Shared bicycle brand identification and density optimization method based on NSGA-III algorithm driving

Abstract

The invention provides a shared bicycle brand identification and density optimization method based on NSGA-III algorithm driving, relates to the field of shared bicycle brand identification management, and realizes intelligent management of shared bicycles through combination of multi-source heterogeneous data acquisition, an image identification technology and an NSGA-III algorithm. Firstly, images and position data of shared bicycles are acquired by using monitoring equipment, brand identification is performed through a deep learning model, and then a density value of shared bicycle brand distribution is calculated based on the brand and position data; multi-objective optimization is carried out through an NSGA-III algorithm, distribution of the shared bicycles is dynamically adjusted, excessive concentration of the shared bicycles of a specific brand is avoided, traffic jam is reduced, a scheduling strategy of the shared bicycles is optimized, and the utilization efficiency of the shared bicycles and urban traffic mobility are improved.

Description

Technical Field

[0001] The present invention relates to the field of shared bicycle brand recognition management, and specifically to a shared bicycle brand recognition and density optimization method driven by an NSGA-III algorithm. Background Art

[0002] At present, with the widespread use of shared bicycles in urban transportation, the distribution and management of shared bicycles have become key issues affecting urban transportation efficiency; the existing shared bicycle management has significant deficiencies in dealing with dynamic urban traffic, especially in brand recognition and density optimization of shared bicycles; due to the lack of accurate recognition of shared bicycle brands, the system cannot effectively balance the distribution of multi-brand shared bicycles, resulting in excessive concentration of certain brands in specific areas, causing waste of resources and traffic congestion.

[0003] Existing technologies usually rely only on simple scheduling rules and are unable to dynamically adjust bicycle distribution to meet real-time demand; the lack of accuracy in brand recognition further limits the optimization effect. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In view of the shortcomings of the prior art, the present invention provides a shared bicycle brand identification and density optimization method driven by the NSGA-III algorithm, which solves the problem of uneven distribution and traffic congestion of shared bicycles due to excessive concentration of a certain brand.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a shared bicycle brand identification and density optimization method driven by NSGA-III algorithm, comprising the following steps:

[0008] S1. Obtaining real-time multi-source heterogeneous data of the target area where the shared bicycles are located through monitoring equipment, wherein the multi-source heterogeneous data includes images and locations of the shared bicycles and traffic flow data of the target area;

[0009] S2. Based on the image data in the multi-source heterogeneous data, an image recognition algorithm is used to extract feature data of the shared bicycle brand, and the image recognition algorithm is combined with the NSGA-III algorithm to perform multi-feature optimization to improve the accuracy of identifying different shared bicycle brands in a complex environment;

[0010] S3. Calculate the density value of each shared bicycle brand distribution in the target area based on the brand recognition results and the collected shared bicycle location data;

[0011] S4. The NSGA-III algorithm is used to optimize the density of shared bicycles. A multi-objective optimization model is generated according to the density value of the shared bicycle brand distribution. Three objective functions are generated according to the multi-objective optimization model to optimize multiple objectives, where the multiple objectives include:

[0012] The balance of resource allocation of shared bicycles of different brands;

[0013] Avoid excessive concentration of shared bikes of a particular brand;

[0014] Reduce traffic congestion caused by uneven distribution of shared bicycles;

[0015] S5. Based on the multi-objective optimization model, generate a shared bicycle dispatching strategy to dispatch shared bicycles of different brands to achieve reasonable allocation of shared bicycles of different brands and smooth traffic.

[0016] Preferably, monitoring equipment is deployed in the target area to collect multi-source heterogeneous data of shared bicycles. The deployed cameras and GPS collect the image data, location data and surrounding traffic flow data of shared bicycles in real time. The collected data is preprocessed, including cleaning, format unification and outlier processing. The preprocessed data is deeply analyzed, and the state and type of bicycles are identified by image recognition technology. The distribution and flow pattern of bicycles are revealed through location data analysis. At the same time, traffic flow analysis is combined to understand the impact of shared bicycles on urban traffic.

[0017] Preferably, in the process of brand recognition of shared bicycles, the bicycle image is captured by a camera, and then the image is standardized through image processing technology to reduce the interference of environmental factors; the standardized image is analyzed by a deep learning model to automatically identify the key features of the bicycle brand, including logos, colors and design patterns; to further improve the recognition performance, the image recognition algorithm is combined with the NSGA-III multi-objective optimization algorithm to optimize multiple features in the recognition process, such as speed, accuracy and resource consumption, to find the best algorithm configuration; this method of combining deep learning and optimization algorithms improves the accuracy of shared bicycle brand recognition, and also ensures the efficiency of the recognition process and the rational use of resources, providing technical support for the operation and management of shared bicycles.

