WO2020006764A1

WO2020006764A1 - Path detection method, related device, and computer readable storage medium

Info

Publication number: WO2020006764A1
Application number: PCT/CN2018/094905
Authority: WO
Inventors: 李业; 廉士国
Original assignee: 深圳前海达闼云端智能科技有限公司
Priority date: 2018-07-06
Filing date: 2018-07-06
Publication date: 2020-01-09
Also published as: CN109074490B; CN109074490A

Abstract

The present application relates to the technical field of computer vision, and relates in particular to a path detection method, a related device, and a computer readable storage medium. The path detection method comprises: establishing a three-dimensional point cloud of a road according to acquired image information; detecting ground information of the road in the three-dimensional point cloud; determining an early warning area according to the ground information of the road; detecting traffic conditions in the early warning area, and determining a path detection result of the road according to the traffic conditions. The described method may be applied to the detection of paths in complex environments, improves user experience, and is simultaneously capable of providing more road condition information by means of a three-dimensional point cloud.

Description

Path detection method, related device and computer-readable storage medium

Technical field

The present application relates to the field of computer vision technology, and in particular, to a path detection method, a related device, and a computer-readable storage medium.

Background technique

Path detection is an extremely important technology in the fields of blindness guidance, robotics, and autonomous driving. Existing path detection is based on vision to detect the road that a vehicle or robot is traveling to improve the safety of the vehicle or robot.

technical problem

The inventor discovered in the process of studying the prior art that the traditional path detection usually sets a two-dimensional detection area in the image, and determines whether the area can pass by judging whether there is an obstacle in the area. However, due to the perspective projection characteristics of the image, if the image detection area is a rectangular detection area, the real world corresponds to a fan-shaped area in front of the camera. When detecting the acquired image, objects on both sides of the passable width will be regarded as obstacles. If the image detection area is a trapezoidal detection area, although the influence of the fan-shaped area can be corrected to a certain extent in the real world, the size and position of the trapezoidal detection area is extremely inconvenient and needs to be adjusted with the lens focal length, camera attitude, etc Change. In addition, traditional path detection usually only provides a rough warning of the obstacle ahead, and cannot provide more detailed road condition information, which makes subsequent decision-making and user experience extremely inconvenient.

Technical solutions

A technical problem to be solved in some embodiments of the present application is to provide a path detection method, a related device, and a computer-readable storage medium to solve the above technical problems.

An embodiment of the present application provides a path detection method, including:

Build a 3D point cloud of the road based on the acquired image information;

Detecting ground information of roads in 3D point clouds;

Determine early warning areas based on road surface information;

Detect the traffic conditions in the early warning area, and determine the road detection results based on the traffic conditions.

An embodiment of the present application further provides a path detection device, including: a establishment module, a first detection module, a determination module, and a second detection module;

A building module for building a three-dimensional point cloud of a road based on the acquired image information;

A first detection module for detecting ground information of a road in a three-dimensional point cloud;

A determining module for determining an early warning area based on road surface information;

The second detection module is used to detect the traffic condition in the early warning area, and determine the road detection result of the road according to the traffic condition.

An embodiment of the present application further provides an electronic device, including: at least one processor; and,

Memory in communication with at least one processor; wherein,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute the foregoing path detection method.

An embodiment of the present application further provides a computer-readable storage medium storing a computer program, and the computer program is executed by a processor to implement the foregoing path detection method.

Beneficial effect

Compared with the prior art, by establishing a three-dimensional point cloud of the road and determining the early warning area based on the three-dimensional point cloud of the road, the problem of inaccurate detection of the road due to the unreasonable early warning area set by the two-dimensional image is avoided, and Detecting the ground information of the road in the 3D cloud, and then determining the early warning area and the traffic conditions of the early warning area, can ensure the reliability of the path detection results, make it suitable for path detection in complex environments, and improve the user experience. At the same time, the 3D point cloud can provide More traffic information.

