embodiment, the execution order of steps S703 and S707 is not limited, and steps S703 and S707 may be executed simultaneously.
[0104] S708. When receiving a tracing request for tracing a target item, obtain a target archive file of the target item from the established archive file, and return the target archive file; wherein the target archive file is used to trace the target item.
[0105] The specific implementation principles and implementation processes of steps S701 to S708 can be referred to the description in the previous embodiment and will not be repeated here.
[0106] The method provided in this embodiment can track and record items based on the archive files of the items, and then when the items are lost, the specific location of the items can be found by tracing back the archive files of the items, so that the items can be traced quickly and accurately. In addition, the method provided in this embodiment can make the visual large model have a higher recognition accuracy for items in the target area by strengthening the training of the visual large model, which can further improve the accuracy of item tracing.
[0107] Corresponding to the above-mentioned embodiment of the object tracing method, the present application also provides an embodiment of an object tracing device.
[0108] The embodiment of the item tracing device provided in this application can be applied to item tracing equipment. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of the item tracing device in which it is located reading the corresponding computer program instructions in the non-volatile memory into the memory and running them. From the hardware level, if Figure 8 As shown, it is a hardware structure diagram of the item tracing device provided by the present application where the item tracing device is located. Figure 8 In addition to the processor, memory, network interface, and non-volatile memory shown, the item tracing device in which the device is located in the embodiment may also include other hardware according to the actual function of the item tracing device, which will not be described in detail.
[0109] Fig. 9 This is a schematic diagram of the structure of the first embodiment of the item tracing device provided by this application. Fig. 9 The device provided in this embodiment may include an acquisition module 910, an identification module 920, a file creation module 930 and a processing module 940, wherein:
[0110] The acquisition module 910 is used to acquire a video image in a target area and extract at least one video frame from the video image.
[0111] The recognition module 920 is used to use the visual big model to recognize each video frame in the at least one video frame to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item contained in each video frame.
[0112] The archiving module 930 is used to create an archive file for each object in the at least one video frame according to the recognition result; wherein the archive file includes a video frame containing the corresponding object.
[0113] The processing module 940 is used to obtain a target archive file of the target item from the established archive files and return the target archive file when receiving a tracing request for tracing the target item; wherein the target archive file is used to trace the target item.
[0114] The device provided in this embodiment can be used to perform Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects, which will not be repeated here.
[0115] The item tracing device provided in this embodiment obtains a video image in a target area, extracts at least one video frame from the video image, and then uses a visual macro model to identify each video frame in the at least one video frame to obtain the identification result of the at least one video frame, wherein the identification result includes the item information of each item contained in each video frame, thereby establishing an archive file of each item in the at least one video frame according to the identification result; wherein the archive file includes a video frame containing the corresponding item. In this way, when a tracing request for tracing a target item is received, the target archive file of the target item can be obtained from the established archive file, and the target archive file can be returned. In this way, the item can be tracked and recorded based on the archive file of the item, and then when the item is lost, the specific location of the item can be found by tracing back the archive file of the item, so that the item can be traced quickly and accurately.
[0116] Fig.10 This is a schematic diagram of the structure of the second embodiment of the item tracing device provided by this application. Fig.10 The device provided in this embodiment, based on the above embodiment, further includes a training module 950, wherein:
[0117] The training module 950 is used to obtain the labeled object images with labeling information corresponding to each object according to the object information of each object contained in each video frame, and use the labeled object images to perform intensive training on the visual large model.
[0118] Optionally, the training module 950 is specifically used to capture the object image of each object from each video frame according to the position information of each object contained in each video frame, and to label the object image of each object in the each object according to the category information of each object to obtain a labeled object image with labeling information.
[0119] Optionally, the training module 950 is specifically used to save the labeled object images into corresponding classification components, and when there is a target classification component in the classification component, use the labeled object images in the target classification component to perform intensive training on the visual large model; wherein the labeling information of the labeled object images included in each of the classification components is the same, and the number of labeled object images saved in the target classification component is greater than a preset threshold.
