Automatic identification method and device
By calculating the matching degree between the object to be identified and the recognition model and selecting the most suitable model for identification, the problems of low recognition accuracy and efficiency in the prior art are solved, and efficient identification of different categories of objects is achieved.
Patent Information
- Application Number
- CN202411049278.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-26
AI Technical Summary
Existing automatic identification technology cannot effectively identify different categories of objects to be identified, resulting in low recognition accuracy and efficiency.
By receiving the attribute feature data of the object to be identified and the feature data of the pre-constructed recognition model, calculate the matching degree, and select the recognition model with the greatest matching degree for identification, determine whether the recognition result meets the preset conditions, execute the corresponding evaluation strategy or place it in Observe the database.
The accuracy and efficiency of identification of objects to be identified in different categories is improved, and the accuracy and processing efficiency of identification results are improved.
Smart Images

Figure CN119027888A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to an automatic identification method and device thereof. Background Art
[0002] Automatic recognition technology is a technology that uses a certain recognition device to automatically obtain relevant information of the identified object through the proximity between the identified object and the recognition device, and provides the information to the background computer processing system to complete the relevant subsequent processing.
[0003] Automatic identification technology integrates computer, optical, electrical, communication and network technologies, and combines them with the Internet, mobile communications and other technologies to achieve global tracking of objects and information sharing, thereby giving objects intelligence and enabling communication and dialogue between people and objects, and between objects. Automatic identification technology can identify and recognize each object, and can update data in real time. It is an important part of building real-time sharing of global object information. Automatic recognition can automatically collect data, automatically identify information, and automatically input it into the computer, allowing humans to process large amounts of data and information in a timely and accurate manner.
[0004] The existing patent with publication number CN116092198A discloses a method for identifying and detecting mining helmets, which includes dynamically and in real time collecting images of on-site workers; identifying whether the on-site workers are wearing helmets based on the images; if they are not wearing helmets, continuing to identify whether they are wearing helmets for the next frame of the image within a preset time period; if they are not wearing helmets, sending an alarm prompt. However, the automatic identification method of the helmet can only automatically identify images, and cannot be used for the identification of multiple different categories of objects to be identified, and the identification accuracy and efficiency of the objects to be identified are both low.
[0005] Therefore, the technical problem that urgently needs to be solved is: how to realize the automatic recognition of different contents of interest in different categories of objects to be identified, and improve the recognition accuracy and efficiency of the objects to be identified. Summary of the invention
[0006] The purpose of the present application is to provide an automatic recognition method and device thereof, so as to realize automatic recognition of different contents of interest in different categories of objects to be recognized, and to improve the recognition accuracy and efficiency of the objects to be recognized.
[0007] To achieve the above-mentioned purpose, the present application provides an automatic recognition method, which includes the following steps: receiving an object to be recognized and content of interest to be recognized; collecting first attribute feature data of the object to be recognized and second attribute feature data of multiple pre-built recognition models, and obtaining recognition verification evaluation data of the recognition model; calculating the matching degree between the object to be recognized and each pre-built recognition model based on the first attribute feature data of the object to be recognized and the second attribute feature data of multiple pre-built recognition models, as well as the recognition verification evaluation data of the recognition model; selecting the recognition module with the largest matching degree to recognize the object to be recognized and obtain the recognition result.
[0008] The automatic identification method as described above, wherein the method further includes the following steps: determining whether the identification result meets the preset review conditions corresponding to the interest content to be identified, if so, executing the corresponding evaluation strategy, otherwise, placing the object to be identified into the observation database.
[0009] The automatic recognition method as described above, wherein the method for pre-building a recognition model includes the following steps: obtaining a training sample data set and a verification sample data set; inputting the training sample data set into a neural network basic learning model for training to obtain a trained recognition model; inputting the verification sample data set into the trained recognition model for recognition to obtain a recognition result; calculating the recognition accuracy of the recognition model based on the recognition result; comparing the recognition accuracy of the recognition model with a preset required threshold, if the recognition accuracy of the recognition model is less than the preset required threshold, optimizing the recognition model, otherwise, there is no need to optimize the recognition model.
