Efficient retrieval and management system and method based on vehicle characteristics

Abstract

The invention provides an efficient retrieval and management system and method based on vehicle features, and relates to the technical field of feature management, and the system comprises a feature extraction module which is used for processing an input vehicle picture and extracting vehicle feature data; the feature management module comprises a binary feature file storage unit; a memory mapping cache unit; a feature insertion and updating unit; the feature comparison and sorting unit is used for comparing the input features with a static feature library by adopting an efficient similarity calculation algorithm; a random feature extraction unit; and the HTTP service interface module is used for providing RESTful interface services of vehicle searching by images, feature comparison, feature sorting, feature insertion and updating. The technical problem of low vehicle feature retrieval and management efficiency in the prior art can be solved, and the technical effect of improving the vehicle feature retrieval and management efficiency is achieved.

Description

Technical Field

[0001] This application relates to the field of feature management technology, and in particular to an efficient retrieval and management system and method based on vehicle features. Background Technology

[0002] With the rapid development of artificial intelligence and intelligent transportation technologies, vehicle feature-based retrieval and management systems are showing increasingly important application value in fields such as smart transportation and security monitoring.

[0003] Currently, existing vehicle feature retrieval systems have significant shortcomings at the core algorithm level. Most systems still employ traditional serial computation modes for feature similarity comparison, failing to effectively utilize the parallel computing capabilities of modern processor architectures. When the feature vector dimension is high and the feature database exceeds one million records, comparison algorithms based on simple loop traversal lead to an exponential increase in computation time. Furthermore, existing vehicle feature retrieval systems lack optimized use of CPU vectorized instruction sets, failing to fully leverage the hardware's computing potential. This results in response times for a single retrieval request potentially reaching tens of seconds or even minutes, severely failing to meet the sub-second response requirements of real-world applications.

[0004] In summary, existing technologies suffer from slow retrieval speeds due to the low computational efficiency of feature comparison algorithms, which further impacts the effectiveness of intelligent transportation systems in vehicle tracking, real-time deployment, and other business operations. Summary of the Invention

[0005] The purpose of this application is to provide an efficient retrieval and management system and method based on vehicle features, in order to solve the technical problem in the prior art where the low computational efficiency of feature comparison algorithms leads to slow retrieval speed, which further affects the business effectiveness of intelligent transportation systems in vehicle tracking, real-time deployment and other tasks.

[0006] In view of the above problems, this application provides an efficient retrieval and management system and method based on vehicle features.

[0007] Firstly, this application provides an efficient retrieval and management system based on vehicle features, comprising: a feature extraction module for processing input vehicle images and extracting vehicle feature data; a feature management module for storing, updating, retrieving, and managing the vehicle feature data, the feature management module comprising: a binary feature file storage unit for storing the vehicle feature data in binary file format; a memory-mapped cache unit for loading the vehicle feature data into a memory Map structure and grouping it according to feature type; a feature insertion and update unit for determining whether the grouped vehicle feature data already exists, and if it exists, locating and updating it; if it does not exist, appending it to the group to obtain input features; a feature comparison and sorting unit for comparing the input features with a static feature library using an efficient similarity calculation algorithm and outputting TopN similarity results; a random feature extraction unit for randomly extracting a preset number of features from the static feature library and sorting them; and an HTTP service interface module for providing RESTful interface services for image-based vehicle search, feature comparison, feature sorting, feature insertion, and updating.

[0008] Preferably, the efficient retrieval and management system based on vehicle features is further configured such that: the feature management module uses a two-layer structure of binary files and memory Maps for feature location, addition, and overwrite updates.

[0009] Preferably, the efficient retrieval and management system based on vehicle features is further configured to: integrate a high-performance computing library into the feature comparison and sorting unit, perform efficient similarity calculation of feature vectors using vectorized operations, and perform TopN retrieval and full-database sorting.

