A vehicle logo detection and recognition method based on vehicle logo repositioning

Abstract

This invention discloses a vehicle logo detection and recognition method based on vehicle logo relocation, belonging to the field of vehicle logo detection and recognition technology. It addresses the problem of insufficient stability and accuracy in vehicle logo detection and positioning due to dynamic factors such as multi-angle changes and susceptibility to interference from vehicle vibration and lighting variations. The method involves acquiring vehicle images via a camera, preprocessing them, and then using an edge detection algorithm to extract vehicle contour information and filter out the target vehicle image range. Within this range, wheel hub information is identified, generating vehicle wheel hub features to determine whether to perform vehicle logo detection and positioning. Positioned vehicle logos are recorded, and vehicle vibration frequency and braking conditions are monitored in real time. Vehicle logo influence features are calculated based on braking conditions to set the vehicle logo influence weight. Vehicles are selected and marked based on vehicle vibration frequency and then sorted. Secondary vehicle logo detection and positioning is performed on the marked vehicles according to the sorting results, thereby achieving accurate vehicle logo recognition.

Description

Technical Field

[0001] This invention relates to the field of vehicle logo detection and recognition technology, and more specifically, to a vehicle logo detection and recognition method based on vehicle logo relocation. Background Technology

[0002] With the continuous development of intelligent transportation systems, vehicle recognition and dynamic detection technologies based on video image processing have been widely applied in scenarios such as road monitoring, vehicle management, and traffic enforcement. Among these, vehicle logos, as an important identifier for vehicle manufacturers, are crucial for tasks such as vehicle attribute extraction, brand classification, and traffic flow feature analysis.

[0003] The existing technology has the following shortcomings:

[0004] Currently, vehicle recognition mainly relies on traditional license plate recognition and whole-vehicle image feature extraction methods. However, in complex environments, such as when license plates are damaged, obscured, or the image is blurred, the accuracy and robustness of license plate recognition significantly decrease. To improve the comprehensiveness and redundancy of recognition, some studies have proposed using vehicle logo recognition as an auxiliary information source. However, vehicle logos themselves are small in area, subject to multi-angle changes, and are easily affected by dynamic factors such as vehicle vibration and changes in lighting, resulting in insufficient stability and accuracy in their detection and localization. Therefore, this paper proposes a vehicle logo detection and recognition method based on vehicle logo relocalization.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a vehicle logo detection and recognition method based on vehicle logo relocation, which solves the problems mentioned in the background art by employing a wheel hub feature-assisted judgment mechanism and a vehicle dynamic behavior feature weighting strategy.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for vehicle logo detection and recognition based on vehicle logo relocation, comprising the following steps:

[0008] Step S1: Capture vehicle images through a camera, and after preprocessing, use an edge detection algorithm to extract vehicle contour information from the vehicle images and filter out the target vehicle image range.

[0009] Step S2: Identify wheel hub information within the target vehicle image range, generate wheel hub features of the vehicle, and determine whether to detect and locate the vehicle logo based on the wheel hub features of the vehicle.

[0010] Step S3: Record the detected and located vehicle logo, monitor the vehicle vibration frequency and vehicle braking status in real time, and calculate the vehicle logo influence characteristics based on the vehicle braking status.

[0011] Step S4: Set the influence weight of vehicle logos based on the characteristics of vehicle logo influence, filter out marked vehicles based on vehicle vibration frequency and sort them for detection, and perform secondary detection and positioning of marked vehicles according to the detection sort.

[0012] In a preferred embodiment, in step S1, the camera continuously captures vehicle images at a preset frame rate and merges the vehicle images into a vehicle image frame sequence in chronological order. The vehicle images are preprocessed by an image processing tool, including grayscale processing and noise suppression processing.

[0013] Grayscale conversion is used to convert vehicle images into grayscale images;

[0014] Noise suppression processing is used to smooth and reduce noise in grayscale images, resulting in filtered vehicle images.

