A road network traffic bottleneck identification method integrating wavelet transform and image binarization

Abstract

The present invention discloses a road network traffic bottleneck identification method integrating wavelet transform and image binarization. The main steps include: inputting speed time series data, searching for speed mutation points according to the wavelet transform principle, and outputting the time and position of the speed mutation points; secondly, sorting the speed values ​​of the speed mutation points in ascending order, selecting the speed threshold according to the quartile method, and converting the speed time series data into a speed spatiotemporal heat map; then, setting the speed threshold as a parameter for image binarization, and dividing the speed spatiotemporal heat map into congested areas and non-congested areas through image binarization; finally, for the irregularly scattered congested areas in the binarized image, considering that they have little impact on the overall congestion of the traffic road network, using median filtering to eliminate irregular, sporadically distributed pixels and small bottleneck points, so as to improve the calculation efficiency and recognition accuracy of frequent traffic bottlenecks in the urban road network.

Description

Technical Field

[0001] The present application relates to the field of traffic information, and in particular to a method for identifying road network traffic bottlenecks by combining wavelet transform and image binarization. Background Art

[0002] Traffic congestion seriously affects the efficiency of urban operations and people's daily travel. Expressways have the characteristics of large capacity and high speed, and are often the first choice for travelers because of their obvious effect on traffic improvement. However, expressway traffic congestion is frequent due to reasons such as poor connection between the main and auxiliary roads, unreasonable entrance and exit traffic organization, mismatched traffic capacity, and imperfect surrounding supporting road networks.

[0003] In practice, GPS data is particularly important for the study of traffic congestion. By processing, analyzing and mining massive GPS data, road network operation status information can be obtained, and road section operation speed analysis can be realized. Traffic management departments in major domestic cities, such as Beijing, Shanghai, Guangzhou, Shenzhen, Wuhan, Hangzhou, etc., have established urban road traffic operation analysis platforms based on local GPS data. Internet companies such as AutoNavi and Didi Chuxing also use GPS data for traffic information services and road operation evaluation. As one of the earliest cities to carry out this work, Beijing continues to track and evaluate road traffic changes with indicators such as traffic congestion index and road network operation speed, providing quantitative evaluation methods for traffic planning and management. At the same time, based on the in-depth mining of GPS data, it is possible to identify the frequently congested sections of expressways during peak hours and analyze the congestion time-space cloud map, further screen the congested sections and congestion bottlenecks of the urban expressway network, and provide technical support for urban congestion relief projects.

[0004] At present, bottleneck identification methods mainly include two categories: 1) bottleneck identification based on fixed point detectors; 2) bottleneck identification based on mobile detection. However, the coverage of trajectory data obtained by fixed point detectors is often very limited, and the installation and maintenance costs are high, while mobile detection can effectively obtain trajectory data over a large range, and often judges and identifies traffic bottlenecks based on traffic indicators such as delay time, speed difference, congestion range, and congestion intensity. In order to accurately identify traffic bottlenecks, the amount of data required is too large, and bottleneck identification often cannot be performed quickly. Bottleneck identification based on speed spatiotemporal graphs is more intuitive and convenient for observing the characteristics and changing trends of traffic congestion. Using the function of image binarization to distinguish between congested areas and non-congested areas and then identifying traffic bottlenecks is a relatively new method in recent years. One of the key factors in identifying bottlenecks based on image binarization is to determine the speed threshold. Only on the basis of selecting a suitable threshold can traffic bottlenecks be accurately and effectively identified. However, most previous studies set speed thresholds based on experience to distinguish congested sections and non-congested sections. This method of setting speed thresholds is relatively subjective, resulting in corresponding errors in the identified traffic bottlenecks. Urban traffic networks are complex and changeable, and a single threshold set based on empirical values ​​often cannot meet the needs of dynamic and accurate bottleneck identification. Therefore, how to scientifically and reasonably set the speed threshold for image binarization has become a key issue in urban road network bottleneck identification. Summary of the invention

[0005] 1. Purpose of the Invention

[0006] Aiming at the problem that current bottleneck identification lacks precision and reliability, the present invention proposes a road network traffic bottleneck identification method integrating wavelet transform and image binarization to achieve the goal of quickly and accurately identifying bottlenecks on urban expressways.