[0010] Preferably, the efficient retrieval and management system based on vehicle features is further configured such that: the memory-mapped cache unit establishes independent grouped feature maps according to feature types.

[0011] Preferably, the efficient retrieval and management system based on vehicle features is further configured to: allow high-concurrency access to the HTTP service interface module, and employ thread pools and asynchronous task scheduling mechanisms for high-concurrency processing of feature extraction and comparison.

[0012] Secondly, this application also provides an efficient retrieval and management method based on vehicle features, including: receiving vehicle images uploaded by users, extracting vehicle feature data through a feature extraction module; storing the vehicle feature data through a feature management module, locating and updating the data if it already exists, and appending it to a binary feature file if it does not exist, and synchronously updating the memory Map cache to obtain input features; receiving feature comparison requests, performing efficient similarity calculation between the input features and the static feature library through a feature comparison and sorting unit, and outputting TopN similarity results; and providing feature insertion, updating, comparison, and feature sorting services through an HTTP service interface module.

[0013] The technical solution provided in this application has at least the following technical effects or advantages: by achieving the technical goal of rapid comparison and retrieval of vehicle feature data, it achieves rapid response in the feature database, improves vehicle recognition accuracy and system throughput, and effectively supports the technical effects of intelligent transportation systems in vehicle tracking and deployment scenarios with extremely high real-time requirements.

[0014] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the structure of an efficient retrieval and management system based on vehicle features according to this application.

[0017] Figure 2 This is a flowchart illustrating an efficient retrieval and management method based on vehicle features, as proposed in this application.

[0018] Figure labeling: Feature extraction module 1, Feature management module 2, Binary feature file storage unit 21, Memory mapping cache unit 22, Feature insertion and update unit 23, Feature comparison and sorting unit 24, Random feature extraction unit 25, HTTP service interface module 3. Detailed Implementation

[0019] This application provides a high-efficiency retrieval and management system and method based on vehicle features, solving the technical problem in existing technologies where the low computational efficiency of feature comparison algorithms leads to slow retrieval speeds, further affecting the operational effectiveness of intelligent transportation systems in vehicle tracking, real-time deployment, and other business scenarios. It achieves the technical goal of rapid comparison and retrieval of vehicle feature data, enabling fast response in the feature database, improving vehicle recognition accuracy and system throughput, and effectively supporting the technical performance of intelligent transportation systems in real-time vehicle tracking and deployment scenarios.

[0020] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0021] Example 1, please refer to the appendix. Figure 1 This application provides an efficient retrieval and management system based on vehicle features, specifically including:

[0022] Feature extraction module 1 is used to process the input vehicle image and extract vehicle feature data.

[0023] Specifically, feature extraction module 1 is a component that processes vehicle images, analyzes and extracts information that can represent the unique identity of a vehicle from the vehicle images, and represents it as a series of values, namely vehicle feature data.

[0024] Feature management module 2 is used to store, update, retrieve, and manage the vehicle feature data. The feature management module includes: a binary feature file storage unit 21, used to store the vehicle feature data in binary file chunks; a memory-mapped cache unit 22, used to load the vehicle feature data into a memory Map structure and manage it in groups according to feature type; a feature insertion and update unit 23, used to determine whether the grouped vehicle feature data already exists; if it exists, it locates and overwrites the data; if it does not exist, it appends the data to the group to obtain the input features; a feature comparison and sorting unit 24, used to compare the input features with a static feature library using an efficient similarity calculation algorithm and output TopN similarity results; and a random feature extraction unit 25, used to randomly extract a preset number of features from the static feature library and sort them.

[0025] Specifically, the feature management module is responsible for managing the extracted digital fingerprints for storage, timely updates, quick searches, and overall maintenance. To achieve the management goals, the feature management module 2 includes several components with clearly defined functions.

[0026] The binary feature file storage unit 21 writes vehicle feature data into a file in a low-level format that a binary computer can directly understand. In order to facilitate management and improve efficiency, the total data is divided into multiple smaller parts for storage.