[0015] In a preferred embodiment, in step S1, edge contour pixels of the vehicle image are extracted using the Cannibal edge detection algorithm. The Cannibal edge detection algorithm includes gradient calculation, non-maximum suppression, and a double threshold connection method. The specific steps are as follows:

[0016] After calculating the gradient intensity and direction of each pixel, an edge response map is generated based on non-maximum suppression, and then edge contour pixels are selected by connecting the two thresholds.

[0017] In a preferred embodiment, in step S1, edge contour pixels of the target vehicle are filtered, strong edges are detected by image processing tools, and the region of closed edge contour in the vehicle image is extracted.

[0018] The minimum bounding rectangle of the contour region is obtained by boundary fitting, and its aspect ratio is calculated. A first aspect ratio threshold and a second aspect ratio threshold are preset.

[0019] If the aspect ratio is between the first and second aspect ratio thresholds, the outline region is determined to be the target vehicle region; otherwise, the outline region is determined to be a non-target vehicle region.

[0020] In a preferred embodiment, in step S2, based on the target vehicle region, the standard wheel hub region is identified by calling the Hough circle transform method using an image processing tool. The identification process satisfies the following rules:

[0021] Rule 1: If the radius of the identified circular region is within the preset radius range, the circular region is retained; otherwise, the circular region is not retained.

[0022] Rule 2: Calculate the edge density if the identified circular region contains edge density; if the edge density is greater than the preset edge density threshold, the circular region is retained; otherwise, the circular region is not retained.

[0023] If all the above rules are met, then the circular area is determined to be the hub area.

[0024] In a preferred embodiment, the size features and texture features of the wheel hub region are extracted, where the wheel hub size features are the radius of the circular region and the wheel hub texture features are the edge density.

[0025] A logistic regression model is constructed based on wheel hub size and texture features to determine whether the target vehicle region is a car region.

[0026] Wheel size features and wheel texture features are normalized and summed as logistic regression parameters for the wheel hub. A logistic regression model is constructed based on the logistic regression parameters of the wheel hub to obtain the vehicle prediction probability.

[0027] If the predicted probability of a car is greater than the preset probability threshold, the target vehicle area is determined to be a car area, and the car logo is detected and located; otherwise, the current area is determined to be a non-car area.

[0028] In a preferred embodiment, in step S3, the vehicle logo is detected and located, the center coordinates of the vehicle logo area are obtained, the center coordinates of the target area of ​​each frame in the vehicle image frame sequence are retrieved, and the change in the position of the vehicle logo between adjacent frames is calculated.

[0029] When the change in the position of the car logo is greater than the preset change threshold, the corresponding vehicle shakes. The vehicle shake frequency in the vehicle image frame sequence is calculated by the number of times the vehicle shakes and the total number of vehicle images in the vehicle image frame sequence.

[0030] The vehicle speed is calculated by the change in the position of the vehicle logo in adjacent frames, and the speeds in the vehicle image frame sequence are merged into a vehicle speed sequence.

[0031] The number of braking operations is the number of times the vehicle's speed decreases in the speed sequence, with continuous decreases counted as only one.

[0032] The ratio of the average speed to the maximum speed of a vehicle in a vehicle image frame sequence is used as the braking speed ratio.

[0033] In a preferred embodiment, in step S3, the influence characteristics of the vehicle logo are evaluated using fuzzy inference, which combines the number of braking attempts and the braking speed ratio.

[0034] The ratio of braking number to braking speed is defined as an input variable, and these variables are divided into different fuzzy sets.

[0035] The influence features of the car logo are defined as output variables and divided into fuzzy sets;

[0036] A set of fuzzy rules is formulated to describe the influence of different input variables on the output variable, and the influence characteristics of the car logo are evaluated based on fuzzy inference.

[0037] In a preferred embodiment, in step S4, the membership values ​​in each fuzzy set are obtained according to a preset membership function as the influence weight of the car logo;

[0038] The product of the vehicle logo's influence weight and the vehicle's vibration frequency is used as the comprehensive score. If the comprehensive score is greater than the comprehensive score threshold, the vehicle is marked; otherwise, the vehicle is not marked.

[0039] The comprehensive score of the marked vehicles is sorted in descending order, and the vehicle logo images of the marked vehicles are then subjected to secondary detection in sequence.