[0010] In the automatic recognition method as described above, the recognition module recognizes the object to be recognized, and the obtained recognition result is the data of interest in the object to be recognized.
[0011] In the automatic identification method as described above, different priorities are set in advance for different categories of the data of interest; the server preferentially processes the data of interest with a higher priority level according to the priorities set in advance for the data of interest.
[0012] In the automatic recognition method as described above, the evaluation strategy includes calculating the degree of congestion of vehicles on a certain road based on the results of recognition by the recognition model.
[0013] In the automatic identification method as described above, the calculation formula for the degree of congestion of vehicles on a certain road is as follows:
[0014]
[0015] Among them, δ≠θ, a≠b;
[0016] Among them, YJ represents the degree of congestion of vehicles on a certain road; r1 represents the weight influencing factor of the number of vehicles on a certain road; Fe represents the width of a certain road; Kn represents the number of vehicles counted on a certain road after the video sampling ends; δ and θ are parameters; r 2 Represents the weighted influencing factor of the distance between vehicles on a certain road; r 1 +r 2 =1; LB indicates the number of vehicles traveling in parallel; μ 1 Indicates the influence weight of the front and rear distance between vehicles; μ 2 Indicates the influence weight of the left and right distance between vehicles; HX ab Indicates the lateral distance between the ath vehicle and the bth vehicle traveling in parallel, that is, the left-right distance; QH δθ represents the distance between the δth vehicle and the θth vehicle.
[0017] In the automatic identification method as described above, the first attribute feature data of the object to be identified includes: the type of the object to be identified, the identification type of the interest content to be identified and the size of the object to be identified.
[0018] As a second aspect of the present application, the present application also provides an automatic identification device, which includes: a data receiving module, used to receive an object to be identified and an interest content to be identified; a data acquisition module, used to collect first attribute feature data of the object to be identified and second attribute feature data of multiple pre-built recognition models, and obtain recognition verification evaluation data of the recognition model; a data processor, based on the first attribute feature data of the object to be identified and the second attribute feature data of multiple pre-built recognition models, as well as the recognition verification evaluation data of the recognition model, calculates the matching degree between the object to be identified and each pre-built recognition model; a recognition module, used to identify the object to be identified and obtain a recognition result.
[0019] The automatic identification device as described above, wherein the device also includes: a judgment module, used to judge whether the identification result meets the preset review conditions corresponding to the interest content to be identified, if so, execute the corresponding evaluation strategy, otherwise, put the object to be identified into the observation database.
[0020] The beneficial effects achieved by this application are as follows:
[0021] (1) The present application calculates the matching degree between the object to be identified and each pre-built recognition model based on the first attribute feature data of the object to be identified and the second attribute feature data of multiple pre-built recognition models, as well as the recognition verification evaluation data of the recognition model; selects the recognition module with the largest matching degree to identify the object to be identified, and obtains the recognition result, thereby improving the recognition accuracy and recognition efficiency of the content of interest to be identified of the object to be identified.
[0022] (2) The present application determines whether the recognition result meets the preset review conditions corresponding to the content of interest to be recognized. If so, the corresponding evaluation strategy is executed. Otherwise, the object to be recognized is placed in the observation database, thereby improving the processing efficiency of the content of interest to be recognized, automatically executing the corresponding evaluation strategy, and improving the evaluation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For those skilled in the art, other drawings can also be obtained based on these drawings.
[0024] Figure 1 The present invention is a flowchart of an automatic identification method according to an embodiment of the present application.
[0025] Figure 2 A flow chart of a method for pre-building a recognition model in an embodiment of the present application.
[0026] Figure 3 A schematic diagram of the structure of an automatic identification device according to an embodiment of the present application.
[0027] Figure numerals: 10 - data receiving module; 20 - data acquisition module; 30 - data processor; 40 - identification module; 50 - judgment module; 100 - automatic identification device. DETAILED DESCRIPTION
[0028] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.
[0029] Embodiment 1
[0030] like Figure 1 As shown, the present application provides an automatic identification method, which includes the following steps:
[0031] Step S1, receiving an object to be identified and interesting content to be identified.