[0018] Preferably, the iterative process of the NSGA-III algorithm starts with defining an initial optimization method that includes different parameters and feature extraction methods; in each round of iteration, the NSGA-III algorithm evaluates the performance of the current optimization method on multiple optimization objectives, and based on the evaluated performance, the NSGA-III algorithm generates new individuals through selection, crossover and mutation operations to simulate the natural selection and genetic mutation process to explore the area in the solution space that is not covered by the current optimization method; the newly generated optimization method is combined with the original optimization method to form a new optimization method, and the NSGA-III algorithm evaluates the performance of these optimization methods again, and through non-dominated sorting and The crowding distance calculation retains the best performing optimization method and eliminates the poorly performing optimization method; as the iteration proceeds, the NSGA-III algorithm continuously adjusts the feature extraction and recognition strategies to adapt to the brand recognition of shared bicycles in complex environments; including changing the image preprocessing steps, optimizing the structure and parameters of the deep learning model, or introducing new image features; in this way, the misrecognition rate is reduced and new image data is responded to in real time; after multiple rounds of iterations, the NSGA-III algorithm will converge to a set of optimal solutions, which represent the recognition strategy that achieves the best balance between multiple optimization objectives; it can maintain high performance even in complex and changing urban environments.

[0019] Preferably, the brand recognition results of shared bicycles are associated with the GPS location data of the shared bicycles, so that the brand information of each shared bicycle is accurately matched with its geographical location; the target area is divided into multiple sub-areas, and the sub-areas are customized according to user needs and urban planning; in each sub-area, the number of bicycles of each brand is counted through data aggregation and classification counting, and the ratio of the number of shared bicycles of each brand to the area of ​​the sub-area is calculated to obtain the density value of the shared bicycle brand distribution; the density value reflects the brand distribution density of shared bicycles in each sub-area, and is an indicator for evaluating the rationality of shared bicycle distribution and the satisfaction of user needs; the operator adjusts the bicycle deployment and scheduling strategy according to the density value, optimizes the distribution of shared bicycles, and improves the user experience and the efficiency of shared bicycle use; by continuously monitoring and analyzing the bicycle distribution density, the operator can promptly discover and solve problems such as bicycle accumulation and uneven brand distribution, and realize more refined operation management.

[0020] Preferably, a multi-objective optimization model is constructed through the density value of the shared bicycle brand distribution, and three objective functions are generated, which are used to measure the balance of shared bicycle resource allocation, the excessive concentration of shared bicycles of a specific brand, and the traffic congestion caused by the uneven distribution of shared bicycles.

[0021] Preferably, the first objective function Used to measure the balance of shared bicycle resource allocation, where x i,j represents the number of bicycles of brand i in region j, is the average number of shared bicycles of all brands in region j; the second objective function Used to avoid excessive concentration of a particular brand, where p i,j represents the proportion of shared bicycles of brand i in region j; the third objective function It is used to reduce traffic congestion caused by uneven distribution of shared bicycles, where dj represents the bicycle density in area j, is the global average density.

[0022] Preferably, the density values ​​of different brands of shared bicycles in the region, user needs and real-time traffic flow data are integrated to generate a relatively reasonable scheduling strategy. i,j Reduce its f1 and reduce the balance of shared bicycle resource allocation; by reducing p i,j Reduce its f2 and reduce the proportion of excessive concentration of certain specific shared bicycle brands; reduce d j This reduces f3 and reduces traffic congestion caused by uneven distribution of shared bicycles. For example, if the density of a certain brand of shared bicycles is relatively high, some shared bicycles need to be transferred away to achieve a balance among all brands of shared bicycles.

[0023] (III) Beneficial effects

[0024] The present invention provides a shared bicycle brand identification and density optimization method driven by the NSGA-III algorithm, which has the following beneficial effects:

[0025] 1. The present invention optimizes the distribution of shared bicycles through the NSGA-III algorithm, realizes the reasonable allocation of resources of shared bicycles of different brands, avoids the excessive concentration of shared bicycles of a specific brand in a specific area, and improves the utilization rate and scheduling efficiency of shared bicycles.

[0026] 2. The present invention reduces traffic congestion caused by uneven distribution of shared bicycles by dynamically adjusting the density of shared bicycles in each area, making urban traffic smoother.

