A method for calculating the spatial coordinate edge of obstacles based on roadside monocular vision

Abstract

The present invention relates to a method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision, which specifically includes the following steps: S1, scene calibration is performed according to the position information of a camera, and the image information taken by the camera and the position information of the obstacle are obtained; S2, a pixel coordinate system and a local plane coordinate system are respectively established according to the image information taken by the camera, and the pixel coordinates corresponding to the obstacle are calculated according to the position information of the obstacle, and converted into local plane coordinates; S3, the local plane coordinates of the obstacle are converted into world plane coordinates; S4, the world plane coordinates of the obstacle are converted into geographic coordinates, and the geographic coordinates are sent to the controller of the intelligent networked bus. Compared with the prior art, the present invention has the advantages of realizing accurate obstacle position reminders and pedestrian warnings, and improving the safety of the intelligent networked bus during driving.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent connected public transportation, and in particular to a method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision. Background Art

[0002] Currently, commonly used visual measurement technologies include monocular vision measurement, binocular vision measurement, and structured light vision measurement. Binocular vision can achieve higher accuracy by fusing information from multiple cameras, but its structure is relatively complex and the measurement time is relatively long. Structured light vision measurement has strong anti-interference ability and good real-time performance, but the calibration of the measurement system is relatively difficult. Monocular vision technology only requires one visual sensor and has the advantages of simple structure, easy mobility, and rapid data acquisition. It is a non-contact measurement technology and has been widely used in object point coordinate measurement.

[0003] Coordinate transformation based on monocular vision usually includes four coordinate systems: geographic coordinate system, world plane coordinate system, local plane coordinate system and pixel coordinate system. The geographic coordinate system is a spherical coordinate system that uses meridians, latitudes, longitudes and latitudes to represent ground points. The world plane coordinate system is defined in a two-dimensional plane. For practical needs, a functional relationship is established between the earth's ellipsoid and the plane to convert geographic coordinates into plane coordinates. The most commonly used projection coordinate system is the Gauss projection coordinate system, with units in meters. The local plane coordinate system is the plane coordinate system where the actual object is located. The origin is the location of the camera, and the unit is meters. The local plane coordinate system and the world plane coordinate system can be transformed into each other through translation and rotation. The pixel coordinate system refers to the coordinate system constructed based on the picture taken by the camera. The origin can be arbitrarily specified at a point in the picture, and the unit is pixel. Summary of the invention

[0004] The purpose of the present invention is to provide a method for calculating the spatial coordinate edge of obstacles based on roadside monocular vision in order to overcome the defects of the above-mentioned prior art. For the operation scene of intelligent networked buses, a mapping model is constructed to convert the pixel coordinates of obstacles in the pictures collected by the camera into the geographic coordinate system, that is, longitude and latitude. This point is transmitted to the intelligent networked bus as longitude and latitude to realize obstacle location reminder and pedestrian warning, and improve the safety of intelligent networked buses during driving.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision specifically comprises the following steps:

[0007] S1. Perform scene calibration according to the position information of the camera to obtain the image information taken by the camera and the position information of the obstacle;

[0008] S2. According to the image information taken by the camera, a pixel coordinate system and a local plane coordinate system are established respectively, and the pixel coordinates corresponding to the obstacle are calculated according to the position information of the obstacle, and converted into local plane coordinates;

[0009] S3, converting the local plane coordinates of the obstacle into world plane coordinates;

[0010] S4. Convert the world plane coordinates of the obstacle into geographic coordinates, and send the geographic coordinates to the controller of the intelligent network-connected bus. The controller controls the intelligent network-connected bus to avoid the obstacle according to the geographic coordinates of the obstacle.