[0007] 2. Technical solution adopted by the present invention

[0008] The bottleneck identification method based on wavelet transform and image binarization proposed in the present invention can be implemented by the following steps:

[0009] (1) Sort the original data, retain the valid data fields, and remove the invalid data. Input the processed speed time series data, find the speed mutation point according to the wavelet transform principle, and output the time and position of the speed mutation point.

[0010] (2) According to the speed mutation points output in step (1), the speed values ​​are sorted in ascending order, the speed threshold is selected according to the quartile method, and the speed time series data is converted into a speed spatiotemporal heat map.

[0011] (3) The speed threshold determined in step (2) is used as a parameter for image binarization, and the speed spatiotemporal heat map is divided into a congested area and a non-congested area by image binarization. The part with a speed greater than the threshold is regarded as a non-congested area, and the part with a speed less than the threshold is regarded as a congested area, that is, the heat map is converted into a binary image to increase the calculation speed.

[0012] (4) For irregularly scattered congested areas in the binary image, median filtering is used to remove irregular, sporadic pixel points and small bottleneck points to improve the calculation efficiency and recognition accuracy of common traffic bottlenecks in urban road networks.

[0013] The step (1) is specifically as follows: Wavelet transform (WT) is an extension of Fourier transform (FT). Wavelet analysis theory can realize the localized description of the signal in time and frequency, and provide a powerful tool for the analysis of signal mutation points. Mutation points usually refer to signal mutations under normal signal conditions. In dynamic systems, signal mutations are relatively fast. The main feature of signal mutations is that the signal has local changes in time and space. According to the speed of signal change, select the appropriate decomposition scale wavelet analysis. Assume that the h(t) function is the convolution of the functions f(t) and g(t), that is:

[0014] h(t)=f(t)×g(t)

[0015] According to the properties of Fourier transform, we can get:

[0016]

[0017] You can get: h ′ (t) = f ′ (t)×g(t)=f(t)×g′(t)

[0018] If the function f(t) is regarded as a signal and g(t) is regarded as a filter, then the convolution result of the derivative of the signal and the filter can be regarded as the convolution of the derivative of the filter and the signal. Therefore, the mutation points and extreme points of the wavelet transform have a corresponding relationship with the mutation points and extreme points of the signal f(t), and the wavelet transform can be used to detect mutation signals. The specific process is as follows:

[0019] Assume that θ(t) is a low-pass stationary function that acts as a smoothing function and satisfies the condition

[0020]

[0021] Usually θ(t) is taken as a Gaussian function, that is

[0022]

[0023] Assume that θ(t) is quadratically differentiable and define

[0024]

[0025] Then the function Satisfy the admissibility conditions of wavelet:

[0026]

[0027] Therefore, it can be used as a wavelet mother function. The method of detecting signal mutation points based on wavelet analysis generally performs multi-scale analysis on the signal. When the signal mutates, its wavelet transformed coefficient has the maximum wavelet energy. Therefore, the time of the mutation point can be determined by detecting the maximum wavelet energy. The detection content of the signal mutation point includes the time of the mutation point, the type of mutation point and the change of amplitude.

[0028] The valid fields in the original data, such as speed, longitude and latitude, and azimuth, are retained. After removing abnormal data, the data is gridded to draw a time-space speed heat map. The speed time series data is input, and after one-dimensional continuous wavelet transform processing, the wavelet energy and the time and space coordinates of the speed mutation point are output. The location where the wavelet energy peaks is the location of the speed mutation point.

[0029] The step (2) is specifically as follows: considering the actual error, the speed mutation point is matched to the speed space-time diagram to verify its accuracy. After verification, it can be concluded that the wavelet energy value of the true speed mutation point is much greater than the false speed mutation point detected due to data error, so a reasonable wavelet energy threshold is selected to screen out the effective speed mutation point. After screening, the speeds of all speed mutation points are arranged in order from small to large. Quartiles are also called quartile points, which divide all data into four equal parts, each of which includes 25% of the data, and the values ​​at each quartile are quartiles. There are three quartiles, the first quartile is what is commonly called the quartile, called the lower quartile, the second quartile is the median, and the third quartile is called the upper quartile, represented by Q1, Q2, and Q3 respectively. The upper quartile is taken as the speed threshold for image binarization to distinguish between congested areas and non-congested areas.

