CN118708746A - Vector data generation method, device, electronic device and storage medium - Google Patents

Abstract

The present application provides a vector data generation method, device, electronic device and readable storage medium, the method comprising: acquiring point cloud data; determining a top view corresponding to the point cloud data according to the point cloud data; loading the top view and point cloud data into the corresponding top view operation window and laser point cloud window respectively; determining a target area image in the top view operation window, or determining a target area image in the laser point cloud window, wherein the target area image contains a traffic identifier; generating vector data corresponding to the traffic identifier based on the target area image in the top view operation window, and displaying the vector data in the top view operation window; or generating vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and displaying the vector data in the laser point cloud window. The present application significantly improves the efficiency and accuracy of map data production through an automated and intelligent data processing method, while reducing costs and operational difficulty.

Description

Vector data generation method, device, electronic device and storage medium

Technical Field

The present application relates to the technical field of vector data generation, and in particular to a vector data generation method, device, electronic device and storage medium.

Background Art

The process of map data production involves the production of identifiers such as traffic signs, road markings, road facilities, traffic lights, zebra crossings, etc. In particular, for the production of zebra crossings, grid no-parking zones and other elements, traditional methods have a series of problems: it takes a long time, requires operators to pick up point cloud coordinates very accurately, and operators need to constantly adjust the point cloud perspective to select the most accurate points. This leads to higher requirements for operators and consumes a lot of manual work time. At present, the vector data of manually drawn traffic identifiers has the problems of low efficiency and low accuracy. Manual work requires a lot of time and energy, and is easily affected by obstructions, which affects the accuracy of the data. Therefore, existing methods cannot meet the needs of efficient and accurate production of map data.

Summary of the invention

In view of this, embodiments of the present application provide a vector data generation method, device, electronic device and readable storage medium to solve the technical problems of low efficiency and low accuracy caused by manual drawing of map data in the prior art.

In a first aspect of an embodiment of the present application, a vector data generation method is provided, including: acquiring point cloud data, which is point cloud data obtained by acquiring the external environment of a vehicle through a vehicle-mounted laser scanner; determining a top view corresponding to the point cloud data based on the point cloud data; loading the top view and the point cloud data into the corresponding top view operation window and laser point cloud window, respectively; determining a target area image in the top view operation window, or determining a target area image in the laser point cloud window, wherein the target area image includes a traffic identifier; generating vector data corresponding to the traffic identifier based on the target area image in the top view operation window, and displaying the vector data in the top view operation window; or generating vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and displaying the vector data in the laser point cloud window.

According to a second aspect of an embodiment of the present application, a vector data generating device is provided, comprising an acquisition module for acquiring point cloud data, wherein the point cloud data is point cloud data obtained by acquiring the external environment of a vehicle through a vehicle-mounted laser scanner; a first determination module for determining a top view corresponding to the point cloud data based on the point cloud data; a loading module for loading the top view and the point cloud data into the corresponding top view operation window and the laser point cloud window, respectively; a second determination module for determining a target area image in the top view operation window, or determining a target area image in the laser point cloud window, wherein the target area image includes a traffic identifier; a generation module for generating vector data corresponding to the traffic identifier based on the target area image in the top view operation window, and displaying the vector data in the top view operation window; or generating vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and displaying the vector data in the laser point cloud window.

According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the method provided in the first aspect are implemented.

According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the window touch unit is controlled to implement the steps of the method provided in the first aspect above.

