An automatic traffic flow energy state recognition method based on unmanned aerial vehicle aerial video

Abstract

The application relates to a traffic flow energy state automatic identification method based on unmanned aerial vehicle aerial video. The method acquires an unmanned aerial vehicle aerial image; image mapping and coordinate conversion methods are used to convert the unmanned aerial vehicle aerial image into actual track data; based on the actual track data and a pre-divided field source level, corresponding continuous energy field strength and discrete energy field strength of each level are acquired; based on the corresponding continuous energy field strength and discrete energy field strength of each level, according to electromagnetic field theory and artificial potential field theory, the environmental influence of a target vehicle is acquired by using a pre-constructed comprehensive field strength calculation model, and then the corresponding traffic flow energy state is obtained. Compared with the prior art, the application has the advantages of effectively improving the accuracy of vehicle traffic flow energy state automatic identification in an interlaced section and the like.

Description

Technical Field

[0001] This invention relates to the field of transportation engineering technology, and in particular to an automatic identification method for traffic flow energy state based on drone aerial video. Background Technology

[0002] Vehicle traffic patterns near urban expressway entrances and exits are highly variable, often involving lane changes and frequent acceleration and deceleration. In cities with scarce land resources, expressways often adopt a "first-in, last-out" pattern, where vehicles on the main line exit the ramp first and then merge onto the next ramp within 200 meters or less. Therefore, vehicle traffic patterns within weaving sections are characterized by "intersecting trajectories and frequent interactions." Consequently, weaving sections have numerous conflict points and are more prone to rear-end collisions and side-impact accidents. To effectively prevent accidents, monitoring traffic flow and identifying energy states within weaving sections is crucial.

[0003] Existing methods for detecting the energy state of vehicle traffic flow in weaving sections typically employ 2-3 drones simultaneously capturing images, which are then stitched together to obtain a video of the entire weaving section. The energy state of the traffic flow in the weaving section is then identified based on the stitched video. However, this method suffers from unavoidable errors due to image stitching and rotation, severely impacting the accuracy of the identification. Therefore, improving the accuracy of automatic identification of the energy state of vehicle traffic flow in weaving sections has become a problem that needs to be solved in this field. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology, such as large errors and low recognition accuracy, and to provide an automatic traffic flow energy state recognition method based on UAV aerial video.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] This invention provides an automatic traffic flow energy state identification method based on UAV aerial video, comprising the following steps: acquiring UAV aerial images; converting the UAV aerial images into actual trajectory data using image mapping and coordinate transformation methods; based on the actual trajectory data and pre-divided field source levels, obtaining the corresponding continuous energy field strength and discrete energy field strength for each level; based on the corresponding continuous energy field strength and discrete energy field strength for each level, and according to electromagnetic field theory and artificial potential field theory, using a pre-constructed comprehensive field strength calculation model to obtain the environmental influence of the target vehicle, thereby obtaining the corresponding traffic flow energy state.

[0007] As a preferred technical solution, the UAV aerial image is converted into actual trajectory data using an image mapping and coordinate transformation method, specifically including: acquiring an orthophoto map; mapping the UAV aerial image to the orthophoto map; and using a coordinate transformation matrix to convert the pixel coordinates in the UAV aerial image into actual coordinates in the orthophoto map.

[0008] As a preferred technical solution, the coordinate transformation matrix includes:

[0009] [P' x ,P' y ,1]=M·[P x ,P y ,1]

[0010]

[0011] In the formula, P x P y To represent the pixel coordinates in the drone aerial image before transformation, P′ x 、P′ y The coordinates are the transformed coordinates, θ is the rotation angle of the coordinate system, and a, b, c”, and d are matrix transformation control parameters.

[0012] As a preferred technical solution, the field source hierarchy includes a facility layer, a rule layer, and a motion layer.

[0013] As a preferred technical solution, the field sources of the facility layer include the road basic environment of the weaving section, the field sources of the rule layer include control measures to guide driver behavior, and the field sources of the motion layer include the vehicle status within the weaving section.