[0030] The step (3) is specifically as follows: Binarization is to make the grayscale value of each pixel in the pixel matrix of the image 0 (black) or 255 (white), so that the amount of data in the image is greatly reduced, so that the outline of the target can be highlighted, that is, the entire image presents only black and white effects. The grayscale value range in the grayscale image is 0 to 255, and the grayscale value range in the image after binarization is 0 or 255. In the speed space-time map, cells greater than the speed threshold will become white, and cells less than the speed threshold will become black. Therefore, the speeds in different cells are all converted to 0 and 1 after binarization. First, the speed space-time heat map is converted into a grayscale map. Then, binarization is performed, the speed threshold is set, and the grayscale map is converted into a binary map.

[0031] The step (4) is specifically as follows: white represents areas with good traffic conditions, and black represents areas with traffic congestion. It is incorrect to regard the irregularly scattered congested areas in the binary image as bottlenecks. Median filtering is a nonlinear signal processing technology based on sorting statistics theory that can effectively suppress noise. Its basic principle is to replace the value of a point in a digital image or digital sequence with the median of the values ​​of each point in a neighborhood of the point, so that the surrounding pixel values ​​are close to the true value, thereby eliminating isolated noise points. It is very effective in smoothing impulse noise. At the same time, it can protect the sharp edges of the image, select appropriate points to replace the values ​​of noise points, and perform well for salt and pepper noise. Using median filtering, irregular and sporadic small bottleneck points are eliminated to improve the calculation efficiency and recognition accuracy of frequent traffic bottlenecks in urban road networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0033] Figure 1 It is a fast road network diagram according to a specific embodiment of the present disclosure;

[0034] Figure 2 It is a time-space velocity heat map according to a specific embodiment of the present disclosure;

[0035] Figure 3 It is a flowchart according to a specific embodiment of the present disclosure;

[0036] Figure 4 It is a time-space velocity diagram after binarization processing according to a specific embodiment of the present disclosure;

[0037] Figure 5 It is a binarized spatiotemporal velocity image processed by median filtering according to the specific implementation mode of the present disclosure. DETAILED DESCRIPTION

[0038] 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, rather than all the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present invention and its application or use. 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.

[0039] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0040] Unless otherwise specifically stated, the relative arrangement of the parts and steps described in these embodiments, numerical expressions and numerical values ​​do not limit the scope of the present invention. At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the accompanying drawings are not drawn according to the actual proportional relationship. The technology, methods and equipment known to ordinary technicians in the relevant field may not be discussed in detail, but in appropriate cases, the technology, methods and equipment should be regarded as a part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a limitation. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once a certain item is defined in an accompanying drawing, it does not need to be further discussed in subsequent drawings.

[0041] The road network traffic bottleneck identification method based on integrated wavelet transform and image binarization proposed by the present invention can be implemented by the following steps:

[0042] (1) Sort the original data, retain valid data fields, and remove invalid data. Input the processed speed time series data, find the speed mutation point according to the wavelet transform principle, and output the time and position of the speed mutation point;

[0043] (2) According to the speed mutation points output in step (1), the speed values ​​are sorted in ascending order, the speed threshold is selected according to the quartile method, and the speed time series data is converted into a speed spatiotemporal heat map;

[0044] (3) The speed threshold determined in step (2) is used as a parameter for image binarization, and the speed spatiotemporal heat map is divided into a congested area and a non-congested area by image binarization. The part with a speed greater than the threshold is regarded as a non-congested area, and the part with a speed less than the threshold is regarded as a congested area. That is, the heat map is converted into a binary image to increase the calculation speed and facilitate analysis;

[0045] (4) For irregularly scattered congested areas in the binary image, median filtering is used to remove irregular, sporadic pixel points and small bottleneck points to improve the calculation efficiency and recognition accuracy of common traffic bottlenecks in urban road networks.

[0046] The following is a detailed introduction to the steps of implementing the method based on actual GPS data.

[0047] Step (1): Study the road sections such as Figure 1 As shown in the figure. The length of the road section is 32km. The data used covers the entire road network of Beijing, including the floating vehicle GPS trajectory data from January 1 to February 14, 2015, a total of 45 days. The GPS trajectory data contains 14 different fields of information, and the data types include trajectory ID, terminal acquisition time, longitude and latitude coordinates, etc. First, the data is preprocessed to prevent data errors from interfering with the recognition results. Secondly, the GPS trajectory data is converted into spatiotemporal grid data, and the highly discrete GPS trajectory points are converted into continuous spatiotemporal speed matrices, which facilitates the systematic disclosure of traffic status.