Compared with the prior art, the embodiments of the present application have at least the following beneficial effects: the embodiments of the present application avoid the repetitiveness and inefficiency in the traditional manual operation process by using the point cloud data obtained by the vehicle-mounted laser scanner. Automated point cloud data acquisition and processing significantly improves the speed of data acquisition and reduces the demand for human resources, thereby achieving high efficiency in map data production. By generating a bird's-eye view using point cloud data, it is ensured that the vector data of the traffic identifier is highly matched with the actual environment, and the error caused by imprecise manual operation is reduced. This method improves the accuracy of map data and ensures the quality and reliability of the map. By loading the point cloud data and the bird's-eye view into the corresponding operation window respectively, the present invention simplifies the operation process, so that the operator can more intuitively identify and process the traffic identifier. Even in the case of obstructions, the required point cloud data can be quickly and accurately selected by adjusting the viewing angle, reducing the requirements for the skills of the operator. In addition, the present application allows the operator to flexibly switch between the bird's-eye view operation window and the laser point cloud window, and select the most suitable view to work according to different situations. This flexibility enables the operator to better process data in complex scenes, improves the adaptability of work and the ability to cope with complex situations.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

FIG1 is a flow chart of a method for generating vector data according to an embodiment of the present application;

FIG2 is a flow chart of another vector data generating method according to an embodiment of the present application;

FIG3 is a schematic diagram of an application scenario of an embodiment of the present application;

FIG4 is a block diagram of a vector data generating device according to an embodiment of the present application;

FIG. 5 is a schematic diagram of the structure of an electronic device according to an embodiment of the present application.

DETAILED DESCRIPTION

In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

FIG1 is a flow chart of a vector data generation method according to an embodiment of the present application. The method provided by the embodiment of the present application can be executed by any electronic device with computer processing capabilities, such as a background server.

As shown in FIG. 1 , the vector data generating method includes steps S110 to S150 .

In step S110, point cloud data is acquired, where the point cloud data is acquired by using a vehicle-mounted laser scanner to acquire the vehicle's external environment.

In step S120, a top view corresponding to the point cloud data is determined based on the point cloud data.

In step S130, the top view and point cloud data are loaded into the corresponding top view operation window and laser point cloud window respectively.

In step S140, a target area image in the top view operation window is determined, or a target area image in the laser point cloud window is determined, wherein the target area image includes a traffic identifier.

In step S150, vector data corresponding to the traffic identifier is generated based on the target area image in the overhead view operation window, and the vector data is displayed in the overhead view operation window; or vector data corresponding to the traffic identifier is generated based on the target area image in the laser point cloud window, and the vector data is displayed in the laser point cloud window.

The method can avoid the repetitiveness and inefficiency of the traditional manual operation process by using the point cloud data obtained by the vehicle-mounted laser scanner. Automated point cloud data acquisition and processing significantly improves the speed of data acquisition and reduces the demand for human resources, thereby achieving high efficiency in map data production. By generating a bird's-eye view using point cloud data, it is ensured that the vector data of the traffic identifier is highly matched with the actual environment, and the error caused by imprecise manual operation is reduced. This method improves the accuracy of map data and ensures the quality and reliability of the map. By loading the point cloud data and the bird's-eye view into the corresponding operation window respectively, the present invention simplifies the operation process, so that the operator can more intuitively identify and process the traffic identifier. Even in the case of obstructions, the required point cloud data can be quickly and accurately selected by adjusting the viewing angle, reducing the requirements for the skills of the operator. In addition, the present application allows the operator to flexibly switch between the bird's-eye view operation window and the laser point cloud window, and select the most suitable view to work according to different situations. This flexibility enables the operator to better process data in complex scenes, improves the adaptability of work and the ability to cope with complex situations.

In some embodiments, when making traffic-related map data, the use of vehicle-mounted laser scanner technology is an advanced data collection method. This device is usually installed in front of the vehicle, and sometimes it may be installed on the roof or other locations to obtain detailed three-dimensional information about the surrounding environment. The laser scanner detects the surrounding environment by emitting a laser beam. When the laser beam hits any object, it is reflected back to the laser scanner. The device determines the distance of the object by calculating the time difference between the emission and reception of the laser beam. In this way, the laser scanner can quickly and continuously collect distance information of the surrounding environment while the vehicle is driving. The collected distance information is converted into point cloud data, which is a data set containing millions of spatial coordinate points. Each point represents the location where the laser beam is reflected back, thus forming a detailed three-dimensional model of the vehicle's external environment. Point cloud data not only contains location information, but may also contain reflectivity information, that is, the intensity of the laser beam reflected back, which can be used to distinguish different types of object surfaces.