BRIEF DESCRIPTION OF THE DRAWINGS

One or more embodiments are exemplified by the pictures in the accompanying drawings. These exemplary descriptions do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the drawings in the drawings do not constitute a limitation on scale.

1 is a flowchart of a path detection method in a first embodiment of the present application;

2 is a relationship diagram between a pixel coordinate system and a camera coordinate system in the first embodiment of the present application;

3 is a relationship diagram between a camera coordinate system and world coordinates in the first embodiment of the present application;

4 is a flowchart of a path detection method in a second embodiment of the present application;

5 is a flowchart of another path detection method in the second embodiment of the present application;

6 is a structural diagram of a path detection device in a third embodiment of the present application;

FIG. 7 is a structural diagram of an electronic device in a fourth embodiment of the present application.

Embodiments of the invention

In order to make the purpose, technical solution, and advantages of the present application clearer, some embodiments of the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application, and are not used to limit the application. However, a person of ordinary skill in the art can understand that in the embodiments of the present application, many technical details are provided in order to make the reader better understand the application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solution claimed in this application can be implemented.

The first embodiment of the present application relates to a path detection method. As shown in FIG. 1, the method includes the following steps:

Step 101: Establish a three-dimensional point cloud of the road according to the acquired image information.

Specifically, the three-dimensional point cloud is a huge collection of points on the surface of the target object. The establishment of the three-dimensional point cloud can determine road information in space. The application can use multiple methods to establish the three-dimensional point cloud. This embodiment does not limit the establishment of three-dimensional points. The specific implementation of the cloud.

In a specific implementation, a three-dimensional point cloud can be established by using a depth map. The specific process of establishing a three-dimensional point cloud includes: obtaining a depth map and the attitude angle of the camera, where the attitude angle is the attitude angle of the camera when the depth map was taken; calculating a scale normalization factor according to the depth map and a preset normalization scale; Calculate the scale normalized depth map according to the depth map and the scale normalization factor; construct a 3D point cloud in the camera coordinate system based on the scale normalized depth map; and according to the 3D point cloud and camera in the camera coordinate system To construct a three-dimensional point cloud in the world coordinate system.

It should be noted that there are many methods for obtaining depth maps, including but not limited to: lidar depth imaging method, computer stereo vision imaging, coordinate measuring machine method, moire fringe method, structured light method, depth is not limited here How to get the graph.

Specifically, using formula 1 to calculate the scale normalization factor, formula 1 is expressed as follows:

S = Norm / max (W, H) (1)

Among them, S represents the scale normalization factor, W represents the width of the depth map, H represents the height of the depth map, and Norm represents a preset normalized scale. Norm is a preset known amount. In specific applications, if it is necessary to process the depth map of continuous frames to establish a three-dimensional point cloud, the normalization scale used in the processing of the depth map of each frame remains unchanged.

The normalized depth map is calculated using Equation 2. Equation 2 is expressed as follows:

Among them, W _S represents the width of the depth map after normalization, and H _S represents the height of the depth map after normalization. The depth-normalized depth map can be determined according to W _S and H _S.

Specifically, a three-dimensional point cloud in a camera coordinate system is constructed according to Equation 3 and a normalized depth map, and the three-dimensional point cloud in the camera coordinate system is represented as P (X _c , Y _c , Z _c ). Each pixel in the picture contains the distance between the camera and the object. Then, the pixel coordinates in the depth map are converted into the coordinates of the camera coordinate system using Equation 3 to form a three-dimensional point cloud in the camera coordinate system. Equation 3 is expressed as follows:

Among them, u and v are the coordinate values of any point P in the normalized depth map, X _c , Y _c , and Z _c are the coordinate values of point P in the camera coordinate system, and M _{3 × 4} is the internal parameter of the camera Matrix, Z _c is the depth value of point P in the depth map after normalization, that is, the distance value from the camera to the shooting object, which is a known quantity.