[0011] The scene calibration in step S1 includes camera position calibration and feature point position calibration. The feature point position calibration process includes selecting multiple feature points at equal intervals on a plane within the camera monitoring range and obtaining the longitude and latitude of the feature points. The camera position calibration specifically includes the following steps:

[0012] S11, locate the latitude and longitude of the camera. Since the world plane coordinate system takes the north direction as the y-axis direction, it is necessary to obtain the angle θ between the camera and the north direction;

[0013] S12, obtaining the height of the camera, the horizontal distance between the nearest point of the image captured by the camera and the camera, and the actual distance of the camera's monitoring range.

[0014] The process of establishing the pixel coordinate system in step S2 includes taking the downstream horizontal line of the image information taken by the camera as the x-axis and the midpoint of the horizontal line as the origin to construct the pixel coordinate system, and the unit is pixel; the process of establishing the local plane coordinate system includes taking the camera as the origin and the horizontal line parallel to the x-axis of the pixel coordinate system as the x-axis to construct the local plane coordinate system, and the unit is meter.

[0015] Furthermore, step S2 further includes calibrating the feature points in the pixel coordinate system and the local plane coordinate system to obtain a first coordinate transformation relationship between the pixel coordinate system and the local plane coordinate system, and transforming the pixel coordinates of the obstacle into local plane coordinates according to the first coordinate transformation relationship. The calculation formula of the first coordinate transformation relationship is as follows:

[0016]

[0017] Among them, (u, v) is the coordinate value of the obstacle in the pixel coordinate system, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, F1 is the horizontal coordinate conversion function, and F0 is the vertical coordinate conversion function.

[0018] The conversion from local plane coordinates to world plane coordinates is a rigid body transformation. The object will not be deformed, and only coordinate rotation and translation are required. The process of converting the local plane coordinates to the world plane coordinates in step S3 includes coordinate rotation and coordinate translation. The second coordinate transformation relationship between the local plane coordinate system and the world plane coordinate system is obtained according to the coordinate rotation and coordinate translation.

[0019] Furthermore, the calculation formula for the coordinate rotation is as follows:

[0020]

[0021] Among them, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, (x′, y′) is the coordinate value of the obstacle in the world plane coordinate system, θ is the angle between the camera and the north direction, and R1 is the rotation matrix;

[0022] The coordinate translation process comprises the following steps:

[0023] S31, calibrating the longitude and latitude of the origin in the local plane coordinate system and the longitude and latitude of the origin in the world coordinate system;

[0024] S32, calculating the distance and azimuth between the origin in the local plane coordinate system and the origin in the world coordinate system;

[0025] S33. Calculate a translation matrix based on the distance and azimuth between the two points.

[0026] Furthermore, the calculation formula for the distance between the two points is as follows:

[0027] D(P,P 0 )=R*arccos(C)*π / 180

[0028] Among them, D(P,P 0 ) is the distance between the two points, P is the origin of the local plane coordinates, P 0 is the latitude and longitude of the origin of the world coordinate system, R is the radius of the earth, and C is the calculation process variable, as follows:

[0029] C = sin(λ*(π / 180))*sin(λ 0 *(π / 180))*cos(ψ*(π / 180)-ψ 0 *(π / 180))+cos(λ*(π / 180))*cos(λ 0 *(π / 180))

[0030] Among them, (ψ,λ) is the latitude and longitude of the origin in the local plane coordinates, (ψ 0 ,λ 0 ) is the latitude and longitude of the origin in the world coordinate system;

[0031] The calculation formula of the azimuth between the two points is as follows:

[0032]

[0033] Where A is the azimuth between the two points;

[0034] The calculation formula of the translation matrix T is as follows:

[0035]

[0036] The calculation formula of the second coordinate transformation relationship is as follows:

[0037]

[0038] Where T is the translation matrix.

[0039] The process of converting the world plane coordinates into geographic coordinates in step S4 includes linear transformation or Gaussian projection inverse transformation according to a specific range, wherein linear transformation is applicable to coordinate transformation within a small range, and Gaussian projection inverse transformation is applicable to coordinate transformation within a large range.