During the driving process of the vehicle, the laser scanner can capture various traffic identifiers including traffic lights, traffic signs, road markings, parking spaces, zebra crossings, no-parking zones, etc. For example, through point cloud data, the position and height of the traffic lights can be identified, and in some cases, the status of the traffic lights (red, yellow, green light) can be determined. The shape, size and position of the sign can be extracted from the point cloud data, and then the type of traffic sign can be identified. The width, length and shape of the road markings can be determined by the linear features in the point cloud data. The boundaries and size of parking spaces can be identified by the markers in the point cloud data. Zebra crossings are usually represented by regular line patterns in the point cloud data, which can be used to determine their position and width. No-parking zones can be identified by analyzing the ground markers and surrounding environmental features in the point cloud data. In this way, on-board laser scanners provide an efficient and accurate method to collect traffic environment information, greatly improving the speed of production and update of traffic-related map data. This data is not only very useful for making ordinary maps, but also crucial for the navigation systems of autonomous vehicles, because these systems rely on accurate environmental information to navigate safely.

In some embodiments, the point cloud data obtained by the above method contains a large number of isolated points and discrete points, which are affected by various factors such as environmental factors, equipment problems, electromagnetic wave diffraction, data splicing and registration, resulting in noise and clutter in the point cloud data. In view of this situation, before determining the top view corresponding to the point cloud data according to the point cloud data, the above method further includes steps S210 and S220, as shown in FIG2 .

In step S210, denoising is performed on the point cloud data to obtain denoised point cloud data.

In step S220, the denoised point cloud data is subjected to clutter removal processing to obtain target point cloud data.

This method can perform denoising on point cloud data to obtain denoised point cloud data; perform clutter removal on the denoised point cloud data to obtain target point cloud data. In this way, after denoising and clutter removal, the target point cloud data obtained will be clearer and more accurate.

In some embodiments, the point cloud data is denoised to obtain denoised point cloud data; the denoised point cloud data is subjected to clutter removal to obtain target point cloud data. For example, first, the original point cloud data is preliminarily checked to remove those points that obviously do not belong to the target object, such as points in the sky or points far from the ground. A filtering algorithm is used to remove noise. Common filtering algorithms include statistical outlier removal (SOR), radius outlier removal, etc. The average distance of each point is calculated, and those points with greatly different distances from surrounding points are eliminated. For each point, only the neighboring points within the specified radius are considered. If the number of neighboring points of a point is less than a certain threshold, it is considered to be noise and removed. These steps are iterated multiple times to ensure that as much noise as possible is removed while retaining the necessary data points. A ground segmentation algorithm is used to separate ground points and non-ground points in the point cloud. Ground points are usually the basis for constructing a bird's-eye view, while non-ground points may contain traffic identifiers. Cluster analysis is performed on non-ground points to group points that are close in space together, which helps identify objects such as vehicles, pedestrians, and traffic signs. Based on the clustering results, point groups that do not belong to the target category (such as traffic identifiers) are identified and removed, which may be caused by other clutter such as trees and buildings. More advanced filtering techniques are applied, such as model-based filtering, in which known traffic identifier shape and size information is used to further clean up mismatched data points. After the above denoising and clutter removal processing, the target point cloud data obtained will be clearer and more accurate.