[0027] 3. The present invention combines real-time data to generate optimized dispatching strategies, which can automatically respond to changes in urban traffic and user needs, enhance the intelligent management level of shared bicycles, and greatly improve operational results. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] Embodiment 1:

[0031] like Figure 1 As shown, an embodiment of the present invention provides a shared bicycle brand identification and density optimization method driven by the NSGA-III algorithm. The development of this method is first based on a large-scale data acquisition and processing system. In modern cities, tens of thousands of shared bicycles use monitoring equipment such as cameras and GPS locators to record the images, geographic locations and traffic flow data in the target area of ​​the shared bicycles in real time. And unified preprocessing operations are performed, including data cleaning, outlier removal and formatting, to ensure the accuracy and consistency of the data.

[0032] In the process of shared bicycle brand identification, the shared bicycle images captured by the monitoring equipment are first standardized using image processing technology to reduce the interference of environmental factors such as lighting, weather, and background complexity on the identification process; the standardized images are analyzed through a deep learning model to extract the brand features of the shared bicycles, which include the brand logo, color, pattern, etc. Brand features are crucial for shared bicycle brand identification and can help the deep learning model accurately identify shared bicycles of different brands in a variety of complex scenarios.

[0033] This method combines the deep learning model with NSGA-III. The NSGA-III algorithm is based on multi-objective optimization. Its multi-objective is to optimize multiple features during the recognition process, such as recognition speed, accuracy, and resource consumption. The NSGA-III algorithm starts with the initial optimization method, defines different parameters and feature extraction methods, and performs selection, crossover, and mutation operations through iteration to generate new individuals. In each round of iteration, the NSGA-III algorithm evaluates the performance of the current optimization method and retains the best performing optimization method through non-dominated sorting. As the iteration proceeds, the NSGA-III algorithm adjusts the image preprocessing steps and optimizes the structure and parameters of the deep learning model, ultimately enabling the NSGA-III algorithm to maintain efficient brand recognition capabilities in complex environments.

[0034] This method calculates the distribution density of shared bicycles of various brands in the target area based on the identified brand information and GPS location data. The core of density estimation is to divide the target area into multiple sub-areas. In each sub-area, the number of shared bicycles of various brands is calculated and compared with the area of ​​the area to obtain the distribution density value of the shared bicycle brand. The density value is an important indicator for evaluating the rationality of the distribution of shared bicycles, which directly affects the user experience and the fluidity of urban traffic.

[0035] Density optimization further improves the distribution of shared bicycles through the NSGA-III algorithm; the NSGA-III algorithm builds a multi-objective optimization model based on multiple optimization objectives, such as the distribution balance of shared bicycle brand resources, the concentration of shared bicycles of a specific brand, and traffic congestion; in each iteration, the NSGA-III algorithm evaluates the distribution of shared bicycles of each brand in different areas and adjusts the distribution according to the optimization objectives; the best performing solution set is identified through non-dominated sorting, and new solution sets are generated through multiple crossover and mutation operations, gradually approaching the optimal shared bicycle distribution plan; this process ensures a more balanced distribution of shared bicycle brands in the city, avoids excessive concentration of shared bicycles of a certain brand in certain areas, increases users' demand for shared bicycles, and reduces the risk of urban traffic congestion.

[0036] The optimized density distribution provides operators with a strategic basis for the scheduling of shared bicycles. In the urban transportation system, the real-time monitoring and data feedback mechanism ensures that operators can grasp the distribution status of shared bicycles in various regions at any time, and dynamically adjust the deployment and scheduling strategies of shared bicycles according to user needs and traffic conditions. For example, during peak hours, shared bicycles of certain brands may be over-concentrated in certain areas, while other areas may be short of shared bicycles. Based on the optimization results of the NSGA-III algorithm, shared bicycles are automatically dispatched to areas where shared bicycles are scarce to meet user needs and avoid traffic congestion caused by excessive concentration of shared bicycles.

[0037] The application of NSGA-III algorithm in density optimization not only improves the resource utilization of shared bicycles, but also improves the response speed of the method. In each iteration of the algorithm, it not only dispatches shared bicycles, but also predicts and analyzes future traffic patterns. By combining traffic flow data with the real-time distribution data of shared bicycles, the algorithm can predict the demand for shared bicycles in a certain period of time in the future, and dispatch them in advance to ensure smooth traffic in the city. Another important result of density optimization is to improve the utilization efficiency of brand resources. By dynamically adjusting the distribution of shared bicycles of different brands, the optimized method can enable more users to find suitable shared bicycles when needed, thereby improving the efficiency of shared bicycle use.