[0040] Furthermore, the linear transformation process comprises the following steps:

[0041] S411, selecting two feature points within the monitoring range of the camera, and obtaining the geographic coordinates and local plane coordinates of the two feature points;

[0042] S412, calculating the latitude corresponding to the unit length of the horizontal coordinate, and the longitude corresponding to the unit length of the vertical coordinate;

[0043] S413, obtaining the geographic coordinates of the origin in the local plane coordinate system, and calculating the geographic coordinates of the obstacle according to the latitude corresponding to the unit horizontal coordinate length and the longitude corresponding to the unit vertical coordinate length. The specific formula is as follows:

[0044]

[0045] Among them, (ψ1, λ1) and (ψ2, λ2) are the geographic coordinates of the feature point, (x1, y1) and (x2, y2) are the local plane coordinates of the feature point, (ψ, λ) are the geographic coordinates of the origin in the local plane coordinate system, (x, y) are the local plane coordinates of the obstacle, and (ψ′, λ′) are the geographic coordinates of the obstacle.

[0046] The calculation formula of the latitude dx corresponding to the unit horizontal coordinate length is as follows:

[0047]

[0048] The calculation formula of the longitude corresponding to the unit ordinate length is as follows:

[0049]

[0050] Furthermore, the process of the Gaussian projection inverse transform includes the following steps:

[0051] S421, setting the starting latitude to the latitude of the bottom point of the world coordinate system, that is, the latitude corresponding to the meridian arc length when x′=X, where x′ is the horizontal coordinate of the coordinate value of the obstacle in the world plane coordinate system, and X is the meridian arc length;

[0052] S422, performing iterative calculation according to the meridian arc length, and making a judgment according to a preset termination condition to obtain a final conversion latitude;

[0053] S423, obtaining the geographic coordinates of the obstacle according to the transformed latitude and the longitude and latitude of the origin in the world coordinate system;

[0054] The formula of the iteration process in step S422 is as follows:

[0055]

[0056]

[0057]

[0058]

[0059] in, is the latitude of the base point, X is the meridian arc length, is the transformation latitude of the i+1th iteration, is the transformation latitude of the i-th iteration, a0, a2, a4, a6 and a8 are process parameters;

[0060] The formula of the termination condition is as follows:

[0061]

[0062] Among them, ε is the calculation accuracy;

[0063] The formula for calculating the geographic coordinates of the obstacle in step S423 is as follows:

[0064]

[0065]

[0066] N f =a(1-e 2 sin2 ψ f ) -1 / 2

[0067] M f =a(1-e 2 sin 2 ψ f ) -3 / 2

[0068] t f =tanψ f

[0069]

[0070] Among them, (ψ′,λ′) is the geographical coordinates of the obstacle, ψ f is the final transformed latitude, (x′, y′) is the coordinate value of the obstacle in the world plane coordinate system, λ0 is the longitude of the central meridian, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, N f 、M f ,t f and is the process variable, a is the major axis of the earth ellipse, f is the flattening of the ellipsoid, e is the first eccentricity of the ellipsoid, e′ is the second eccentricity of the ellipsoid, b is the semi-minor axis of the earth's ellipse.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] The present invention aims at the operation scenario of the intelligent network-connected bus, constructs a mapping model, converts the pixel coordinates of the obstacles in the pictures collected by the camera into the geographic coordinate system, i.e., longitude and latitude, through the local plane coordinate system and the world coordinate system, and transmits this point to the intelligent network-connected bus, effectively avoiding the limitations caused by line of sight obstruction when the intelligent network-connected bus actively obtains images through its own device, realizes accurate obstacle position reminder and pedestrian warning, and improves the safety of the intelligent network-connected bus during driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a schematic diagram of the process of the present invention;

[0074] Figure 2 is a schematic diagram of a pixel coordinate system of the present invention;

[0075] Figure 3 is a schematic diagram of a local plane coordinate system of the present invention;

[0076] Figure 4 It is a schematic diagram of coordinate rotation of the present invention. DETAILED DESCRIPTION

[0077] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0078] Example

[0079] like Figure 1 As shown, a method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision specifically includes the following steps:

[0080] S1. Perform scene calibration according to the position information of the camera to obtain the image information taken by the camera and the position information of the obstacle;

[0081] S2. According to the image information taken by the camera, a pixel coordinate system and a local plane coordinate system are established respectively, and the pixel coordinates corresponding to the obstacle are calculated according to the position information of the obstacle, and converted into local plane coordinates;

[0082] S3, converting the local plane coordinates of the obstacle into world plane coordinates;

[0083] S4. Convert the world plane coordinates of the obstacle into geographic coordinates, and send the geographic coordinates to the controller of the intelligent connected bus. The controller controls the intelligent connected bus to avoid the obstacle according to the geographic coordinates of the obstacle.

[0084] The scene calibration in step S1 includes camera position calibration and feature point position calibration. The feature point position calibration process includes selecting m×n feature points with equal spacing on a plane within the camera monitoring range and obtaining the longitude and latitude of the feature points. The camera position calibration specifically includes the following steps:

[0085] S11, locate the latitude and longitude of the camera. Since the world plane coordinate system takes the north direction as the y-axis direction, it is necessary to obtain the angle θ between the camera and the north direction;

[0086] S12, obtaining the height of the camera, the horizontal distance between the nearest point of the image captured by the camera and the camera, and the actual distance of the camera's monitoring range.

[0087] The process of establishing the pixel coordinate system in step S2 includes taking the downstream horizontal line of the image information captured by the camera as the x-axis and the midpoint of the horizontal line as the origin to construct the pixel coordinate system, with the unit being pixel. Figure 2 As shown; the process of establishing a local plane coordinate system includes taking the camera as the origin and the horizontal line parallel to the x-axis of the pixel coordinate system as the x-axis, constructing a local plane coordinate system in meters, as shown Figure 3 shown.

[0088] Step S2 also includes calibrating the feature points in the pixel coordinate system and the local plane coordinate system, specifically including:

[0089] In the pixel coordinate system, the pixel coordinates of m×n feature points are calibrated respectively (u i , v i ), the pixel coordinate set of the feature point is expressed as (U, V);

[0090] In the local plane coordinates, the local plane coordinates (x i ,y i ), the plane coordinate set of the feature points is expressed as (X, Y); thus, the first coordinate transformation relationship between the pixel coordinate system and the local plane coordinate system is obtained, and the pixel coordinates of the obstacle are transformed into the local plane coordinates according to the first coordinate transformation relationship. The calculation formula of the first coordinate transformation relationship is as follows:

[0091]

[0092] Among them, (u, v) is the coordinate value of the obstacle in the pixel coordinate system, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, F1 is the horizontal coordinate conversion function, and F0 is the vertical coordinate conversion function.

[0093] The conversion from local plane coordinates to world plane coordinates is a rigid body transformation. The object will not be deformed, and only coordinate rotation and translation are required. The process of converting local plane coordinates to world plane coordinates in step S3 includes coordinate rotation and coordinate translation. The second coordinate transformation relationship between the local plane coordinate system and the world plane coordinate system is obtained according to the coordinate rotation and coordinate translation.

[0094] like Figure 4 As shown in the figure, the calculation formula for coordinate rotation is as follows:

[0095]

[0096] Among them, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, (x′, y′) is the coordinate value of the obstacle in the world plane coordinate system, θ is the angle between the camera and the north direction, and R1 is the rotation matrix;

[0097] The process of coordinate translation includes the following steps:

[0098] S31, calibrating the longitude and latitude of the origin in the local plane coordinate system and the longitude and latitude of the origin in the world coordinate system;

[0099] S32, calculating the distance and azimuth between the origin in the local plane coordinate system and the origin in the world coordinate system;

[0100] S33. Calculate a translation matrix based on the distance and azimuth between the two points.