In some embodiments, according to the point cloud data, determining the top view corresponding to the point cloud data includes: converting the point cloud data by a preset conversion method to obtain pixel position data corresponding to the point cloud data; generating a top view corresponding to the point cloud data based on the pixel position data corresponding to the point cloud data. For example, the point cloud data may be the point cloud data after the above preprocessing. When making traffic-related map data, generating a two-dimensional top view from three-dimensional point cloud data is a common data processing method. This process usually involves orthographic projection of the point cloud data onto a plane (usually the XOY plane) and converting the points in the three-dimensional space into pixels on the two-dimensional plane. First, the vehicle coordinate system must be determined, usually with the X-axis pointing to the front of the vehicle (for example, the direction in which the vehicle is moving), the Y-axis pointing to the left of the vehicle, and the Z-axis pointing vertically to the ground and upward. In most cases, the XOY plane is selected as the projection plane, so that a top view of the map can be obtained, which contains all important information on the horizontal plane of the vehicle. A conversion method is preset, that is, how to map points in the three-dimensional space to pixel positions on the two-dimensional plane. This usually involves the X and Y coordinates of the point, because these two coordinates define the position of the point on the horizontal plane. Orthographically project the preprocessed point cloud data (the data after denoising and clutter removal) along the Z axis onto the XOY plane. In this process, the Z coordinate of each point in the point cloud is ignored, and the X and Y coordinates are used to determine its position on the two-dimensional plane. Convert the X and Y coordinates to pixel positions. This usually involves setting the scale, that is, the conversion ratio between the actual distance and the pixel distance. For example, the point cloud data unit can be set to meters and the image resolution can be set to 5 cm. Create a two-dimensional array or matrix to represent the top-view image. Each pixel position corresponds to an element in the matrix. Fill the pixel values in the image matrix according to the properties of the point in the point cloud data (such as reflection intensity, color, etc.). If multiple points in the point cloud data are projected to the same pixel position in the image, the final value of the pixel can be determined by the average, maximum, or other suitable method. Render an image based on the filled pixel values. This image is the two-dimensional top view corresponding to the point cloud data. Through the above steps, a top view containing the location information of all important traffic identifiers can be obtained. This top view can be used for a variety of applications, such as road monitoring, automated driving assistance systems, traffic planning, etc. This 2D representation simplifies complex 3D data, making the detection and recognition of traffic identifiers much easier.

Based on the above embodiment, the parameters involved in the above conversion process are as follows:

(1) The X-axis of the laser point cloud is the forward direction of the car, the Y-axis is to the left, and the Z-axis is upward;

(2) The X, Y, and Z coordinates of the laser point cloud can be either positive or negative;

(3) The X-axis of the image starts from the origin in the upper left corner and increases to the right, while the Y-axis increases downward;

(4) The X and Y coordinates of the image are both positive numbers;

(5) Pitch angle: The angle formed by connecting the coordinates of the laser point cloud with the coordinates of the origin and the XOY plane is the pitch angle. The formula for the pitch angle is:

Pitch angle = arcsin (the distance between the point cloud coordinate point and the projection point on the XOY plane / the distance between the point cloud coordinate point and the origin)

(6) Deflection angle: The angle between the coordinates of the laser point cloud and the projection point of the XOY platform and the origin and the X-axis is the deflection angle. The deflection angle formula is: deflection angle = arctan (the distance from the projection point of the point cloud coordinates to the X-axis/the distance from the projection point of the point cloud coordinates to the Y-axis).

Point cloud data is a set of discrete points in three-dimensional space acquired by a laser scanner (such as a laser radar or other sensor). Each point contains its xyz coordinate information, RGB color, grayscale value, etc. Common point cloud coordinate systems include scanner coordinate system, inertial navigation coordinate system, local horizontal coordinate system, and Earth-centered Earth-fixed coordinate system. Point clouds in different coordinate systems need to be converted to the same coordinate system for processing. Coordinate conversion is generally achieved through matrix transformation. If the point cloud data is in meters and a resolution of 5 cm (res = 0.05) is required, the implementation is as follows:

To convert point cloud coordinates to image coordinates, you need to take the negative value of the Y-axis coordinate and divide it by the resolution to convert it to the X-coordinate of the image. To convert point cloud coordinates to image coordinates, you need to take the negative value of the X-axis coordinate and divide it by the resolution to convert it to the Y-coordinate of the image. The conversion formula is:

imageX＝(-y_points/res).astype(np.int32)

imageY＝(-x_points/res).astype(np.int32)

Since the X and Y coordinates of the point cloud can be negative, and the image coordinates must be positive, the image coordinates need to be offset to ensure that all coordinates are positive. Assuming that the span range of left and right and front and back is 10 meters, then leftRight＝(-20,20), fowardBackward＝(-20,20), and the image coordinates are offset so that the origin (0,0) corresponds to the upper left corner of the image. The offset formula is:

imageX-=int(np.floor(leftRight[0]/res))

imageY+=int(np.ceil(fowardBackward[1]/res)).