[0040] The technical effects and advantages of this invention are as follows:

[0041] 1. This invention acquires vehicle images via a camera, preprocesses them, and uses an edge detection algorithm to extract vehicle contour information and filter out the target vehicle image range. Within the target vehicle image range, wheel hub information is identified, generating wheel hub features to determine whether to detect and locate the vehicle logo. Detected and located logos are recorded. The vehicle vibration frequency and braking status are monitored in real time. Based on the vehicle braking status, the vehicle logo influence features are calculated to set the logo influence weight. Combined with the vehicle vibration frequency, marked vehicles are filtered and sorted for detection. The marked vehicles are then subjected to secondary detection and positioning according to the detection ranking, achieving accurate logo recognition. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating the implementation of a vehicle logo detection and recognition method based on vehicle logo relocation according to the present invention.

[0043] Figure 2 This is a schematic diagram illustrating the steps of a vehicle logo detection and recognition method based on vehicle logo relocation according to the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] Example 1: A method for vehicle logo detection and recognition based on vehicle logo relocation, such as... Figures 1 to 2 As shown, it includes the following steps:

[0046] Step S1: Capture vehicle images through a camera, and after preprocessing, use an edge detection algorithm to extract vehicle contour information from the vehicle images and filter out the target vehicle image range.

[0047] Step S2: Identify wheel hub information within the target vehicle image range, generate wheel hub features of the vehicle, and determine whether to detect and locate the vehicle logo based on the wheel hub features of the vehicle.

[0048] Step S3: Record the detected and located vehicle logo, monitor the vehicle vibration frequency and vehicle braking status in real time, and calculate the vehicle logo influence characteristics based on the vehicle braking status.

[0049] Step S4: Set the influence weight of vehicle logos based on the characteristics of vehicle logo influence, filter out marked vehicles based on vehicle vibration frequency and sort them for detection, and perform secondary detection and positioning of marked vehicles according to the detection sort.

[0050] The specific implementation is as follows:

[0051] In step S1, the camera continuously captures vehicle images at a preset frame rate and merges the vehicle images into a vehicle image frame sequence in chronological order. The vehicle images are preprocessed by image processing tools, including grayscale processing and noise suppression processing.

[0052] Grayscale processing is used to convert vehicle images into grayscale images, removing the interference of color information on edge detection. A weighted average method is used to calculate the grayscale value by proportionally weighting the red, green, and blue channels. Where a, b, and c are preset weighting coefficients. G and B are the pixel values ​​corresponding to the red, green, and blue channels, respectively. pixel coordinates A grayscale image is formed by arranging the grayscale values ​​of each pixel in a two-dimensional matrix, where the grayscale values ​​are the pixel coordinates.

[0053] Noise suppression processing is used to smooth and reduce noise in grayscale images. A Gaussian function is used to remove random noise from the grayscale image to obtain the processed vehicle image. The Gaussian function determines the weighting factor, and the weighted sum of the pixel values ​​in the neighborhood is used as the center pixel value. ,in, Standard deviation The weighting factor is multiplied by each neighborhood of the grayscale image and summed. ,in, Let be the grayscale value of a neighboring pixel in the grayscale image, and k be the preset kernel radius. Image of a vehicle;

[0054] The edge contour pixels of vehicle images are extracted using the Canni edge detection algorithm. The Canni edge detection algorithm includes gradient calculation, non-maximum suppression, and a double threshold connection method. The specific steps are as follows:

[0055] Gradient calculation method: First-order differential operators are used to convolve the X and Y axes of the filtered image respectively to calculate the gray-level change rate of each pixel in the vehicle image along the X and Y axes. ,in, and These represent the grayscale change rates of the vehicle image along the x and y axes, respectively.

[0056] Gradient strength of pixels is obtained based on Euclidean distance: Where G is the pixel gradient intensity value; the pixel gradient intensity value represents the degree of grayscale change at the corresponding pixel. The larger the pixel gradient intensity value, the more likely it is to be an edge point.

[0057] The direction of the corresponding gradient is determined by the arctangent function: ,in, The gradient direction of the pixel edge.