[0101] The distance between two points is calculated as follows:

[0102] D(P,P 0 )=R*arccos(C)*π / 180

[0103] Among them, D(P,P 0 ) is the distance between the two points, P is the origin of the local plane coordinates, P 0 is the latitude and longitude of the origin of the world coordinate system, R is the radius of the earth, and C is the calculation process variable, as follows:

[0104] C = sin(λ*(π / 180))*sin(λ 0 *(π / 180))*cos(ψ*(π / 180)-ψ 0 *(π / 180))+cos(λ*(π / 180))*cos(λ 0 *(π / 180))

[0105] Among them, (ψ,λ) is the latitude and longitude of the origin in the local plane coordinates, (ψ 0 ,λ 0 ) is the latitude and longitude of the origin in the world coordinate system;

[0106] The formula for calculating the azimuth between two points is as follows:

[0107]

[0108] Where A is the azimuth between the two points;

[0109] The calculation formula of the translation matrix T is as follows:

[0110]

[0111] The calculation formula of the second coordinate transformation relationship is as follows:

[0112]

[0113] Where T is the translation matrix.

[0114] The process of converting the world plane coordinates into geographic coordinates in step S4 includes linear transformation or Gaussian projection inverse transformation according to the specific range, wherein linear transformation is applicable to coordinate transformation within a small range, and Gaussian projection inverse transformation is applicable to coordinate transformation within a large range.

[0115] Linear transformation does not require converting local plane coordinates to world plane coordinates. It can be directly converted from local plane coordinates to geographic coordinates, including the following steps:

[0116] S411, selecting two feature points within the monitoring range of the camera, and obtaining the geographic coordinates and local plane coordinates of the two feature points;

[0117] S412, calculating the latitude corresponding to the unit length of the horizontal coordinate, and the longitude corresponding to the unit length of the vertical coordinate;

[0118] S413, obtaining the geographic coordinates of the origin in the local plane coordinate system, and calculating the geographic coordinates of the obstacle according to the latitude corresponding to the unit horizontal coordinate length and the longitude corresponding to the unit vertical coordinate length. The specific formula is as follows:

[0119]

[0120] Among them, (ψ1, λ1) and (ψ2, λ2) are the geographic coordinates of the feature point, (x1, y1) and (x2, y2) are the local plane coordinates of the feature point, (ψ, λ) are the geographic coordinates of the origin in the local plane coordinate system, (x, y) are the local plane coordinates of the obstacle, and (ψ′, λ′) are the geographic coordinates of the obstacle.

[0121] The calculation formula for the latitude dx corresponding to the unit horizontal coordinate length is as follows:

[0122]

[0123] The calculation formula for the longitude corresponding to the unit ordinate length is as follows:

[0124]

[0125] The process of inverse Gaussian projection transformation includes the following steps:

[0126] S421, setting the starting latitude to the latitude of the bottom point of the world coordinate system, that is, the latitude corresponding to the meridian arc length when x′=X, where x′ is the horizontal coordinate of the coordinate value of the obstacle in the world plane coordinate system, and X is the meridian arc length;

[0127] S422, performing iterative calculation according to the meridian arc length, and making a judgment according to a preset termination condition to obtain a final conversion latitude;

[0128] S423, obtaining the geographic coordinates of the obstacle according to the transformed latitude and the longitude and latitude of the origin in the world coordinate system;

[0129] The formula of the iterative process in step S422 is as follows:

[0130]

[0131]

[0132]

[0133]

[0134] in, is the latitude of the base point, X is the meridian arc length, is the transformation latitude of the i+1th iteration, is the transformation latitude of the i-th iteration, a0, a2, a4, a6 and a8 are process parameters;

[0135] The formula for the termination condition is as follows:

[0136]

[0137] Among them, ε is the calculation accuracy;

[0138] The formula for calculating the geographic coordinates of the obstacle in step S423 is as follows:

[0139]

[0140]

[0141] N f =a(1-e 2 sin 2 ψ f ) -1 / 2

[0142] M f =a(1-e 2 sin 2 ψ f ) -3 / 2

[0143] t f =tanψ f

[0144]

[0145] Among them, (ψ′,λ′) is the geographical coordinates of the obstacle, ψ f is the final transformed latitude, (x′, y′) is the coordinate value of the obstacle in the world plane coordinate system, λ0 is the longitude of the central meridian, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, N f 、M f ,t f and is the process variable, a is the major axis of the earth ellipse, f is the flattening of the ellipsoid, e is the first eccentricity of the ellipsoid, e′ is the second eccentricity of the ellipsoid, b is the semi-minor axis of the earth's ellipse.