Through the above steps, the three-dimensional point cloud data can be converted into two-dimensional image coordinates and the corresponding top view can be generated. The image will be used for further analysis and processing, such as traffic sign detection, road condition assessment, etc.

In some embodiments, the top view and point cloud data are loaded into the corresponding top view job window and laser point cloud window, respectively. For example, the top view is a two-dimensional image generated from point cloud data. This image provides a plan view of ground features. To load the top view, you first need to start the software or application used and open the top view job window. This window is specially designed to display and edit two-dimensional map data. Then, the user imports the top view file into the job window through the file menu or drag and drop operation. The file can be a format supported by a geographic information system (GIS) (such as GeoTIFF). After loading, the top view is displayed in the job window. The user can zoom in, zoom out, and pan the view, as well as perform further image processing or map editing tasks. Point cloud data is generated by a laser scanner and contains the three-dimensional coordinates of a large number of points in space. These points are combined to form an accurate three-dimensional representation of the scanned object or scene. Loading point cloud data usually requires specialized software or modules that can process and render a large number of three-dimensional coordinate points. The user needs to open the laser point cloud window, which is an interface designed for displaying and analyzing three-dimensional point cloud data. Point cloud data files (such as LAS, LAZ, PLY, etc.) are imported into the laser point cloud window through the file menu or drag and drop operation. When the point cloud data is loaded, the user can view the point cloud in three-dimensional space, rotate the perspective, zoom in, zoom out, and select and highlight specific data points. Through these two windows, the user can view the top view of the ground and/or the three-dimensional view of the point cloud data. This combination of views can help users better understand and analyze ground features and terrain, especially in urban planning, land surveying, environmental monitoring and other professional fields that require accurate ground information.

In some embodiments, determining the target area image in the top view operation window includes: responding to user operations, recording the operation start point and operation end point in the top view, and generating the target area image in the top view operation window based on the operation start point and operation end point in the top view. For example, when creating a top view operation window, the user focuses on a specific area, which usually involves selecting a target area in the top view. The user specifies a specific area in the top view through a graphical user interface (GUI) or other input methods, such as mouse clicks, touch screen operations, or command input. After receiving the user operation instruction, the coordinates of the operation start point and operation end point specified by the user are recorded. In the top view, the user can define a rectangular area, which is usually completed by selecting two points on the diagonal of the rectangle. A corner of the rectangular area defined by the user in the top view (e.g., the upper left corner). The diagonal corner of the rectangular area defined by the user in the top view (e.g., the lower right corner). When the operation start point and end point are recorded, the size and position of the target area are calculated according to the coordinates of the two points. The target area image is a subset of the user-specified area in the top view, and the boundary of the area is determined by the operation start point and the operation end point. According to the calculated area size and position, the target area is cropped from the original top view to generate the target area image. Through the above steps, users can effectively filter out specific areas from the entire top view for detailed observation and analysis, which is very useful for processing large-scale point cloud data and focusing on specific details.

In some embodiments, determining the target area image in the laser point cloud window includes: in response to user operation, recording the operation start point and operation end point in the point cloud data, and generating the target area image in the laser point cloud window based on the operation start point and operation end point in the point cloud data. For example, when creating a laser point cloud window, the user focuses on a specific area, which usually involves selecting a target area in the point cloud data. The user selects an area in the point cloud data set through an interactive interface (such as mouse click, drag, etc.). This usually occurs in a three-dimensional visualization tool, where the user can freely navigate and select in three-dimensional space. The user's selection behavior triggers the software to record the selected area. In three-dimensional space, this usually means selecting a cube or cuboid area. The point marked in the point cloud data set when the user starts selecting the area defines a boundary of the target area. The point marked in the point cloud data set when the user ends selecting the area, together with the operation start point, defines the opposite boundary of the target area. When the operation start point and end point are determined, the spatial range between the two points is calculated, and this range defines a three-dimensional sub-area, namely the target area. Then all point cloud data in this three-dimensional sub-area are extracted to create a target area image. This involves filtering, cropping, and downsampling the point cloud data. The target region image is a subset of the point cloud data that contains only the region of interest to the user.