[0038] After multiple rounds of development and testing, the shared bicycle brand identification and density optimization method driven by the NSGA-III algorithm has achieved efficient brand identification, accurate shared bicycle density estimation, and intelligent density optimization scheduling; it not only improves the operational efficiency of shared bicycles, but also improves the user experience by reducing urban traffic congestion and improving the rational allocation of shared bicycle resources, providing technical support for urban intelligent traffic management.

[0039] Embodiment 2:

[0040] In a large city, the shared bicycle brand identification and density optimization method driven by the NSGA-III algorithm has been actually applied to the daily management of shared bicycles.

[0041] First, multi-source heterogeneous data collection equipment, including high-resolution cameras, GPS locators, and traffic flow monitoring equipment, have been deployed in major blocks, commercial areas, and residential areas in the city. The monitoring equipment has established a huge shared bicycle management database by capturing real-time images of shared bicycles on the streets, the locations of shared bicycles, and surrounding traffic flow data.

[0042] In daily use, the method first analyzes the captured shared bicycle images in real time through image recognition technology to identify the brand, model and specific location of each shared bicycle in the city. This process combines deep learning algorithms and NSGA-III optimization algorithms to ensure accurate identification of shared bicycles of different brands in complex urban environments; for example, in a busy commercial area, the method can quickly identify shared bicycles of different brands through optimized image recognition algorithms, providing efficient brand recognition results.

[0043] After the method identifies the locations of shared bicycles of various brands, it calculates the distribution density of shared bicycles in each area in real time. For dense commercial areas or transportation hubs, the method continuously monitors the distribution of shared bicycles to ensure that shared bicycles are not over-concentrated in a certain area, thereby affecting traffic flow. The method uses the multi-objective optimization model in the NSGA-III algorithm to dynamically adjust the deployment and scheduling strategies of shared bicycles according to the distribution balance of shared bicycle brand resources, the concentration of shared bicycles of a specific brand, and traffic congestion. For example, when it is detected that shared bicycles of a certain brand are over-concentrated in a certain area, a scheduling instruction will be immediately generated to deploy some shared bicycles to other areas with higher demand but fewer shared bicycles. This scheduling process is completed automatically without human intervention, which greatly improves operational efficiency.

[0044] This method performs particularly well during peak hours in the morning and evening. During peak hours on weekdays in cities, especially near subway stations, shared bicycles of certain brands tend to gather in large numbers in a short period of time, making it difficult for users to find shared bicycles of other brands, which usually causes traffic jams. The NSGA-III algorithm analyzes user demand data and the real-time distribution of shared bicycles to optimize the scheduling strategy of shared bicycles. This method automatically deploys over-concentrated shared bicycles to areas with higher demand, avoiding the accumulation of shared bicycles, while also improving the utilization rate of shared bicycles of other brands.

[0045] The method continuously collects real-time traffic data through traffic flow monitoring equipment, combines it with the distribution density of shared bicycles for analysis, and predicts future traffic conditions. For example, during large-scale events or holidays, the method predicts the increase in demand for shared bicycles in certain areas in advance, and deploys shared bicycles in advance to ensure that users can still easily find shared bicycles during peak hours. In this case, the predictive function and scheduling ability of the method greatly alleviate traffic pressure and make urban traffic smoother.

[0046] Through practical application, the method has demonstrated excellent results; the distribution of shared bicycles of different brands in the city has become more balanced, and there is no longer an excessive concentration of shared bicycles of a specific brand; because the method can dispatch shared bicycles, users can more easily find the required brand shared bicycles in various areas, which improves user satisfaction; at the same time, because the optimized distribution of shared bicycles reduces interference with urban traffic, traffic congestion has been effectively alleviated and urban traffic mobility has been significantly improved.