According to the coordinate conversion relationship between the camera coordinate system and the world coordinate system, the three-dimensional point cloud P (X _c , Y _c , Z _c ) in the camera coordinate system is converted into a three-dimensional point cloud P (X _w , X _w , Y _w , Z _w ), the conversion relationship is expressed by Equation 4:

Among them, X _w , Y _w and Z _w are the coordinate values of any point P in the three-dimensional point cloud in the world coordinate system, X _c , Y _c and Z _c are the coordinate values of point P in the camera coordinate system, and α is the camera The angle between the camera and the X _w axis in the world coordinate system, β is the angle between the camera and the Y _w axis in the world coordinate system, and γ is the angle between the camera and the Z _w axis in the world coordinate system.

Among them, assuming that the image coordinate system is o ₁ -xy, the relationship between the camera coordinate system O _c -X _c Y _c Z _c and the pixel coordinate system o-uv is shown in FIG. 2, and the camera coordinate system O _c -X _c Y _c The relationship between Z _c and world coordinates O _w -X _w Y _w Z _w is shown in FIG. 3.

Among them, as shown in FIG. 2, a rectangular coordinate system o-uv in pixels, which is established with the upper left corner of the depth map as the origin, is used as the pixel coordinate system. The abscissa u represents the number of pixel columns where the pixels are located, and the ordinate v represents the pixels The number of pixel rows at. The intersection of the camera optical axis and the depth map plane is defined as the origin o ₁ of the image coordinate system o ₁ -xy, and the x-axis is parallel to the u-axis and the y-axis is parallel to the v-axis. The camera coordinate system O _c -X _c Y _c Z _c uses the camera optical center O _c as the origin, the X _c axis and Y _c axis are respectively parallel to the x and y axes in the image coordinate system, and the Z _c axis is the light of the camera The axis is perpendicular to the image plane and intersects at o ₁ point.

Among them, as shown in FIG. 3, the origin O _{w of the} world coordinate system O _w -X _w Y _w Z _w coincides with the origin O _{c of the} camera coordinate system, both of which are camera light centers, and the horizontal direction to the right is the positive direction of the X _w axis. Vertically downward is the positive direction of the Y _w axis, perpendicular to the X _w O _w Y _w plane and pointing straight ahead is the positive direction of the Z _w axis, and a world coordinate system is established.

It is worth mentioning that the construction of three-dimensional point clouds based on image information is not limited to the construction of depth maps. For example, laser point cloud data can also be directly obtained by lidar, and three-dimensional point clouds are constructed based on point cloud data. The construction of a three-dimensional point cloud is an exemplary description, and this embodiment does not limit the specific method adopted for constructing the three-dimensional point cloud.

Step 102: Detect the ground information of the road in the three-dimensional point cloud.

In a specific implementation, the specific implementation process of this step is: detecting the ground height in the three-dimensional point cloud; determining obstacle information on the ground height; and using the ground height and the obstacle information as ground information.

It should be noted that determining the ground height and detecting obstacle information on the ground height in a three-dimensional point cloud makes it possible to determine the specific conditions of the road and provide a possibility for ensuring the accuracy of the detection results.

It is worth mentioning that after determining the ground height, pothole detection can be performed on the road to determine the pothole status of the road, and the pothole status of the road is taken as a part of the ground information. In practice, other road-related detections are also performed, such as detecting the types of roads, including blind roads, sidewalks, pedestrian zebra crossings, etc. For example, if this method is applied to blind sticks, it is necessary to determine the specific road category of those currently using blind sticks. Therefore, in practice, more ground information can be detected as needed, which is not limited here.

Step 103: Determine an early warning area according to the ground information of the road.

Specifically, the space coordinates of the early warning area are constructed; the height position of the early warning area in space coordinates is determined according to the ground height; the width and distance of the early warning area in space coordinates are determined according to the obstacle information, thereby determining the early warning area.