Based on the above embodiment, after the user creates a top view window or a laser point cloud window, the user can select the corresponding traffic identifier button according to actual needs. After the traffic identifier is selected, the corresponding recognition strategy can be triggered to identify the above-mentioned target area image. For example, the button corresponding to the traffic identifier includes any of the following: a traffic sign button, a road marking button, a parking space button, a zebra crossing button, a no-parking zone button, and a traffic light button. Specifically, after the user creates a top view window or a laser point cloud window, the user identifies and analyzes specific traffic identifiers in these views. To achieve this, the user interface usually provides a series of buttons, each corresponding to a different type of traffic identifier. The user interface provides a set of buttons, each representing a traffic identifier, such as a traffic sign, a road marking, a parking space, a zebra crossing, a no-parking zone, and a traffic light. The user clicks the corresponding button to select the type of traffic identifier to be identified according to actual needs. When the user selects a specific traffic identifier button, the system activates the recognition strategy associated with the button. The recognition strategy is a predefined algorithm or a series of steps specifically used to detect and identify selected types of traffic identifiers. After a traffic identifier is selected and a recognition strategy is triggered, the system automatically applies this strategy to the previously defined target area image. The recognition process includes techniques such as image processing, pattern recognition, machine learning or deep learning to ensure accuracy.

In some embodiments, generating vector data corresponding to a traffic identifier based on a target area image in a top view operation window includes: calling a recognition strategy corresponding to a button according to a button corresponding to a traffic identifier selected by a user; and performing recognition processing on the target area image in the top view operation window through the recognition strategy corresponding to the button to obtain vector data corresponding to the traffic identifier. For example, in the top view operation window, generating vector data corresponding to a traffic identifier is a process from recognition to data conversion. This process involves user interface interaction, image recognition technology, and vectorization processing. In the top view operation window, a user can browse to a plurality of buttons representing different traffic identifiers on a user interaction interface. The user clicks a corresponding button according to the type of traffic identifier to be recognized. For example, if the user wants to recognize a traffic light, the user can click the "traffic light button" at this time. When the user clicks the button, the system calls a preset recognition strategy according to the button. Each button is associated with a specific recognition strategy, which is specially configured for recognizing a specific type of traffic identifier. The recognition strategy analyzes the target area image in the top view operation window. Through this process, complex point cloud or image data can be converted into a more accurate and usable vector format, thereby achieving efficient data utilization and analysis in various applications.

Based on the above embodiment, the target area image in the overhead view operation window is identified and processed by the recognition strategy corresponding to the button, and the vector data corresponding to the traffic identifier is obtained, including: matching the target area image in the overhead view operation window with the preset traffic identifier template to determine multiple key pixel points in the target area image; assigning elevation to the initial coordinates of each key pixel point to obtain the target coordinates of each key pixel point; and generating the vector data corresponding to the traffic identifier based on the target coordinates of each key pixel point. For example, in the overhead view operation window, the system first needs to identify the traffic identifier. This is usually achieved by matching the target area image with a preset traffic identifier template. The template is a standardized image or feature set for a specific traffic identifier, which can be a combination of shape, color, symbol, etc. The matching process involves image processing technology, such as feature detection, edge recognition, template matching algorithm, etc., to determine the part of the image that matches the template. When the part matching the template is found, the system then determines the key pixel points in the target area image. These key pixel points are important pixels that constitute the boundaries and features of the traffic identifier, and their positions are clear in the image. After identifying the key pixel points, the system needs to convert these two-dimensional coordinates into coordinates in three-dimensional space. The initial coordinates of each key pixel point are assigned an elevation value, that is, an elevation value (z) is added to the known two-dimensional coordinates (x, y), so as to obtain the target coordinates (x, y, z) in three-dimensional space. The elevation value can come from terrain data, point cloud data or other elevation information sources. After obtaining the three-dimensional coordinates of the key pixel points, the system converts these points into vector data. The generated vector data can accurately describe the shape, size and position of traffic identifiers, and can be used in a variety of applications such as navigation, map making and urban planning. Through this process, users can extract accurate vector data from the image data in the overhead view operation window, and then use it in a variety of professional applications to improve work efficiency and data practicality.