[0120] Optionally, the processing module 940 is further used to select a specified number of target video frames from the target archive file in a sequence from the back to the front of the shooting time, and return the target video frames.
[0121] Optionally, the acquisition module 910 is specifically configured to perform frame extraction processing on the video image to obtain the at least one video frame.
[0122] Optionally, the processing module 940 is further used to display each labeled object image saved in the specified classification file to the user in response to a display operation on the specified classification file, and to delete the target labeled object image in response to a deletion operation on the target labeled object image in each displayed labeled object image; wherein the target labeled object image is a labeled object image whose real category in the specified classification file is not the labeling information corresponding to the specified classification file.
[0123] Please continue to refer to Figure 8 The present application also provides an item tracing device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any one of the methods provided in the first aspect of the present application are implemented.
[0124] The present application also provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the steps of any one of the methods provided in the first aspect of the present application are implemented.
[0125] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.
[0126] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application scheme. A person of ordinary skill in the art can understand and implement it without paying any creative work.
[0127] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.
Claims
1. A method for tracing an item, characterized in that: The method comprises: Acquire a video image within the target area, and extract at least one video frame from the video image; Using the visual macro model to recognize each of the at least one video frame, to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item contained in each video frame; According to the recognition result, establishing an archive file for each object in the at least one video frame; wherein the archive file includes the video frame containing the corresponding object; When a tracing request for tracing a target item is received, a target archive file of the target item is obtained from the established archive files, and the target archive file is returned; wherein the target archive file is used to trace the target item.
2. The method according to claim 1, characterized in that: After obtaining the recognition result of the at least one video frame, the method further includes: According to the item information of each item contained in each video frame, obtaining an annotated item image with annotated information corresponding to each item; The labeled object images are used to perform intensive training on the visual large model.
3. The method according to claim 2, characterized in that The item information includes category information and location information; the step of obtaining the labeled item images with labeled information corresponding to each item includes: According to the position information of each object contained in each video frame, intercepting the object image of each object from each video frame; For the object image of each of the objects, the object image is labeled according to the category information of each object to obtain a labeled object image with the labeled information.
4. The method according to claim 2 or 3, characterized in that: The step of using the standard object image to perform enhanced training on the visual large model includes: Saving the annotated object images into corresponding classification components; wherein the annotated object images included in each classification component have the same annotated information; In the case where there is a target classification component in the classification component, the labeled object images in the target classification component are used to perform intensive training on the visual large model; wherein the number of labeled object images stored in the target classification component is greater than a preset threshold.
5. The method according to claim 4, characterized in that The method of saving the labeled object image to the corresponding classification component also includes: In response to a display operation on a designated classification file, display each labeled object image stored in the designated classification file to a user; In response to a deletion operation on a target labeled object image in each displayed labeled object image, the target object image is deleted; wherein the target labeled object image is a labeled object image whose true category in the specified classification file is not the labeling information corresponding to the specified classification file.
6. The method according to any one of claims 1 to 5, characterized in that: The returning the target archive file comprises: Selecting a specified number of target video frames from the target archive file in a sequence from the last to the first of the shooting time; Returns the target video frame.
7. The method according to claim 1, characterized in that The extracting at least one video frame from the video image comprises: Perform frame extraction processing on the video image to obtain the at least one video frame.
8. An item tracing device, characterized in that: The device includes an acquisition module, an identification module, a file creation module and a processing module; wherein, The acquisition module is used to acquire a video image within a target area and extract at least one video frame from the video image; The recognition module is used to recognize each of the at least one video frame using the visual macro model to obtain a recognition result of the at least one video frame, wherein the recognition result includes item information of each item contained in each video frame; The archiving module is used to create an archive file for each object in the at least one video frame according to the recognition result; wherein the archive file includes the video frame containing the corresponding object; The processing module is used to obtain a target archive file of the target item from the established archive files and return the target archive file when receiving a tracing request for tracing the target item; wherein the target archive file is used to trace the target item.
9. An item tracing device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.