[0058] Non-maximum suppression method: retain the pixel with the largest gradient value among the adjacent pixels in the gradient direction of the pixel edge, and suppress the gradient response in non-edge regions; compare the gradient intensity value of the current pixel with the gradient intensity values ​​of the two adjacent pixels along the corresponding direction to obtain the edge response map: if the gradient intensity value of the current pixel is not the maximum value in the corresponding positive direction, then set the current gradient intensity value to zero; otherwise, retain the current gradient intensity value.

[0059] Dual-threshold connection method: A first gradient intensity threshold and a second gradient intensity threshold are preset, with the first gradient intensity threshold being greater than the second gradient intensity threshold. Pixel gradient intensity values ​​in the edge response map are divided into different edge types, and edge contour pixels are filtered out.

[0060] If the pixel gradient intensity value is greater than the first gradient intensity threshold, it is marked as a strong edge pixel; if the pixel gradient intensity value is less than or equal to the first gradient intensity threshold and greater than the second gradient intensity threshold, it is marked as a weak edge pixel; if the pixel gradient intensity value is less than or equal to the second gradient intensity threshold, it is marked as a non-edge pixel.

[0061] When a pixel is marked as a strong edge pixel, the corresponding pixel is an edge contour; when a pixel is marked as a weak edge pixel, if a strong edge pixel exists in the pixel's neighborhood, the corresponding pixel is an edge contour, otherwise it is marked as a non-edge pixel.

[0062] Further filter the edge contour pixels of the target vehicle, use image processing tools to detect the edge contours, and extract the regions with closed edge contours in the vehicle image;

[0063] Calculate the aspect ratio of the contour region to determine if it is a vehicle region: Obtain the minimum bounding rectangle by fitting the boundary of the contour region. The height of the minimum bounding rectangle is used as the numerator, and the width of the minimum bounding rectangle is used as the denominator. The ratio of the two is used as the aspect ratio. Preset aspect ratio threshold 1 and aspect ratio threshold 2. If the aspect ratio is between the aspect ratio threshold 1 and aspect ratio threshold 2, the contour region is determined to be the target vehicle region. Otherwise, the contour region is determined to be a non-target vehicle region.

[0064] For example, if the aspect ratio threshold 1 and aspect ratio threshold 2 are set to 1.2 and 4 respectively, and the aspect ratio is 2, then the outline region is determined to be the vehicle outline.

[0065] By using structured image preprocessing and edge analysis, background interference is effectively eliminated, the target vehicle region is accurately extracted, and subsequent steps are focused on the effective vehicle image region, thereby improving detection efficiency and accuracy.

[0066] It should be noted that the preset frame rate is a parameter configured by the camera, representing the number of images captured continuously per second, denoted as F. For example, 30 FPS means capturing 30 frames of vehicle images per second, and the time interval between two adjacent frames is: The standard deviation in the Gaussian function controls the degree of blurring. A larger standard deviation results in a smoother weight distribution, stronger filtering, and weaker preservation of image details. This setting is determined by professionals and will not be elaborated upon here. The preset kernel radius controls the size of the pixel neighborhood participating in the convolution operation and has a linear relationship with the standard deviation. This setting is also determined by professionals. The image processing tool is an open-source computer vision library capable of preprocessing and feature extraction of vehicle images. The Cannibal edge detection algorithm is a multi-stage edge detection algorithm. It uses a Gaussian function for noise suppression, combines a first-order differential operator to extract image gradient information, and uses non-maximum suppression to accurately preserve the true edge structure. Finally, it uses a dual-threshold connection strategy to classify and determine edge pixels, thereby efficiently extracting vehicle contour boundaries. The preset first and second gradient strength thresholds are parameters for edge pixel classification and are set by professionals. For example, they can be set based on experience or dynamically calculated based on the local or global gradient distribution of the image.

[0067] In step S2, based on the target vehicle area, the standard wheel hub area is identified by calling the Hough circle transform method using an image processing tool. The identification process follows these rules:

[0068] Rule 1: If the radius of the identified circular region is within the preset radius range, the circular region is retained; otherwise, the circular region is not retained.