In some embodiments, generating vector data corresponding to a traffic identifier based on a target area image in a laser point cloud window and displaying the vector data in a laser point cloud window includes: calling a recognition strategy corresponding to a button according to a button corresponding to a traffic identifier selected by a user; and performing recognition processing on the target area image in the laser point cloud window through the recognition strategy corresponding to the button to obtain the vector data displayed in the laser point cloud window. For example, a user selects a type of traffic identifier to be identified in a laser point cloud window, which is usually done by clicking a button corresponding to a specific traffic identifier. Each button is bound to a predefined recognition strategy, which is configured to identify and process the type of traffic identifier corresponding to the button. When a user clicks a specific button, the system calls the recognition strategy bound to the button. The recognition strategy includes algorithms and processing steps, such as feature extraction, pattern recognition, machine learning models, etc., all of which are intended to accurately identify traffic identifiers in point cloud data. The recognition strategy is applied to the target area image in the laser point cloud window, which is generated by point cloud data and reflects the three-dimensional structure in the actual environment. The system analyzes the point cloud data and identifies a point cloud set corresponding to the traffic identifier. This involves point cloud segmentation, clustering, and comparison with predefined models. When traffic identifiers are identified in the point cloud, the system converts the identified point cloud into vector data. The generated vector data is then displayed in the laser point cloud window, which can be superimposed on the original point cloud image or displayed as a different layer.

Based on the aforementioned embodiment, the target area image in the laser point cloud window is identified and processed by the recognition strategy corresponding to the button, and the laser point cloud window display vector data includes: matching the target area image in the laser point cloud window with the preset traffic identifier template to determine multiple key pixels in the target area image; assigning elevation to the initial coordinates of each key pixel to obtain the target coordinates of each key pixel; generating vector data corresponding to the traffic identifier based on the target coordinates of each key pixel. For example, the user first selects a specific traffic identifier button in the laser point cloud window. Each button is associated with a recognition strategy, which contains algorithms and parameter settings for detecting and identifying a specific type of traffic identifier. When the button is triggered, the corresponding recognition strategy is called to prepare for processing the target area image. The recognition strategy first needs to analyze the target area image in the laser point cloud window to detect the traffic identifier. This usually involves matching the target area image with a preset traffic identifier template. These templates are standardized representations of traffic identifiers and can contain features such as shape, color, and pattern. The matching process uses image recognition and computer vision technologies such as edge detection, feature point matching, and deep learning models. When a traffic identifier matching the template is identified in the image, the system determines the key pixels in the target area image. Key pixels are important pixels that constitute the boundary and internal features of the traffic identifier, and they are the key output of the recognition process. After the key pixels are identified, the next step is to map these points from 2D space to 3D space. This is done by assigning an elevation value to each key pixel, which is usually derived from the laser point cloud data itself, because the point cloud data provides the 3D coordinates (x, y, z) of each point. In this way, the initial 2D coordinates (x, y) of each key pixel are expanded into 3D coordinates (x, y, z). After obtaining the 3D coordinates, the system converts these points into vector data format. Vector data describes the exact location, shape and size of the traffic identifier and can be integrated with other Geographic Information System (GIS) data for various applications. The generated vector data is then displayed in the laser point cloud window. This allows the user to visually see the location and shape of the traffic identifier and perform further analysis or editing. This process enables the traffic identifier to be extracted from the laser point cloud data and converted into vector data that can be used for a variety of purposes. This data is useful for urban planning, updating navigation systems and self-driving cars.