[0048] Step (2): According to the principle of wavelet transform to check singular points, that is, the peak position of wavelet energy corresponds to the position of speed mutation point. Input time speed series data, detect speed mutation points, and output the time and space coordinates of speed mutation points. In order to improve the accuracy of speed mutation points, check whether the speed mutation points are in the bottleneck area, that is, the time and space coordinates of the speed mutation points are calibrated in the speed time and space diagram. After inspection, it was found that due to data errors, the detected speed mutation points were not accurate enough. A small number of speed mutation points appeared in places with smooth traffic, which was inconsistent with the facts, and the wavelet energy values ​​of these "false" speed mutation points were much smaller than the "true" speed mutation points. Therefore, the threshold of wavelet energy was set to improve the accuracy of speed mutation points. Here, the threshold of wavelet energy was set to 12000, that is, speed mutation points with wavelet energy less than 12000 were eliminated. After the screening, all the remaining speed mutation points were arranged in order of speed from small to large, and the upper quartile of the speed value was taken as the threshold of image binarization. Here, according to the method of this paper, the upper quartile of the speed mutation point was set to 33km / h.

[0049] Step (3): Image binarization: By setting a speed threshold, the spatiotemporal speed heat map can be converted into a binary image, that is, the color of the pixels with speed greater than the threshold is changed to white, and the color of the pixels with speed less than the threshold is changed to black. The white area is the area with smooth traffic, and the black area indicates traffic congestion. Figure 4 This is a space-time speed map after image binarization. The black part is the congested area, and the white part is the non-congested area. The binarized space-time speed map is more intuitive and clear than the heat map.

[0050] Step (4): The binarized spatiotemporal speed map shows large congested areas and scattered small bottlenecks. In view of the definition of traffic bottlenecks, this paper considers that the impact of local small bottlenecks on traffic is small and is not enough to be considered as bottlenecks. The scattered small bottlenecks meet the characteristics of salt and pepper noise, so median filtering is used to remove these small bottlenecks to improve recognition efficiency and accuracy. Figure 5 This is the space-time speed map after median filtering. It can be clearly seen that the bottleneck area with high congestion intensity is retained in the map, and the scattered black pixels are filtered out, making the overall recognition result clearer.

[0051] In this embodiment, a traffic bottleneck identification method based on wavelet transform and image binarization of GPS trajectory data set is utilized. The results show that this method can better identify traffic bottlenecks on urban expressways. It is intuitive and effective, and is convenient for comparing and analyzing the characteristics and changing trends of traffic bottlenecks in time and space dimensions.

Claims

1. A road network traffic bottleneck identification method integrating wavelet transform and image binarization, characterized in that: The following steps are involved: Step 1: Input speed time series data, find speed mutation points according to the wavelet transform principle, and output the time and position of speed mutation points; Step 2: Sort the speed values ​​of the speed mutation points in ascending order, select the speed threshold according to the quartile method, and convert the speed time series data into a speed spatiotemporal heat map; Step 3: Set the speed threshold determined in step 2 as the parameter of image binarization, and divide the speed spatiotemporal heat map into congested areas and non-congested areas through image binarization; Step 4: For irregularly scattered congested areas in the binary image, use median filtering to remove irregular, sporadic pixels and small bottleneck points to improve the calculation efficiency and recognition accuracy of common traffic bottlenecks in urban road networks.

2. The method according to claim 1, characterized in that The step 1 includes data sorting of the original data, retaining valid data fields, eliminating invalid data, inputting processed speed time series data, finding the speed mutation point according to the position where the wavelet energy peak appears, that is, the position of the speed mutation point, and outputting the time and position of the speed mutation point.

3. The method according to claim 1, characterized in that Step 2 sets the wavelet energy threshold to detect the authenticity of the speed mutation point. After screening, the speeds of all speed mutation points are arranged in ascending order, and the upper quartile is taken as the speed threshold for image binarization to distinguish between congested areas and non-congested areas.

4. The method according to claim 1, characterized in that The speed spatiotemporal heat map is converted into a grayscale image, and then binarized. The speed threshold is set and the grayscale image is converted into a binary image, that is, the entire image is presented in black and white. The white area is the area with smooth traffic, and the black area is the congested area.

5. The method according to claim 1, characterized in that For the irregularly scattered congested areas in the binary image, median filtering is used to remove irregular and sporadic small bottleneck points to improve the calculation efficiency and recognition accuracy of common traffic bottlenecks in urban road networks.

Images