[0027] The memory-mapped cache unit 22 maps the binary feature data stored on the hard disk to the extremely fast memory and organizes it in an efficient data structure called Map. It can be classified into different groups according to the feature type, such as whether it is front or rear information, so as to achieve rapid positioning.

[0028] Based on this, the feature insertion and update unit 23 is responsible for handling the addition of new features or the modification of old features. When new feature data arrives, it first checks whether it has been recorded in the grouped management Map structure. If it is found to exist, it finds its exact location and overwrites the old data with the new data to achieve the update. If it is confirmed to be a brand new feature, it adds the new data to the end of the binary file and creates a new entry in the memory Map accordingly. The feature being processed is called the input feature.

[0029] When a search is required, the feature comparison and sorting unit 24 uses an optimized, fast similarity calculation algorithm to compare the user-provided input features with all features in a pre-prepared, static feature library that does not change frequently. Then, it finds the most similar results and sorts them from high to low similarity, for example, giving the top ten results with the highest similarity.

[0030] In addition, the random feature extraction unit 25 provides an auxiliary function that can randomly extract a certain number of feature samples from the static feature library according to a pre-set quantity, for testing or evaluating the effect of the feature ranking algorithm.

[0031] HTTP service interface module 3 is used to provide RESTful interface services for image-based vehicle search, feature comparison, feature sorting, feature insertion and updating.

[0032] Specifically, the HTTP service interface module 3 provides a series of standardized network service entry points for external users or applications. The network services follow the RESTful design specifications, enabling users to conveniently use core functions such as image-based vehicle search, feature comparison, feature sorting, and feature insertion and updating through network requests.

[0033] Furthermore, this application also includes: the feature management module 2 adopts a two-layer structure of binary files and memory Map for feature location, appending, and overwriting updates.

[0034] Specifically, Feature Management Module 2 is responsible for handling the lifecycle of vehicle feature data and employs a two-tier design combining persistent storage and high-speed memory access. The first layer is a binary file, which writes data directly to disk in the computer's low-level format—a sequence of 0s and 1s—ensuring data is not lost even after power failure. The second layer is an in-memory Map structure, a key-value data structure that organizes data in the computer's fast memory, enabling extremely fast querying and access.

[0035] A two-tiered structure of binary files and in-memory Maps is used for feature location, appending, and overwriting updates. Feature location refers to quickly finding the location of specific feature data in storage based on a given identifier, primarily relying on the in-memory Map structure for instantaneous lookup. Feature appending refers to adding the data to the end of the binary file when a completely new feature is identified, thus persisting the new data. Feature overwriting updates refer to finding the corresponding position of the feature in both the binary file and the in-memory Map when it is discovered that the feature already exists, and replacing the old version with the new data version to ensure the data is up-to-date.

[0036] Furthermore, this application also includes: the feature comparison and sorting unit 24 integrates a high-performance computing library, uses vectorized operations to perform efficient similarity calculation of feature vectors, and performs TopN retrieval and full-database sorting.

[0037] Specifically, the feature comparison and ranking unit 24 is the key component responsible for feature search and ranking. The feature comparison and ranking unit 24 integrates a high-performance computing library, a deeply optimized software toolset that maximizes the computational potential of computer hardware, especially the central processing unit (CPU). The core technology employed by the feature comparison and ranking unit 24 is vectorized computation, which allows the CPU to perform the same mathematical operations on an entire set of data—a vector—simultaneously within a single instruction cycle, instead of processing data one by one as in the traditional method, thereby improving computational throughput.

[0038] High-performance computing libraries are used for efficient similarity calculations between feature vectors. A feature vector is an array of numerical values ​​extracted from a vehicle image, representing the vehicle's features. Similarity calculation is a mathematical method used to measure the closeness of two different feature vectors in numerical space; the closer the values, the more similar the two vehicles appear. Through vectorized operations, large-scale similarity comparisons between an input feature vector and a massive database of feature vectors can be performed quickly in a single operation.