It should be noted that the determination of the early warning area is based on the three-dimensional point cloud in the world coordinate system. Specifically, the Y _w O _w Z _w plane of the world coordinate system is a symmetrical plane, and the early warning area is constructed in the positive direction of the Z _w axis. The three-dimensional space area is the early warning area. The space area of the early warning area is represented as vBox (x, y, z), where x, y, and z respectively represent the width, height, and distance of the early warning area. The distance is determined by the speed of the user, the width and height of the early warning area is determined according to the shape of the user, and the early warning area is not less than the minimum space that the user can pass through. For example, the height of one user is 1.5m, the weight is 90kg (kg), and the speed of action is slow. The warning zone can be set to vBox (100, 170, 150) in cm (centimeter); the height of the other user is 1.9m , Weight 55kg, agile speed, early warning area can be set to vBox (60, 210, 250), unit cm.

It is worth mentioning that in this embodiment, path detection can be performed in the image information of consecutive frames. For depth maps, the coordinate system needs to be converted for each frame of the depth map, but the coordinate values of the early warning area can remain unchanged. It is necessary to determine the location of the early warning area according to the three-dimensional point clouds corresponding to different frames of images.

In addition, the road is not a flat road, and the ground information includes the ground height. After the early warning area is determined, the position of the early warning area needs to be adjusted according to the ground height. For example, roads can be divided into uphill sections, downhill sections and flat sections with different ground heights. The location of the early warning area is adjusted based on the ground height in the ground information, and the traffic conditions of the road need to be detected based on the obstacle information and the size of the early warning area.

In a specific implementation, the real-time ground height is determined according to an adaptive ground detection method, or the real-time ground height is determined based on point cloud data indicating road information in a three-dimensional point cloud, and the position of the early warning area is dynamically adjusted according to changes in ground height. After the adjustment, the early warning area can be ensured to be directly above the ground. In this way, not only can energy efficiency be avoided to avoid ground interference, but also low-level obstacles will not be missed. Specifically, the adjustment early warning area can be determined by Formula 5, specifically expressed as follows:

vBox ₁ = vBox (x, H + y + σ, z) (5)

Among them, H represents the real-time ground height, σ represents the dynamic adjustment margin, vBox ₁ represents the adjusted early warning area, and x, y, and z represent the width, height, and distance of the early warning area, respectively.

Step 104: Detect the traffic conditions in the early warning area, and determine the road detection result according to the traffic conditions.

Specifically, the traffic conditions in the early warning area can be detected based on the obstacle information of the road, and the traffic conditions can specifically indicate the information such as the position of the traffic area and the width and height of the traffic area. After detecting the traffic conditions in the early warning area, determine whether the traffic conditions indicate that the road is passable; if so, determine the planned route in the early warning area and determine the road detection result based on the traffic route; otherwise, determine that the road detection result is impassable.

Specifically, after the path detection result is determined, early warning information is sent according to the path detection result. The early warning information includes, but is not limited to, obstacle information, traffic conditions, and ground height.

The early warning information may be one or a combination of sound information, image information, or light information. For example, if the method is applied to an intelligent robot, after the path detection result is obtained, it may be converted into machine language, so that the The intelligent robot can determine the path condition in the current frame.

It should be noted that the channel detection result may also be reminded to the user in other forms, or prompted to the user after performing appropriate information conversion, which is not specifically limited here.

Compared with the prior art, by establishing a three-dimensional point cloud of the road and determining the early warning area based on the three-dimensional point cloud of the road, the problem of inaccurate detection of the road due to the unreasonable setting of the two-dimensional image warning area is avoided, and Detecting the ground information of roads in the cloud, and then determining the early warning area and the traffic conditions of the early warning area, can ensure the reliability of the road detection results, make it suitable for the road detection in complex environments, and improve the user experience. At the same time, 3D point clouds can provide more More traffic information.

The second embodiment of the present application relates to a path detection method. This embodiment is substantially the same as the first embodiment. The main difference is that this embodiment specifically describes the specific implementation of determining the ground height in a three-dimensional point cloud. The specific implementation of the path detection method is shown in FIG. 4 and includes the following steps:

It should be noted that step 201 is the same as step 101 in the first embodiment, and steps 209 and 210 are the same as step 103 and step 104 in the first embodiment, respectively, and the same steps are not described herein again.