[0014] As a preferred technical solution, the expression for the continuous energy field strength is:

[0015]

[0016] In the formula, E c For continuous energy field strength, c is a preset parameter based on the field source type, λ is the field strength entropy parameter per meter, and M... c The equivalent mass of different types of continuous energy, x is the lateral distance from the vehicle to the obstacle or lane line, ε is the ratio of the width of this lane to the standard lane, m c For the actual physical mass of different energy sources, v c This represents the average speed.

[0017] As a preferred technical solution, the expression for the discrete energy field strength is:

[0018]

[0019] In the formula, Ed For discrete energy field strength, c' is a parameter preset according to the field source type, M d The equivalent mass of the energy source is ω, where ω is the weighting coefficient and θ is the weighting coefficient. r α1 and α2 are the angle between the velocity direction and the energy source, r is the distance from the vehicle to the discrete energy source, and α1 and α2 are undetermined parameters of the discrete energy field strength.

[0020] As a preferred technical solution, the expression for the comprehensive field strength calculation model is:

[0021]

[0022] In the formula, m ego Indicates the mass of this vehicle, Δ c Δ represents the proportion of spatial movement of the vehicle relative to the continuous field source per unit time. d This represents the proportion of spatial motion of the vehicle relative to the ramp or surrounding vehicles per unit time. α3 and α4 are undetermined parameters for calculating the overall field strength.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. The method provided by this invention, when using only one drone to take aerial photos, matches the video images obtained by the drone with a high-precision orthophoto map, and maps the aerial images to the orthophoto map to obtain high-precision vehicle behavior data. This solves the problem of the accuracy of the original image and avoids the original error in the data acquisition stage. Based on this, the energy state identification can be carried out, which can effectively improve the accuracy of automatic identification of the energy state of vehicle traffic flow in weaving sections.

[0025] 2. The method provided by this invention divides a typical interlacing segment scene into three layers: facility layer, rule layer, and motion layer. Dividing the scene into facility layer, rule layer, and motion layer helps researchers to analyze and understand the various aspects of the scene more comprehensively. Through this layered approach, the various factors affecting the magnitude of the field strength in the scene can be more clearly identified and described, which helps to deeply explore the connotation and characteristics of the scene.

[0026] 3. The method provided by this invention further optimizes the energy field model for the characteristics of intersecting sections, and constructs a comprehensive field strength calculation model for continuous facilities and discrete objects. Compared with existing field strength calculation models, this invention hierarchically divides the field source types and establishes a more complete "field strength-pressure-energy" model, which makes the calculation more accurate and thus achieves more accurate automatic identification of the energy state of the entire road section. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the method of the present invention. Detailed implementation manners

[0028] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented on the premise of the technical solution of the present invention, and provides detailed implementation manners and specific operation processes. However, the protection scope of the present invention is not limited to the following embodiments.

[0029] Embodiment

[0030] As Figure 1 shown, this embodiment provides an automatic recognition method for the energy state of traffic flow based on UAV aerial video, and the specific steps are as follows:

[0031] Step S1: Control a UAV to take high-altitude photos to obtain aerial videos and images;

[0032] Step S2: Match the aerial images with a high-precision map of an orthophoto image, map the aerial images to the orthophoto image map to obtain high-precision vehicle weaving behavior data, including vehicle coordinates, trajectories, speeds, and lane-changing directions, etc. Combine the pixel coordinates in the aerial video images with the orthophoto image through coordinate transformation to obtain actual trajectory data. Among them, speed and acceleration can be used for subsequent calculation of the magnitude of the motion layer field strength. The coordinate transformation matrix is as follows:

[0033] [P’ x ,P’ y ,1] = M · [P x ,P y ,1] (1)

[0034]