[0032] Specifically, the objects to be identified include pictures, videos or texts, etc. The types of the content of interest to be identified include digital content, text content, graphic content and attention targets, etc. For example, the text content in the picture, the type of graphics contained in the picture, the type of target contained in the video (such as people, vehicles, animals, trees, etc.), and the keywords in the text, etc.
[0033] Step S2, collecting first attribute feature data of the object to be identified and second attribute feature data of multiple pre-built recognition models, and obtaining recognition verification evaluation data of the recognition models.
[0034] Specifically, the first attribute feature data of the object to be identified includes: the type of the object to be identified, the identification type of the interest content to be identified, and the size of the object to be identified.
[0035] The types of objects to be identified include pictures, videos or texts.
[0036] Specifically, the second attribute feature data of the pre-built recognition model includes: the object type recognized by the recognition model, the content type recognized by the recognition object, the object size range recognized by the recognition object, and the like.
[0037] Among them, the content types recognized by the recognition object include recognition of digital content, recognition of text content, recognition of graphic content and recognition of attention targets, etc.
[0038] The recognition verification evaluation data of the recognition model is obtained based on the recognition results of the recognition model on the object to be recognized in the verification sample data set. The recognition verification evaluation data of the recognition model includes: the time it takes for the recognition model to recognize the object to be recognized, the accuracy of the recognition model's recognition results, and the response time slot of the recognition model. The response time slot is the response recognition time length of the recognition model to the next input object to be recognized after it has recognized the current object to be recognized.
[0039] like Figure 2 As shown, the method for pre-building a recognition model includes the following steps:
[0040] Step S210, obtaining a training sample data set and a verification sample data set.
[0041] The training sample data set includes objects to be identified and interesting contents to be identified.
[0042] Step S220, input the training sample data set into the neural network basic learning model for training to obtain a trained recognition model.
[0043] Step S230: input the verification sample data set into the trained recognition model for recognition to obtain a recognition result.
[0044] Step S240, calculating the recognition accuracy of the recognition model according to the recognition result.
[0045] Among them, according to whether the recognition result of the object to be recognized in the verification sample data set is accurate, the number of accurate recognitions of the object to be recognized is obtained, and according to the number of accurate recognitions of the object to be recognized, the recognition accuracy rate of the recognition model is calculated.
[0046] Specifically, the recognition accuracy of the recognition model is equal to the number of objects to be recognized accurately divided by the total number of objects to be recognized in the verification sample data set.
[0047] Step S250, comparing the recognition accuracy of the recognition model with a preset threshold value, if the recognition accuracy of the recognition model is less than the preset threshold value, the recognition model is optimized, otherwise, there is no need to optimize the recognition model.
[0048] In the present application, when the recognition accuracy of the recognition model is less than a preset required threshold, the recognition model is optimized to improve the recognition accuracy of the recognition model.
[0049] Step S3, calculating the matching degree between the object to be identified and each pre-built recognition model according to the first attribute feature data of the object to be identified and the second attribute feature data of multiple pre-built recognition models, as well as the recognition verification evaluation data of the recognition model.
[0050] Specifically, the calculation formula for the matching degree between the object to be identified and each pre-built identification model is as follows:
[0051]
[0052] Where P represents the matching degree between the object to be identified and each pre-built identification model; H represents the total number of types of the first attribute feature data of the object to be identified; Rz h The factor representing the match between the hth first attribute characteristic data and the corresponding second attribute characteristic data; if the hth first attribute characteristic data matches the corresponding second attribute characteristic data, then Rz h =1, otherwise, Rz h =0;λ 1 represents the influence weight of the time required to obtain the recognition result of the object to be recognized; N represents the total number of objects to be recognized in the verification sample data set; Ts i Indicates the time it takes for the recognition model to identify the i-th object to be identified; Tx i represents the response time slot after the recognition model recognizes the i-th object to be recognized; Ty represents the time required by the object to be recognized to obtain the recognition result; λ 2 It represents the influence weight of the recognition accuracy rate of the recognition model required by the object to be recognized; Nd represents the number of accurate recognition results of the recognition model for the objects to be recognized in the verification sample data set; Ys represents the recognition accuracy rate value of the recognition model required by the object to be recognized.