[0039] The feature comparison and ranking unit 24 provides two main output modes to meet different application scenarios. One is TopN retrieval, where N is a user-defined number. It quickly finds the top N results most similar to the input feature from the entire feature library and directly returns this short list in descending order of similarity, emphasizing speed and the accuracy of core results. The other is full-library ranking, which calculates and compares each feature in the high-performance computing library with the input feature, and then generates a complete list arranged in descending order of similarity score.

[0040] Furthermore, this application also includes: the memory-mapped cache unit 22 establishes independent grouped feature maps according to feature types.

[0041] Specifically, the memory-mapped cache unit 22 is responsible for mapping feature data stored on the hard disk to faster memory for high-speed access. The core data structure of the memory-mapped cache unit 22 is a data organization method that allows for quick lookup of corresponding values ​​using a key.

[0042] The memory-mapped cache unit 22 is logically divided according to feature type. Feature type refers to the artificial classification of vehicle features, such as distinguishing features into front features representing the front of the vehicle and rear features representing the rear of the vehicle. An independent and complete grouped feature map is created and managed for each feature type. That is, there is a front feature map for storing all front features and another rear feature map specifically for storing all rear features, which are isolated from each other and do not interfere with each other.

[0043] Furthermore, this application also includes: the HTTP service interface module 3 performs high-concurrency access and uses a thread pool and asynchronous task scheduling mechanism for high-concurrency processing of feature extraction and comparison.

[0044] Specifically, the HTTP service interface module 3 is the window for communicating with the outside world. It provides a series of network addresses that conform to the HTTP protocol, allowing many users or applications to send requests to the system via the Internet at the same time.

[0045] To effectively handle such high-concurrency scenarios and avoid system crashes due to excessive requests, HTTP service interface module 3 employs two coordination mechanisms. The first is a thread pool. When a new network request arrives, the thread pool allocates a thread specifically to handle that request, saving preparation time and enabling a rapid response to the influx of traffic.

[0046] The second mechanism is asynchronous task scheduling. When a complex user request is received, such as a car search by image, the request may contain time-consuming steps, such as feature extraction and feature comparison. The asynchronous mechanism first registers the task, and then immediately releases the thread to go back to handle other new requests. The registered time-consuming tasks will be scheduled and executed in the background by a dedicated system. When the task is finally completed, the result will be returned to the user.

[0047] In summary, the efficient retrieval and management system based on vehicle features provided in this application has the following technical effects: by achieving the technical goal of rapid comparison and retrieval of vehicle feature data, it achieves rapid response in the feature database, improves vehicle recognition accuracy and system throughput, and effectively supports the technical effects of intelligent transportation systems in vehicle tracking and deployment scenarios with extremely high real-time requirements.

[0048] Example 2: Based on the same inventive concept as the efficient retrieval and management system based on vehicle features in the foregoing examples, this application also provides an efficient retrieval and management method based on vehicle features. Please refer to the appendix. Figure 2 The system includes: receiving vehicle images uploaded by users; extracting vehicle feature data through a feature extraction module; storing the vehicle feature data through a feature management module, locating and updating the data if it already exists, and appending it to a binary feature file if it does not exist, while simultaneously updating the memory Map cache to obtain input features; receiving feature comparison requests; performing efficient similarity calculation between the input features and the static feature library through a feature comparison and sorting unit, and outputting TopN similarity results; and providing feature insertion, updating, comparison, and feature sorting services through an HTTP service interface module.

[0049] Furthermore, the efficient retrieval and management method based on vehicle features also includes: using a high-performance computing library to calculate feature similarity.

[0050] Furthermore, the efficient retrieval and management method based on vehicle features also includes: using a two-layer structure of binary files and memory Maps for feature location, appending, and overwriting updates.