[0035] In the formula, θ is the rotation angle of the coordinate system, P x ,P y is the video pixel coordinate before conversion, P′ x ,P′ y is the coordinate after conversion, and a, b, c”, and d are all matrix transformation control parameters. Among them, a controls the scaling ratio on the x-axis; b controls the rotation and shear on the x-axis; c” controls the rotation and shear on the y-axis; d controls the scaling ratio on the y-axis. The combination of these coefficients can achieve different types of coordinate transformations such as translation, rotation, scaling, and shear. The specific control process is as follows:

[0036] a: Control the scaling ratio on the x-axis. When a > 1, it means that the scaling on the x-axis direction becomes larger; when 0 < a < 1, it means that the scaling on the x-axis direction becomes smaller; when a = 1, it means that there is no scaling on the x-axis direction.

[0037] b: Controls rotation and shear on the x-axis. When b≠0, it represents rotation and shear in the x-axis direction; when b = 0, it represents no rotation and shear in the x-axis direction.

[0038] c": Controls rotation and shear on the y-axis. When c"≠0, it represents rotation and shear in the y-axis direction; when c" = 0, it represents no rotation and shear in the y-axis direction.

[0039] d: Controls the scaling ratio on the y-axis. When d>1, it represents an increase in scaling in the y-axis direction; when 0<d<1, it represents a decrease in scaling in the y-axis direction; when d = 1, it represents no scaling in the y-axis direction.

[0040] Step S3: Hierarchically classify the field source types of the vehicle energy field. According to the energy source, it can be divided into the facility layer, the rule layer, and the motion layer.

[0041] Facility layer: The road foundation environment of the weaving section, including guardrails, ramps, median strips, etc.

[0042] Rule layer: The control measures that guide the driver's behavior, including speed limit signs, lane lines, channelizing lines, variable message signs, etc.

[0043] Motion layer: Refers to the vehicle state information within the weaving section, including vehicle type, speed, acceleration, position, etc. 0]]

[0044] Step S4: Calculate the field strength magnitude corresponding to each layer respectively. The formula for the continuous energy field strength is:

[0045]

[0046]

[0047] In the formula, λ is the field strength entropy parameter per meter, M c is the equivalent mass of different types of continuous energy sources, x is the lateral distance from the vehicle to the obstacle or lane line, ε is the ratio of the width of this lane to the standard lane width, equivalent to the permittivity in electromagnetics, m c represents the actual physical mass of different energy sources, equivalent to the linear charge density, v c represents the average road speed. The value of c takes different values in different scenarios. The specific rules are as follows:

[0048] c = c b = 1 represents a continuous isolation facility;

[0049] c = c s = 1 represents a solid line;

[0050] c = c d = -1 represents a dashed line;

[0051] c = c sd =1 or -1 indicates a dashed or solid line;

[0052] Among them, c b c represents the gravity coefficient of the isolation facility. s The solid line represents the gravity coefficient, c. d The dotted line represents the gravity coefficient, c. sd This represents the gravity coefficient of the dashed and solid lines.

[0053] The formula for calculating discrete energy field strength is:

[0054]

[0055] In the formula, M d It is the equivalent mass of the energy source, θ r ω is the angle between the velocity direction and the energy source. r is the distance from the vehicle to the discrete energy source, α1 and α2 are undetermined parameters of the discrete energy field strength. ω is the weighting coefficient, L represents the number of ramp lanes, and a v This refers to the acceleration of surrounding vehicles. If the energy source is the ramp, ω = L; if the energy source is a vehicle, ω = a v c' = 1 indicates an entrance ramp, a prohibition sign, and surrounding vehicles; c' = -1 indicates an exit ramp.

[0056] Step S5: Based on electromagnetic field theory and artificial potential field theory, construct a comprehensive field strength calculation model to calculate the influence of the surrounding environment on a specific vehicle.