[0053] It should be explained that if the type of the object to be identified is consistent with the type of the object identified by the recognition model, then Rz h =1, otherwise, Rz h= 0; if the identification type of the interest content to be identified is consistent with the type identified by the recognition model, then Rz h =1, otherwise, Rz h = 0; if the size of the object to be identified is within the size range of the object identified by the recognition model, then Rz h =1, otherwise, Rz h = 0. The present application improves the calculation accuracy of the matching degree between the object to be identified and each pre-built identification model.
[0054] Step S4, selecting the recognition module with the largest matching degree to recognize the object to be recognized and obtaining the recognition result.
[0055] As a specific embodiment of the present invention, if the object to be identified is a video, the video is identified by using a pre-built recognition model to obtain the target in the video.
[0056] As an embodiment of the present invention, if the interest content to be identified is to identify whether there are abnormal keywords in the object to be identified, if so, the recognition result indicates that there are abnormal phenomena in the object to be identified, otherwise, the recognition result indicates that there are no abnormal phenomena in the object to be identified.
[0057] As a specific embodiment of the present invention, the recognition module recognizes the object to be recognized, and the recognition result obtained is the data of interest in the object to be recognized. The data of interest are, for example, buildings, people, vehicles, etc. in the video. Preferably, different priorities are set in advance for different categories of the data of interest. The server processes the data of interest with a higher priority level according to the priority set in advance for the data of interest, for example, audits and evaluates the data of interest according to the corresponding evaluation strategy.
[0058] Step S5, determining whether the recognition result meets the preset review condition corresponding to the interest content to be recognized, if so, executing the corresponding evaluation strategy, otherwise, putting the object to be recognized into the observation database.
[0059] Specifically, the preset audit conditions are different according to the different interest contents to be identified. For example, if the interest content to be identified is to identify abnormal text, the corresponding preset audit condition is: if there are more than 3 abnormal texts, then execute the evaluation strategy, and the corresponding evaluation strategy is, for example, to evaluate the abnormality of the object to be identified. If the interest content to be identified is: to identify the number of vehicles in the object to be identified, if the number of vehicles exceeds 50, then execute the evaluation strategy, and the corresponding evaluation strategy is, for example, to evaluate the congestion of vehicles on a certain road.
[0060] Specifically, the object to be identified is put into an observation database so as to be observed in the observation database for future use.
[0061] As a specific embodiment of the present invention, the evaluation strategy includes calculating the degree of congestion of vehicles on a certain road according to the result of recognition by the recognition model.
[0062] Specifically, the calculation formula for the degree of congestion of vehicles on a certain road is as follows:
[0063]
[0064] Among them, δ≠θ, a≠b;
[0065] Among them, YJ represents the degree of congestion of vehicles on a certain road; r 1 represents the weight influencing factor of the number of vehicles on a certain road; Fe represents the width of a certain road; Kn represents the number of vehicles counted on a certain road after the video sampling ends; δ and θ are parameters; r 2 Represents the weighted influencing factor of the distance between vehicles on a certain road; r 1 +r 2 =1; LB indicates the number of vehicles traveling in parallel; μ 1 Indicates the influence weight of the front and rear distance between vehicles; μ 2 Indicates the influence weight of the left and right distance between vehicles; HX ab Indicates the lateral distance between the ath vehicle and the bth vehicle traveling in parallel, that is, the left-right distance; QH δθ represents the distance between the δth vehicle and the θth vehicle.
[0066] It needs to be explained that by setting up a video monitor on the road to collect surveillance videos on the road, the surveillance videos are input as objects to be identified into the recognition model with the largest matching degree for identification, the vehicle targets in the surveillance videos are obtained, the number of vehicles in the surveillance videos is counted, and the recognition model identifies the distance between the vehicles in the surveillance videos, and outputs the lateral spacing distance or the front and rear distance between the vehicles. How the recognition model identifies the lateral spacing distance or the front and rear distance between the vehicle targets and the vehicles in the surveillance videos is achieved through pre-training of the recognition model, so that the recognition model can identify the lateral spacing distance or the front and rear distance between the vehicle targets and the vehicles in the surveillance videos. The present application improves the calculation accuracy of the degree of congestion of vehicles on a certain road. The training method of the recognition model can also adopt the existing training method, as long as the recognition model can recognize the lateral spacing distance or the front and rear distance between the vehicle targets and the vehicles in the surveillance video, which will not be repeated here.