[0069] Rule 2: Determine whether the identified circular region contains edge density, excluding smooth contours or pseudo-circles that are mistakenly identified as wheel hubs, and calculate the edge density index: ,in, This represents the number of edge pixels within the circular region. Let D be the total number of pixels within the circular area, and D be the edge density; if the edge density is greater than a preset edge density threshold... If the condition is not met, the circular area will be retained; otherwise, the circular area will not be retained.

[0070] If all the above rules are met, then the circular area is determined to be the hub area;

[0071] Extract the size features and texture features of the wheel hub region. The size features are the radius of the circular region identified in rule 1 above, and the texture features are the edge density in rule 2 above.

[0072] A logistic regression model is constructed based on wheel hub size and texture features to determine whether the target vehicle region is a car region.

[0073] The wheel hub size and texture features are normalized and summed to serve as the logistic regression parameters for the wheel hub. A logistic regression model is then constructed using these parameters. , where z is the logistic regression parameter of the wheel hub, e is the natural base, and L is the logistic regression model calculation result of the corresponding wheel hub. The logistic regression model calculation result of the wheel hub is used as the vehicle prediction probability.

[0074] If the predicted probability of a car is greater than the preset probability threshold, the target vehicle area is determined to be a car area, and the car logo is detected and located; otherwise, the current area is determined to be a non-car area.

[0075] By utilizing the stability and unique structure of vehicle wheel hubs as a preliminary screening criterion, target vehicles can be initially screened at the image structure level, effectively reducing the number of times invalid images are processed and lowering the false detection rate of vehicle logos.

[0076] It should be noted that the Hough circle transform is an image circular region detection algorithm based on a parameter space voting mechanism. It utilizes the geometric properties of a circle to transform the circular detection problem in image space into an accumulation voting process in parameter space, which can be implemented through open-source computer vision libraries. The preset probability threshold is set by professionals and will not be elaborated here.

[0077] In step S3, the car logo is detected and located, and the center coordinates of the car logo area are obtained and denoted as follows: Retrieve the center coordinates of the target region in each frame of the vehicle image frame sequence, and calculate the change in the position of the vehicle logo between adjacent frames: , This represents the change in the position of the car logo.

[0078] When the change in the position of the vehicle logo exceeds a preset threshold, the corresponding vehicle shakes. The vehicle shake frequency in the vehicle image frame sequence is calculated as follows: ,in, The number of times the vehicle shook is denoted by N, and N is the total number of vehicle images in the vehicle image frame sequence. The frequency of vehicle vibration;

[0079] Calculate vehicle speed by the change in the position of the vehicle logo between adjacent frames: ,in, The time interval between adjacent frames. For speed, the speeds in the vehicle image frame sequence are merged into a vehicle speed sequence;

[0080] The number of times the vehicle speed decreases in the speed sequence is counted as the number of braking actions, with consecutive decreases counted only once. The formula for the number of braking actions is: ,in, If the speed is lower at a certain moment than at the previous moment, Number of braking cycles;

[0081] The ratio of the average speed to the maximum speed of a vehicle in a sequence of vehicle image frames is used as the braking speed ratio: ,in, The maximum speed in the vehicle image frame sequence. R represents the average speed in the vehicle image frame sequence, and R is the braking speed ratio.

[0082] The influence characteristics of the vehicle logo are evaluated using fuzzy inference based on the ratio of braking frequency to braking speed. The specific steps are as follows:

[0083] The ratio of braking number to braking speed is defined as the input variable, and they are divided into different fuzzy sets. For example, "high" and "low" are for the number of braking, and "high" and "low" are for the braking speed ratio.

[0084] Define the influence features of the car logo as output variables and divide them into fuzzy sets, such as "low", "medium", and "high" for the influence features of the car logo.

[0085] A set of fuzzy rules is formulated to describe the influence of different input variables on the output variable. The influence characteristics of the car logo are evaluated based on fuzzy inference. These fuzzy rules can be formulated according to the actual situation, for example:

[0086] When the number of braking cycles and the braking speed ratio are both high, the impact of the vehicle logo is high.