[0146] In addition, it should be noted that the specific embodiments described in this specification may be named differently, and the above contents described in this specification are merely examples of the structure of the present invention. Any equivalent changes or simple changes made based on the structure, features and principles of the present invention are included in the protection scope of the present invention. Those skilled in the art of the present invention may make various modifications or supplements to the specific examples described or adopt similar methods, as long as they do not deviate from the structure of the present invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision, characterized in that: The specific steps include: S1. Perform scene calibration according to the position information of the camera to obtain the image information taken by the camera and the position information of the obstacle; S2. According to the image information taken by the camera, a pixel coordinate system and a local plane coordinate system are established respectively, and the pixel coordinates corresponding to the obstacle are calculated according to the position information of the obstacle, and converted into local plane coordinates; S3, converting the local plane coordinates of the obstacle into world plane coordinates; S4, converting the world plane coordinates of the obstacle into geographic coordinates, and sending the geographic coordinates to the controller of the intelligent networked bus; The process of converting the local plane coordinates into the world plane coordinates in step S3 includes coordinate rotation and coordinate translation, and a second coordinate transformation relationship between the local plane coordinate system and the world plane coordinate system is obtained according to the coordinate rotation and coordinate translation; The calculation formula for the coordinate rotation is as follows: Among them, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, (x′, y′) is the coordinate value of the obstacle in the world plane coordinate system, θ is the angle between the camera and the north direction, and R1 is the rotation matrix; The coordinate translation process comprises the following steps: S31, calibrating the longitude and latitude of the origin in the local plane coordinate system and the longitude and latitude of the origin in the world coordinate system; S32, calculating the distance and azimuth between the origin in the local plane coordinate system and the origin in the world coordinate system; S33, calculating a translation matrix according to the distance and azimuth between the two points; The calculation formula for the distance between the two points is as follows: D(P,P 0 )=R*arccos(C)*π / 180 Among them, D(P,P 0 ) is the distance between the two points, P is the origin of the local plane coordinates, P 0 is the latitude and longitude of the origin of the world coordinate system, R is the radius of the earth, and C is the calculation process variable, as follows: C=sin(λ*(π / 180))*sin(λ 0 *(π / 180))*cos(ψ*(π / 180)-ψ 0 *(π / 180))+cos(λ*(π / 180))*cos(λ 0 *(π / 180)) Among them, (ψ,λ) is the latitude and longitude of the origin in the local plane coordinates, (ψ 0 ,λ 0 ) is the latitude and longitude of the origin in the world coordinate system; The calculation formula of the azimuth between the two points is as follows: Where A is the azimuth between the two points; The calculation formula of the translation matrix T is as follows: The calculation formula of the second coordinate transformation relationship is as follows: Where T is the translation matrix.

2. The method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision according to claim 1, characterized in that: The scene calibration in step S1 includes camera position calibration and feature point position calibration. The feature point position calibration process includes selecting multiple feature points at equal intervals on a plane within the camera monitoring range and obtaining the longitude and latitude of the feature points. The camera position calibration specifically includes the following steps: S11, locate the latitude and longitude of the camera, and obtain the angle between the camera and the north direction; S12, obtaining the height of the camera, the horizontal distance between the nearest point of the image captured by the camera and the camera, and the actual distance of the camera's monitoring range.