[0067] The present invention may also use existing technologies to evaluate the degree of congestion of vehicles on a road and the degree of abnormality of the object to be identified, which will not be described in detail here.
[0068] Embodiment 2
[0069] like Figure 3 As shown, the present application provides an automatic identification device 100, which includes:
[0070] The data receiving module 10 is used to receive the object to be identified and the interest content to be identified.
[0071] The data collection module 20 is used to collect the first attribute feature data of the object to be identified and the second attribute feature data of multiple pre-built recognition models, and obtain the recognition verification evaluation data of the recognition models.
[0072] The data processor 30 calculates the matching degree between the object to be identified and each pre-built recognition model based on the first attribute feature data of the object to be identified, the second attribute feature data of multiple pre-built recognition models, and the recognition verification evaluation data of the recognition models.
[0073] The recognition module 40 is used to recognize the object to be recognized and obtain a recognition result.
[0074] The judgment module 50 is used to judge whether the recognition result meets the preset review conditions corresponding to the interest content to be recognized. If so, the corresponding evaluation strategy is executed; otherwise, the object to be recognized is put into the observation database.
[0075] Specifically, the calculation formula for the matching degree between the object to be identified and each pre-built identification model is as follows:
[0076]
[0077] Where P represents the matching degree between the object to be identified and each pre-built identification model; H represents the total number of types of the first attribute feature data of the object to be identified; Rz h The factor representing the match between the hth first attribute characteristic data and the corresponding second attribute characteristic data; if the hth first attribute characteristic data matches the corresponding second attribute characteristic data, then Rz h =1, otherwise, Rz h =0;λ 1 represents the influence weight of the time required to obtain the recognition result of the object to be recognized; N represents the total number of objects to be recognized in the verification sample data set; Ts i Indicates the time it takes for the recognition model to identify the i-th object to be identified; Tx i represents the response time slot after the recognition model recognizes the i-th object to be recognized; Ty represents the time required by the object to be recognized to obtain the recognition result; λ 2 It represents the influence weight of the recognition accuracy rate of the recognition model required by the object to be recognized; Nd represents the number of accurate recognition results of the recognition model for the objects to be recognized in the verification sample data set; Ys represents the recognition accuracy rate value of the recognition model required by the object to be recognized.
[0078] It is understandable that in order to realize the above functions, hardware structures and / or software modules corresponding to the execution of each function are included. Those skilled in the art should easily realize that, in combination with the modules and algorithm steps of each example described in the embodiments disclosed in this application, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0079] An embodiment of the present invention provides a processor for processing the above-mentioned automatic identification method.
[0080] In the embodiment of the present invention, the processor may be an integrated circuit chip having the ability to process signals. The processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0081] The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor for execution, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The processor reads the information in the storage medium and completes the steps of the above method in combination with its hardware.
[0082] The storage medium may be a memory, which may be, for example, a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memory.
[0083] Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct RAM bus random access memory (DRRAM).
[0084] The beneficial effects achieved by this application are as follows:
[0085] (1) The present application calculates the matching degree between the object to be identified and each pre-built recognition model based on the first attribute feature data of the object to be identified and the second attribute feature data of multiple pre-built recognition models, as well as the recognition verification evaluation data of the recognition model; selects the recognition module with the largest matching degree to identify the object to be identified, and obtains the recognition result, thereby improving the recognition accuracy and recognition efficiency of the content of interest to be identified of the object to be identified.
[0086] (2) The present application determines whether the recognition result meets the preset review conditions corresponding to the content of interest to be recognized. If so, the corresponding evaluation strategy is executed. Otherwise, the object to be recognized is placed in the observation database, thereby improving the processing efficiency of the content of interest to be recognized, automatically executing the corresponding evaluation strategy, and improving the evaluation efficiency.
[0087] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.
[0088] In the description of the present application, the word "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.