Step 202: Perform automatic threshold segmentation in the height direction on the three-dimensional point cloud to obtain a first ground area.

Step 203: Perform fixed threshold segmentation on the distance direction of the three-dimensional point cloud to obtain a second ground area.

Step 204: Determine an initial ground area according to the first ground area and the second ground area.

Step 205: Calculate the inclination of the initial ground area.

Step 206: Determine the ground height of the ground area according to the inclination.

Step 207: Determine obstacle information at the ground height.

Step 208: Use the ground height and obstacle information as ground information.

Steps

207 and 208 have been described in the first embodiment, and are not repeated here.

Specifically, in this embodiment, the three-dimensional point cloud in the world coordinate system is divided in the height direction and the horizontal direction. It should be noted that the three-dimensional point cloud in the world coordinate system defines Y _w as a coordinate in the height direction. Z _w is a set of coordinates in the distance direction, and X _w is a set of coordinates in the width direction. Then step 203 is to perform division in the direction designated by the Y _w axis, and step 204 is to perform division in the direction indicated by the Z _w axis.

In a specific implementation, the specific process of obtaining the first ground region is: according to the height of the region of interest (ROI) selected by the user in the three-dimensional point cloud in the world coordinate system, calculating and obtaining the first segment Threshold; calculate the second segmentation threshold based on the ground height of the previous depth map of the current depth map; and perform automatic threshold segmentation in the height direction of the 3D point cloud in the world coordinate system based on the first and second segmentation thresholds The specific segmentation process can be expressed by Equation 6:

Y _mask = a * ThdY _roi + b * ThdY _pre (6)

Among them, Y _mask represents the first ground area, ThdY _roi is the first segmentation threshold, ThdY _pre is the second segmentation threshold, a and b are weighting coefficients, and the specific values of a and b are set by the user according to actual needs.

It should be noted that when obtaining the first segmentation threshold and the second segmentation threshold, the automatic threshold segmentation algorithms that can be used include the mean method, Gauss method, or Otsu method. Since the automatic threshold segmentation algorithm is relatively mature, in this embodiment, This will not be repeated here.

Specifically, the specific segmentation of the second ground region is obtained as follows: the minimum coordinate value of the distance direction selected by the user in the three-dimensional point cloud in the world coordinate system is set as the third segmentation threshold and set to Z _min ; The maximum coordinate value of the distance direction selected in the three-dimensional point cloud in the coordinate system is set as Z _max as the fourth segmentation threshold; according to the third segmentation threshold and the fourth segmentation threshold, the three-dimensional point cloud in the world coordinate system is distanced. A fixed threshold segmentation in the direction is used to obtain a second ground area, which is set to Z _mask , that is, a region obtained by retaining a Z _w value between Z _min and Z _max is a second ground area.

Specifically, after the first ground area and the second ground area are obtained, the initial ground area can be determined, and the first ground area and the second ground area can be determined to determine the initial ground area. The specific ground area can be determined through formula 7. The initial ground area is expressed as follows:

Gnd ₀ = Y _mask ∩Z _mask (7)

Among them, Gnd ₀ is the initial ground area, Y _mask is the first ground area, and Z _mask is the second ground area. The specific physical meaning of the formula is that the suspected ground area in the height direction can be determined through the first ground area, and the range of the first ground area in the distance direction can be further limited through the second ground area, thereby ensuring the final acquisition. Accuracy of the initial ground area.

Specifically, when calculating the inclination of the initial ground area, the plane where the initial ground area is located must be determined first, that is, the plane fitting of the initial ground area is performed. The inclination of the plane and the coordinate axis determined according to the plane fitting is the initial ground area. Angle of inclination.

It should be noted that in the plane fitting, the points on the initial ground area are used as the known quantity, and the least square method or random sampling consistency algorithm is used to perform the plane fitting on the initial ground area to obtain the initial ground area The general equation of the plane in which it lies. Of course, other fitting methods may also be used to perform plane fitting on the initial ground area, and the specific method of plane fitting is not limited in the embodiments of the present application.