[0051] Furthermore, the efficient retrieval and management method based on vehicle features also includes: grouping and managing features according to feature type, and storing and retrieving the front and rear features of the vehicle independently.

[0052] Furthermore, the efficient retrieval and management method based on vehicle features also includes: using a thread pool and asynchronous task scheduling mechanism for high-concurrency processing of feature extraction and comparison.

[0053] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The efficient retrieval and management system based on vehicle features and specific examples in the foregoing embodiment one are also applicable to the efficient retrieval and management method based on vehicle features in this embodiment. Through the foregoing detailed description of the efficient retrieval and management system based on vehicle features, those skilled in the art can clearly understand the efficient retrieval and management method based on vehicle features in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0054] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0055] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. An efficient retrieval and management system based on vehicle features, characterized in that, include: The feature extraction module is used to process the input vehicle images and extract vehicle feature data; The feature management module is used to store, update, retrieve, and manage the vehicle feature data. The feature management module includes: A binary feature file storage unit is used to store the vehicle feature data in segments as binary files; The memory-mapped cache unit is used to load the vehicle feature data into a memory Map structure and manage it in groups according to feature type; The feature insertion and update unit is used to determine whether the vehicle feature data after group management already exists. If it exists, it locates and overwrites the data; if it does not exist, it appends the data to the group and obtains the input features. The feature comparison and ranking unit is used to compare the input features with the static feature library using an efficient similarity calculation algorithm, and output the Top N similarity results; The random feature extraction unit is used to randomly extract a preset number of features from the static feature library and sort the features. The HTTP service interface module provides RESTful interface services for image-based vehicle search, feature comparison, feature sorting, feature insertion, and update.

2. A high efficient retrieval and management system based on vehicle features as claimed in claim 1 wherein, The feature management module uses a two-layer structure of binary files and memory Maps to locate, append, and update features.

3. A high efficient retrieval and management system based on vehicle features as claimed in claim 1 wherein, The feature comparison and sorting unit integrates a high-performance computing library, uses vectorized operations to perform efficient similarity calculation of feature vectors, and performs TopN retrieval and full-database sorting.

4. A high efficient retrieval and management system based on vehicle features as claimed in claim 1 wherein, The memory-mapped cache unit establishes independent grouped feature maps according to feature type.

5. A high efficient retrieval and management system based on vehicle features as claimed in claim 1 wherein, The HTTP service interface module performs high-concurrency access and uses a thread pool and asynchronous task scheduling mechanism for high-concurrency processing of feature extraction and comparison.

6. A method for efficient retrieval and management based on vehicle features, characterized in that, Executed by any one of claims 1 to 5, a high-efficiency retrieval and management system based on vehicle features, comprising: Receive vehicle images uploaded by users and extract vehicle feature data through the feature extraction module; The vehicle feature data is stored through the feature management module. If it already exists, it is located and overwritten. If it does not exist, it is appended to the binary feature file and the memory Map cache is updated synchronously to obtain the input features. It receives feature comparison requests, performs efficient similarity calculation between input features and static feature library through feature comparison and sorting unit, and outputs TopN similarity results; The HTTP service interface module provides services for feature insertion, updating, comparison, and feature sorting.

7. The efficient retrieval and management method based on vehicle features as described in claim 6, characterized in that, Feature similarity calculation is performed using a high-performance computing library.

8. The efficient retrieval and management method based on vehicle features as described in claim 6, characterized in that, A two-layer structure of binary files and in-memory Maps is used for feature location, appending, and overwriting updates.

9. The efficient retrieval and management method based on vehicle features as described in claim 6, characterized in that, The features are grouped and managed according to their type, and the features of the front and rear of the vehicle are stored and retrieved independently.

10. The efficient retrieval and management method based on vehicle features as described in claim 6, characterized in that, High-concurrency processing of feature extraction and comparison is achieved by using thread pools and asynchronous task scheduling mechanisms.

Images