[0047] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A shared bicycle brand recognition and density optimization method driven by NSGA-III algorithm, characterized in that: The following steps are involved: S1. Obtaining real-time multi-source heterogeneous data of the target area where the shared bicycles are located through monitoring equipment, wherein the multi-source heterogeneous data includes images and locations of the shared bicycles and traffic flow data of the target area; S2. Based on the image data in the multi-source heterogeneous data, the feature data of the shared bicycle brand is extracted using an image recognition algorithm, and the image recognition algorithm is combined with the NSGA-III algorithm to perform multi-feature optimization to improve the accuracy of identifying different shared bicycle brands in a complex environment; S3. Calculate the density value of each shared bicycle brand distribution in the target area based on the brand recognition results and the collected shared bicycle location data; S4. The density of shared bicycles is optimized using the NSGA-I II algorithm. A multi-objective optimization model is generated based on the density values ​​of the shared bicycle brand distribution. Three objective functions are generated based on the multi-objective optimization model to optimize multiple objectives, where the multiple objectives include: The balance of resource allocation of shared bicycles of different brands; Avoid excessive concentration of shared bikes of a particular brand; Reduce traffic congestion caused by uneven distribution of shared bicycles; S5. Based on the multi-objective optimization model, generate a shared bicycle dispatching strategy to dispatch shared bicycles of different brands to achieve reasonable allocation of shared bicycles of different brands and smooth traffic.

2. The shared bicycle brand identification and density optimization method based on NSGA-III algorithm drive according to claim 1 is characterized by: By deploying monitoring equipment in the target area, multi-source heterogeneous data of shared bicycles can be obtained in real time. The monitoring equipment includes cameras and GPS. Multi-source heterogeneous data is collected by monitoring equipment to understand the distribution of shared bicycles in the target area.

3. The shared bicycle brand identification and density optimization method based on NSGA-III algorithm drive according to claim 1 is characterized by: In the brand recognition of shared bicycles, image data is standardized through image processing technology, and the processed images are analyzed through deep learning models to extract the unique features of shared bicycle brands, including logos, colors, and design patterns; The image recognition algorithm is combined with the NSGA-III algorithm for multi-feature optimization.

4. The method for shared bicycle brand identification and density optimization based on NSGA-III algorithm according to claim 1 is characterized in that: During the optimization process, multiple features are optimized, including improving recognition accuracy, reducing false recognition rate, and improving the real-time and robustness of the NSGA-III algorithm. The NSGA-III algorithm searches for the optimal solution set through iterative optimization, and the solution set represents the recognition that achieves the best balance between different objectives. The NSGA-III algorithm continuously adjusts feature extraction and recognition to adapt to brand recognition of shared bicycles in complex environments.

5. The shared bicycle brand identification and density optimization method based on NSGA-III algorithm drive according to claim 1 is characterized by: The brand recognition results of shared bicycles are matched with the GPS location data of shared bicycles to accurately match the brand information of each shared bicycle with its geographical location; the target area is divided into multiple sub-areas, and the sub-areas are customized based on the actual needs of users; for each sub-area, the number of shared bicycles of each brand is counted through data aggregation and classification counting. After the statistics are completed, the ratio of the number of shared bicycles of each brand in each sub-area to the area of ​​the sub-area is calculated to obtain the density value of the shared bicycle brand distribution.

6. The method for shared bicycle brand identification and density optimization based on NSGA-III algorithm according to claim 1 is characterized in that: By calculating the density value of the shared bicycle brand distribution in the target area, a multi-objective optimization model is constructed and three objective functions are generated; the three objective functions are used to optimize multiple objectives; each objective function calculates the distribution of shared bicycles in each area and obtains the optimization values ​​of its multiple objectives; through non-dominated sorting, the optimization values ​​are sorted according to their performance in the objective function, so as to identify the optimization values ​​that are not dominated by other solutions and form a solution set; in this process, crossover and mutation operations are performed to generate new solutions and help explore a better solution set; Through multiple iterations, the NSGA-III algorithm gradually approaches the true optimal solution set and finally obtains a set of optimization solutions.

7. The method for shared bicycle brand identification and density optimization based on NSGA-III algorithm according to claim 6 is characterized in that: The first objective function Used to measure the balance of shared bicycle resource allocation, where x i,j represents the number of shared bicycles of brand i in region j, is the average number of shared bicycles of all brands in region j; the second objective function It is used to avoid excessive concentration of shared bicycles of a specific brand, where p i,j represents the proportion of shared bicycles of brand i in region j; the third objective function It is used to reduce traffic congestion caused by uneven distribution of shared bicycles, where d j represents the shared bicycle density in region j, is the global average density; where M is the sub-region and N is the target region.

8. The method for shared bicycle brand identification and density optimization based on NSGA-III algorithm according to claim 7 is characterized in that: Integrate the density values, user demand and traffic flow data of different brands of shared bicycles in the target area, adjust the optimization values ​​of objective functions f1, f2 and f3 through a multi-objective optimization model, and generate a relatively reasonable scheduling strategy.

Images