In a specific implementation, after fitting the initial plane to obtain the general equation of the plane where the initial ground area is located, AX + BY + CZ = D, the normal vector of the plane can be determined:

Further, the inclination angle of the initial ground area can be determined according to the normal vector. Specifically, the normal vector of the fitting plane and the vertical upward unit vector are used.

The included angle is the horizontal tilt angle θ of the initial ground, and the tilt angle θ is calculated by Equation 8:

Where θ is the inclination of the initial ground area,

Is the normal vector of the initial ground area,

Is a vertical upward unit vector,

Express

Mold

Express

Of mold.

The ground height of the ground area can be determined according to the inclination of the initial ground area, which can be the ground height of a point on the ground area, or the real-time ground height.

It is worth mentioning that the path detection method in this embodiment is based on the detection of image data of continuous frames. The specific implementation process of path detection of image data of continuous frames is shown in FIG. 5 and includes the following implementation steps:

Step 301: Initialize the system.

Step 302: Establish a three-dimensional point cloud of the road according to the acquired image information.

Step 303: Detect the ground information of the road in the three-dimensional point cloud.

Step 304: Determine an early warning area according to the ground information of the road.

Step 305: Detect the traffic condition in the early warning area, and determine whether it is passable. If yes, go to step 306; otherwise, go to step 307.

Step 306: Determine the passing route planned in the early warning area and determine the detection result of the road according to the passing route.

Step 307: Determine that the detection result of the road is impassable.

Step 308: Send a warning message according to the detection result of the path.

Step 309: Determine whether there is image information of the next frame. If yes, go to step 302; otherwise, end the path detection.

The steps of the various methods above are just for clarity of description. During implementation, they can be combined into one step or some steps can be split and divided into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this patent. ; Add insignificant modifications to the algorithm or process or introduce insignificant designs, but the core design of the algorithm and process that does not change is within the scope of this patent.

The third embodiment of the present application relates to a path detection device. The specific structure is shown in FIG. 6 and includes: a establishing module 601, a first detection module 602, a determination module 603, and a second detection module 604.

A establishing module 601 is configured to establish a three-dimensional point cloud of a road according to the acquired image information. The first detection module 602 is configured to detect ground information of a road in a three-dimensional point cloud. A determining module 603 is configured to determine an early warning area according to ground information of a road. The second detection module 604 is configured to detect a traffic condition in the early warning area, and determine a road detection result of the road according to the traffic condition.

It is not difficult to find that this embodiment is a device embodiment corresponding to the first or second embodiment, and this embodiment can be implemented in cooperation with the first or second embodiment. The related technical details mentioned in the first or second embodiment are still valid in this embodiment. In order to reduce repetition, details are not repeated here.

The fourth embodiment of the present application relates to an electronic device. The specific structure is shown in FIG. 7 and includes: at least one processor 701; and a memory 702 communicatively connected to the at least one processor 401; Instructions executed by the at least one processor 701, and the instructions are executed by the at least one processor 701, so that the at least one processor 701 can execute the path detection method in the first or second embodiment.

The memory and the processor are connected in a bus manner. The bus may include any number of interconnected buses and bridges. The bus links one or more processors and various circuits of the memory together. The bus can also link various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art, so they are not described further herein.

The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory can be used to store data used by the processor when performing operations.

A fifth embodiment of the present application relates to a computer-readable storage medium. The readable storage medium is a computer-readable storage medium, and the computer-readable storage medium stores computer instructions that enable a computer to execute the first The method for path detection involved in one or the second method embodiments.

It should be noted that those skilled in the art can understand that the display method in the above embodiments is implemented by a program instructing related hardware. The program is stored in a storage medium and includes several instructions for making a device (may It is a single-chip microcomputer, a chip, or the like) or a processor that executes all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random-Access Memory), magnetic disks or optical disks and other media that can store program codes .