[0089] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims
1. An automatic identification method, It is characterized in that The method comprises the following steps: Receiving an object to be identified and content of interest to be identified; Collecting first attribute feature data of the object to be identified and second attribute feature data of a plurality of pre-built recognition models, and obtaining recognition verification evaluation data of the recognition models; Calculating the matching degree between the object to be identified and each pre-built recognition model according to the first attribute feature data of the object to be identified and the second attribute feature data of the pre-built multiple recognition models, as well as the recognition verification evaluation data of the recognition model; The recognition module with the largest matching degree is selected to recognize the object to be recognized and obtain the recognition result.
2. The automatic identification method according to claim 1, It is characterized in that The method further comprises the steps of: Determine whether the recognition result meets the preset review conditions corresponding to the interest content to be identified. If so, execute the corresponding evaluation strategy; otherwise, put the object to be identified into the observation database.
3. The automatic identification method according to claim 1, It is characterized in that The method of pre-building a recognition model includes the following steps: Obtain training sample data set and verification sample data set; The training sample data set is input into the neural network basic learning model for training to obtain the trained recognition model; Input the verification sample data set into the trained recognition model for recognition to obtain the recognition result; According to the recognition results, the recognition accuracy of the recognition model is calculated; The recognition accuracy of the recognition model is compared with the preset requirement threshold. If the recognition accuracy of the recognition model is less than the preset requirement threshold, the recognition model is optimized; otherwise, there is no need to optimize the recognition model.
4. The automatic identification method according to claim 1, It is characterized in that The recognition module recognizes the object to be recognized, and the obtained recognition result is the data of interest in the object to be recognized.
5. The automatic identification method according to claim 4, It is characterized in that Set different priorities for different categories of interesting data in advance; The server processes the interested data with a higher priority level first according to the priority level pre-set for the interested data.
6. The automatic identification method according to claim 2, It is characterized in that The evaluation strategy includes calculating the degree of vehicle congestion on a certain road based on the results of the recognition model.
7. The automatic identification method according to claim 6, characterized in that: in, The calculation formula for the degree of vehicle congestion on a certain road is as follows: Among them, δ≠θ, a≠b; Among them, YJ represents the degree of congestion of vehicles on a certain road; r 1 represents the weight influencing factor of the number of vehicles on a certain road; Fe represents the width of a certain road; Kn represents the number of vehicles counted on a certain road after the video sampling ends; δ and θ are parameters; r 2 Represents the weighted influencing factor of the distance between vehicles on a certain road; r 1 +r 2 =1; LB indicates the number of vehicles traveling in parallel; μ 1 Indicates the influence weight of the front and rear distance between vehicles; μ 2 Indicates the influence weight of the left and right distance between vehicles; HX ab Indicates the lateral distance between the ath vehicle and the bth vehicle traveling in parallel, that is, the left-right distance; QH δθ represents the distance between the δth vehicle and the θth vehicle.
8. The automatic identification method according to claim 2, It is characterized in that The first attribute feature data of the object to be identified includes: the type of the object to be identified, the identification type of the interest content to be identified, and the size of the object to be identified.
9. An automatic identification device, It is characterized in that The method according to any one of claims 1 to 8 is performed, the device comprising: A data receiving module, used for receiving an object to be identified and interesting content to be identified; A data collection module, used to collect first attribute feature data of the object to be identified and second attribute feature data of multiple pre-built recognition models, and obtain recognition verification evaluation data of the recognition models; A data processor, which calculates the matching degree between the object to be identified and each pre-built recognition model according to the first attribute feature data of the object to be identified and the second attribute feature data of the pre-built multiple recognition models, and the recognition verification evaluation data of the recognition model; The recognition module is used to recognize the object to be recognized and obtain the recognition result.
10. The automatic identification device according to claim 9, It is characterized in that The device also includes: The judgment module is used to judge whether the recognition result meets the preset review conditions corresponding to the interest content to be identified. If so, the corresponding evaluation strategy is executed; otherwise, the object to be identified is placed in the observation database.
Citation Information
Cited By
Method and system for quickly identifying oversized ground object image on edge side
CN120279416A
A method and system for rapidly recognizing large objects on the edge
CN120279416B