[0087] When the number of braking attempts and the braking speed ratio are both moderate, the impact on the vehicle logo is moderate.

[0088] When the number of braking cycles and the braking speed ratio are both low, the impact of the vehicle logo is low. ...

[0090] This process is repeated to form a set of fuzzy rules.

[0091] By calculating the vehicle status and the position of the vehicle logo, and combining braking behavior analysis, a fuzzy rule set is constructed to obtain the influence characteristics of the vehicle logo, providing key parameter support for subsequent secondary detection.

[0092] It should be noted that the division of fuzzy sets can be adjusted according to the actual situation. For example, if the number of braking times of a vehicle is less than 3, it is labeled as a "low" fuzzy set; if the number of braking times is between 3 and 7, it is labeled as a "medium" fuzzy set. The preset change threshold and preset probability threshold are set by professionals and will not be elaborated here.

[0093] In step S4, the preset membership function is the degree of belonging of the input variable in the fuzzy rule to different fuzzy sets. The membership value in each fuzzy set obtained according to the preset membership function is used as the influence weight of the car logo and is denoted as W.

[0094] For example, when the influence of the car logo is high, the membership degree is 0.7; when the influence of the car logo is medium, the membership degree is 0.2; and when the influence of the car logo is low, the membership degree is 0.1.

[0095] A comprehensive score is calculated by combining the weighting of the vehicle logo and the frequency of vehicle vibration, and the vehicle undergoing secondary testing is then marked.

[0096] The comprehensive score of vehicles marked as having undergone secondary inspection is calculated and sorted. Where S is the overall score;

[0097] If the overall score is greater than the overall score threshold, the vehicle will be marked; otherwise, the vehicle will not be marked.

[0098] The comprehensive score of the marked vehicles is sorted in descending order, and the vehicle logo images of the marked vehicles are then subjected to secondary detection in sequence.

[0099] Introducing the influence weight of vehicle logos as a reference for re-identification priority enhances the adaptability of the recognition system to dynamic and complex scenarios, avoids the missed or false detection of important target vehicle logos due to dynamic disturbances, and improves the system's recognition accuracy and reliability in practical applications.

[0100] It should be noted that the membership function is used to describe the degree to which the input variable value belongs to a certain fuzzy set. It ranges between 0 and 1. The preset membership function is determined by domain professionals based on vehicle operation data and experience knowledge. The starting point, peak point and ending point of the fuzzy set corresponding to each input variable are determined, thus forming a set of continuous membership function curves. The comprehensive scoring threshold is set by professionals, which will not be elaborated here.

[0101] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0102] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0103] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0104] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0105] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0106] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0107] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0108] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0109] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for vehicle logo detection and recognition based on vehicle logo relocation, characterized in that: Includes the following steps: Step S1: Capture vehicle images through a camera, and after preprocessing, use an edge detection algorithm to extract vehicle contour information from the vehicle images and filter out the target vehicle image range. Step S2: Identify wheel hub information within the target vehicle image range, generate wheel hub features of the vehicle, and determine whether to detect and locate the vehicle logo based on the wheel hub features of the vehicle. Step S3: Record the detected and located vehicle logo, monitor the vehicle vibration frequency and vehicle braking status in real time, and calculate the vehicle logo influence characteristics based on the vehicle braking status. In step S3, the vehicle logo is detected and located, the center coordinates of the vehicle logo area are obtained, the center coordinates of the target area of ​​each frame in the vehicle image frame sequence are retrieved, and the change in the position of the vehicle logo between adjacent frames is calculated. When the change in the position of the car logo is greater than the preset change threshold, the corresponding vehicle shakes. The vehicle shake frequency in the vehicle image frame sequence is calculated by the number of times the vehicle shakes and the total number of vehicle images in the vehicle image frame sequence. The vehicle speed is calculated by the change in the position of the vehicle logo in adjacent frames, and the speeds in the vehicle image frame sequence are merged into a vehicle speed sequence. The number of braking operations is the number of times the vehicle's speed decreases in the speed sequence, with continuous decreases counted as only one. The ratio of the average speed to the maximum speed of a vehicle in a vehicle image frame sequence is used as the braking speed ratio; In step S3, the influence characteristics of the vehicle logo are evaluated using fuzzy inference, which combines the number of braking attempts and the braking speed ratio. The ratio of braking number to braking speed is defined as an input variable, and these variables are divided into different fuzzy sets. The influence features of the car logo are defined as output variables and divided into fuzzy sets; A set of fuzzy rules is formulated to describe the influence of different input variables on the output variable, and the influence characteristics of the car logo are evaluated based on fuzzy inference. Step S4: Set the influence weight of vehicle logos based on the influence characteristics of vehicle logos, filter out marked vehicles and sort them according to the vehicle vibration frequency, and perform secondary detection and positioning of marked vehicles according to the detection sort. In step S4, the membership values ​​in each fuzzy set are obtained according to the preset membership function and used as the influence weight of the car logo; The product of the vehicle logo's influence weight and the vehicle's vibration frequency is used as the comprehensive score. If the comprehensive score is greater than the comprehensive score threshold, the vehicle is marked; otherwise, the vehicle is not marked. The comprehensive score of the marked vehicles is sorted in descending order, and the vehicle logo images of the marked vehicles are then subjected to secondary detection in sequence.