Those of ordinary skill in the art can understand that the foregoing embodiments are specific embodiments for implementing the present application, and in practical applications, various changes can be made in form and details without departing from the spirit and range.

Claims

A path detection method, including:

Build a 3D point cloud of the road based on the acquired image information;

Detecting ground information of the road in the three-dimensional point cloud;

Determining an early warning area according to the ground information of the road;

Detect a traffic condition in the early warning area, and determine a road detection result of the road according to the traffic condition.
The path detection method according to claim 1, wherein the detecting ground information of the road in the three-dimensional point cloud specifically comprises:

Detecting a ground height in the three-dimensional point cloud;

Determining obstacle information at the ground height;

Use the ground height and the obstacle information as the ground information.
The path detection method according to claim 2, wherein the determining an early warning area based on ground information of the road specifically comprises:

Constructing the spatial coordinates of the early warning area;

Determining a height position of the early warning area under the spatial coordinates according to the ground height;

Determining the width and distance of the early warning area in the spatial coordinates according to the obstacle information.
The method for detecting a path according to claim 2 or 3, wherein after determining an early warning area based on the ground information of the road and before detecting the traffic condition of the early warning area, the method for detecting a route further comprises:

Adjusting the position of the early warning area according to the ground height.
The path detection method according to any one of claims 1-4, wherein the determining a path detection result of the road according to the traffic condition specifically includes:

Judging whether the traffic condition indicates that the road is passable;

If yes, determine a passing route planned in the early warning area, and determine a detection result of the road according to the passing route;

Otherwise, it is determined that the detection result of the road is impassable.
The path detection method according to claim 2 or 3, wherein the detecting the ground height in the three-dimensional point cloud specifically comprises:

Performing automatic threshold segmentation of the three-dimensional point cloud in the height direction to obtain a first ground area;

Performing a fixed threshold segmentation in the distance direction on the three-dimensional point cloud to obtain a second ground area;

Determining an initial ground area according to the first ground area and the second ground area;

Calculating the inclination of the initial ground area;

A ground height of the ground area is determined according to the inclination angle.
The path detection method according to any one of claims 1-6, wherein the image information includes: a depth map and a pose angle of a camera.
The path detection method according to claim 7, wherein the establishing a three-dimensional point cloud of a road based on the acquired image information specifically comprises:

Calculating a scale normalization factor according to the depth map and a preset normalization scale;

Calculating a scale-normalized depth map according to the depth map and the scale-normalization factor;

Constructing a three-dimensional point cloud in a camera coordinate system according to the scale-normalized depth map;

According to the three-dimensional point cloud in the camera coordinate system and the attitude angle of the camera, a three-dimensional point cloud in the world coordinate system is constructed.
The path detection method according to claim 2 or 3, wherein the determining the obstacle information at the ground height specifically comprises:

Determining a ground position of the road according to the ground height;

Performing pothole detection on the ground position of the road to obtain a pothole detection result;

According to the ground height and the pothole detection result, obstacle information on the ground height is generated.
The path detection method according to any one of claims 1-9, wherein the path detection result includes at least one of a position of an obstacle on a road, a type of the obstacle on the road, and a decision suggestion.
The path detection method according to any one of claims 1 to 9, wherein after the path detection result of the road is determined according to the traffic condition, the path detection method further comprises:

An early warning is issued based on the result of the path detection of the road.
A path detection device, including: a establishment module, a first detection module, a determination module, and a second detection module;

The establishing module is configured to establish a three-dimensional point cloud of a road according to the acquired image information;

The first detection module is configured to detect ground information of the road in the three-dimensional point cloud;

The determining module is configured to determine an early warning area according to the ground information of the road;

The second detection module is configured to detect a traffic condition in the early warning area, and determine a road detection result of the road according to the traffic condition.
An electronic device including:

At least one processor; and

A memory connected in communication with the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute the method according to any one of claims 1-11. Access detection method.
A computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the path detection method according to any one of claims 1-11 is implemented.