[0057]

[0058] In the formula, E c E represents the continuous field strength. d Denotes discrete field strength, m ego Indicates the mass of this vehicle, Δ c Δ represents the proportion of spatial movement of the vehicle relative to the continuous field source per unit time. d This represents the proportion of spatial motion of the vehicle relative to vehicles on the ramp or around it per unit time. α3 and α4 are undetermined parameters for calculating the overall field strength. ego Indicates the speed of this vehicle, a ego Let x represent the acceleration of the vehicle, x be the relative distance between the vehicle and the continuous energy source, r be the distance between the vehicle and the discrete energy source, and Δt be the unit of time.

[0059] The automatic traffic flow energy state identification method based on UAV aerial video provided in this embodiment only requires one UAV for high-altitude shooting, while previous studies typically used 2-3 UAVs for simultaneous shooting. This method effectively avoids image stitching to obtain traffic flow video of the entire weaving section, directly obtaining the complete traffic flow video of the weaving section and reducing errors. Furthermore, this method, through a comprehensive field strength calculation model, can more accurately calculate the influence of the surrounding environment on specific vehicles. These influences include road conditions, the impact of other vehicles, infrastructure, and changes in vehicle motion states. By monitoring these influences, automatic identification of the energy state of the entire road segment can be achieved.

[0060] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for automatic identification of traffic flow energy state based on drone aerial video, characterized in that, Includes the following steps: Acquire drone aerial images; The drone aerial images are converted into actual trajectory data using image mapping and coordinate transformation methods. Based on the actual trajectory data and the pre-divided field source levels, the corresponding continuous energy field strength and discrete energy field strength of each level are obtained respectively. Based on the continuous and discrete energy field strengths corresponding to each level, the environmental influence of the target vehicle is obtained by using a pre-constructed comprehensive field strength calculation model according to electromagnetic field theory and artificial potential field theory, thereby obtaining the corresponding traffic flow energy state. The expression for the continuous energy field strength is: In the formula, It is a continuous energy field with strong energy. c These are parameters preset based on the source type. λ The field strength entropy parameter per meter. For the equivalent quality of different types of continuous energy, x The lateral distance from the vehicle to an obstacle or lane line. ε This is the ratio of the width of this lane to the width of the standard lane. The actual physical mass of different energy sources. Average road speed; The field source hierarchy includes a facility layer, a rule layer, and a motion layer; The field sources of the facility layer include the basic road environment of the weaving section; the field sources of the rule layer include control measures that guide driver behavior; and the field sources of the motion layer include the vehicle status within the weaving section. The expression for the discrete energy field strength is: In the formula, It is a discrete energy field with strong intensity. c' These are parameters preset based on the source type. The equivalent mass of the energy source. These are the weighting coefficients. The angle between the velocity direction and the energy source. r The distance from the vehicle to the discrete energy source. and These are undetermined parameters for discrete energy field strength.

2. The method for automatic identification of traffic flow energy state based on UAV aerial video as described in claim 1, characterized in that, The drone aerial images are converted into actual trajectory data using image mapping and coordinate transformation methods, specifically including: Obtain orthophoto maps; Map the drone aerial images onto the orthophoto map; The pixel coordinates in the UAV aerial image are converted into the actual coordinates in the orthophoto map using a coordinate transformation matrix.

3. The method for automatic identification of traffic flow energy state based on UAV aerial video as described in claim 2, characterized in that, The coordinate transformation matrix includes: In the formula, To convert the pixel coordinates in the drone aerial image before transformation, The coordinates are the transformed coordinates. θ Let be the rotation angle of the coordinate system. a, b, c'', d These are all matrix transformation control parameters.

4. The method for automatic identification of traffic flow energy state based on UAV aerial video as described in claim 1, characterized in that, The expression for the comprehensive field strength calculation model is as follows: In the formula, This indicates the mass of the vehicle. This indicates the proportion of spatial movement of the vehicle relative to the continuous field source per unit time. This indicates the proportion of spatial movement of the vehicle relative to the ramp or surrounding vehicles per unit time. and These are the parameters to be determined for calculating the overall field strength.

Images