2. The method for vehicle logo detection and recognition based on vehicle logo relocation according to claim 1, characterized in that: In step S1, the camera continuously captures vehicle images at a preset frame rate and merges the vehicle images into a vehicle image frame sequence in chronological order. The vehicle images are preprocessed by image processing tools, including grayscale processing and noise suppression processing. Grayscale conversion is used to convert vehicle images into grayscale images; Noise suppression processing is used to smooth and reduce noise in grayscale images, resulting in filtered vehicle images.

3. The method for vehicle logo detection and recognition based on vehicle logo relocation according to claim 2, characterized in that: In step S1, the edge contour pixels of the vehicle image are extracted using the Cannibal edge detection algorithm. The Cannibal edge detection algorithm includes gradient calculation, non-maximum suppression, and a double threshold connection method. The specific steps are as follows: After calculating the gradient intensity and direction of each pixel, an edge response map is generated based on non-maximum suppression, and then edge contour pixels are selected by connecting the two thresholds.

4. The method for vehicle logo detection and recognition based on vehicle logo relocation according to claim 3, characterized in that: In step S1, edge contour pixels of the target vehicle are filtered, and strong edges are detected by image processing tools to extract the region of closed edge contour in the vehicle image. The minimum bounding rectangle of the contour region is obtained by boundary fitting, and its aspect ratio is calculated. A first aspect ratio threshold and a second aspect ratio threshold are preset. If the aspect ratio is between the first and second aspect ratio thresholds, the outline region is determined to be the target vehicle region; otherwise, the outline region is determined to be a non-target vehicle region.

5. The method for vehicle logo detection and recognition based on vehicle logo relocation according to claim 4, characterized in that: In step S2, based on the target vehicle area, the standard wheel hub area is identified by calling the Hough circle transform method using an image processing tool. The identification process follows these rules: Rule 1: If the radius of the identified circular region is within the preset radius range, the circular region is retained; otherwise, the circular region is not retained. Rule 2: Calculate the edge density if the identified circular region contains an edge density; if the edge density is greater than a preset edge density threshold... If the condition is not met, the circular area will be retained; otherwise, the circular area will not be retained. If all the above rules are met, then the circular area is determined to be the hub area.

6. The method for vehicle logo detection and recognition based on vehicle logo relocation according to claim 5, characterized in that: Extract the size features and texture features of the wheel hub region. The size features are the radius of the circular region, and the texture features are the edge density. A logistic regression model is constructed based on wheel hub size and texture features to determine whether the target vehicle region is a car region. Wheel size features and wheel texture features are normalized and summed as logistic regression parameters for the wheel hub. A logistic regression model is constructed based on the logistic regression parameters of the wheel hub to obtain the vehicle prediction probability. If the predicted probability of a car is greater than the preset probability threshold, the target vehicle area is determined to be a car area, and the car logo is detected and located; otherwise, the current area is determined to be a non-car area.

Images