3. The method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision according to claim 1, characterized in that: The process of establishing the pixel coordinate system in step S2 includes taking the downstream horizontal line of the image information taken by the camera as the x-axis and the midpoint of the horizontal line as the origin to construct the pixel coordinate system; the process of establishing the local plane coordinate system includes taking the camera as the origin and the horizontal line parallel to the x-axis of the pixel coordinate system as the x-axis to construct the local plane coordinate system.

4. The method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision according to claim 2, characterized in that: The step S2 further includes calibrating the feature points in the pixel coordinate system and the local plane coordinate system to obtain a first coordinate transformation relationship between the pixel coordinate system and the local plane coordinate system, and transforming the pixel coordinates of the obstacle into local plane coordinates according to the first coordinate transformation relationship. The calculation formula of the first coordinate transformation relationship is as follows: Among them, (u, v) is the coordinate value of the obstacle in the pixel coordinate system, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, F1 is the horizontal coordinate conversion function, and F0 is the vertical coordinate conversion function.

5. The method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision according to claim 1, characterized in that: The process of converting the world plane coordinates into geographic coordinates in step S4 includes linear transformation or Gaussian projection inverse transformation according to a specific range, wherein linear transformation is applicable to coordinate transformation within a small range, and Gaussian projection inverse transformation is applicable to coordinate transformation within a large range.

6. The method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision according to claim 5, characterized in that: The linear transformation process comprises the following steps: S411, selecting two feature points within the monitoring range of the camera, and obtaining the geographic coordinates and local plane coordinates of the two feature points; S412, calculating the latitude corresponding to the unit length of the horizontal coordinate, and the longitude corresponding to the unit length of the vertical coordinate; S413, obtaining the geographic coordinates of the origin in the local plane coordinate system, and calculating the geographic coordinates of the obstacle according to the latitude corresponding to the unit horizontal coordinate length and the longitude corresponding to the unit vertical coordinate length. The specific formula is as follows: Among them, (ψ1, λ1) and (ψ2, λ2) are the geographic coordinates of the feature point, (x1, y1) and (x2, y2) are the local plane coordinates of the feature point, (ψ, λ) are the geographic coordinates of the origin in the local plane coordinate system, (x, y) are the local plane coordinates of the obstacle, and (ψ′, λ′) are the geographic coordinates of the obstacle.

7. The method for calculating the spatial coordinate edge of an obstacle based on roadside monocular vision according to claim 5, characterized in that: The process of the Gaussian projection inverse transformation comprises the following steps: S421, setting the starting latitude to the latitude of the bottom point of the world coordinate system, that is, the latitude corresponding to the meridian arc length when x′=X, where x′ is the horizontal coordinate of the coordinate value of the obstacle in the world plane coordinate system, and X is the meridian arc length; S422, performing iterative calculation according to the meridian arc length, and making a judgment according to a preset termination condition to obtain a final conversion latitude; S423, obtaining the geographic coordinates of the obstacle according to the transformed latitude and the longitude and latitude of the origin in the world coordinate system; The formula of the iteration process in step S422 is as follows: in, is the latitude of the base point, X is the length of the meridian arc, is the transformation latitude of the i+1th iteration, is the transformation latitude of the i-th iteration, a0, a2, a4, a6 and a8 are process parameters; The formula for calculating the geographic coordinates of the obstacle in step S423 is as follows: N f =a(1-e 2 sin 2 ψ f ) -1 / 2 M f =a(1-e 2 sin 2 ψ f ) -2 / 2 t f =tanψ f Among them, (ψ′,λ′) is the geographical coordinates of the obstacle, ψ f is the final transformed latitude, (x′, y′) is the coordinate value of the obstacle in the world plane coordinate system, λ0 is the longitude of the central meridian, (x, y) is the coordinate value of the obstacle in the local plane coordinate system, N f 、M f ,t f and is the process variable, a is the major axis of the earth ellipse, f is the flattening of the ellipsoid, e is the first eccentricity of the ellipsoid, e′ is the second eccentricity of the ellipsoid, b is the semi-minor axis of the earth's ellipse.

Images