Fig. 3 is a schematic diagram of an application scenario of an embodiment of the present application. In this embodiment, the point cloud data acquired by the laser scanner can be uploaded to the vehicle equipment for processing, or uploaded to the backend server for processing.

Taking the background server as an example, after the background server receives the point cloud data, it can perform the following steps:

1. Preprocess the point cloud data to remove noise and clutter;

2. Generate a top view based on the preprocessed point cloud data;

3. Start the production system, which loads the top view and pre-processed point cloud data;

4. Users can select elements to be identified on the interactive interface, including traffic signs, zebra crossings, road markings, parking spaces, traffic lights, no-parking zones and other traffic identifier buttons;

5. In the top view or laser point cloud window, the user can left-click on the interactive interface to select a starting point, then left-click again to select an end point, and finally right-click to end. In this way, the recognition area in the operation window can be determined, that is, the above-mentioned target area image; for example, the coordinates selected by the user are obtained, and the borders of different proportions are automatically generated according to the traffic identifier button category. In order to avoid the inaccuracy of the points selected by the user, the program automatically expands outward by 1 meter to ensure the integrity of data recognition. According to the border range, the image is intercepted from the top view.

6. The background automatically recognizes the traffic identifier button selected by the user and the target area image selected by the user, and obtains the coordinates of the key points in the target area image. For example, according to the recognition type selected by the user, different models of image recognition are started. For example, if the user selects a road arrow, the road arrow template is compared with the image selected by the user, and the best matching result is output.

7. Since different models have different coordinate points, the pixel coordinates of the key points of the image are converted into point cloud coordinates. For example, the key point coordinates are converted by assigning elevations to the generated key point coordinates. The converted key coordinate points are connected in a clockwise order to generate surface data, that is, vector data of the traffic identifier in the target area image. Among them, the converted key point coordinates are assigned elevations, and the matching points corresponding to the point cloud are preferentially found according to the key points. If no matching points are found, the elevations are assigned according to the adjacent point rule.

8. Display the connected surface data in the laser point cloud window and/or the top view operation window.

By combining the overhead view and point cloud data and taking advantage of automated recognition and manual editing, the efficiency of producing vector data corresponding to traffic identifiers can be improved while ensuring accuracy. This is very valuable for projects that require large amounts of geospatial data processing.

FIG4 is a block diagram of a vector data generating device according to an embodiment of the present application. As shown in FIG4 , the vector data generating device 400 includes an acquisition module 410 , a first determination module 420 , a loading module 430 , a second determination module 440 and a generation module 450 .

Specifically, the acquisition module 410 is used to acquire point cloud data, where the point cloud data is obtained by acquiring the external environment of the vehicle through a vehicle-mounted laser scanner.

The first determining module 420 is used to determine the top view corresponding to the point cloud data according to the point cloud data.

The loading module 430 is used to load the top view and point cloud data into the corresponding top view operation window and laser point cloud window respectively.

The second determination module 440 is used to determine the target area image in the top view operation window, or determine the target area image in the laser point cloud window, and the target area image includes a traffic identifier.

Generation module 450 is used to generate vector data corresponding to the traffic identifier based on the target area image in the overhead view operation window, and display the vector data in the overhead view operation window; or to generate vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and display the vector data in the laser point cloud window.

The vector data generating device 400 can avoid the repetitiveness and inefficiency in the traditional manual operation process by using the point cloud data obtained by the vehicle-mounted laser scanner. Automated point cloud data acquisition and processing significantly improves the speed of data acquisition and reduces the demand for human resources, thereby achieving high efficiency in map data production. By generating a bird's-eye view using point cloud data, it is ensured that the vector data of the traffic identifier is highly matched with the actual environment, and the error caused by imprecise manual operation is reduced. This method improves the accuracy of map data and ensures the quality and reliability of the map. By loading the point cloud data and the bird's-eye view into the corresponding operation window respectively, the present invention simplifies the operation process, so that the operator can more intuitively identify and process the traffic identifier. Even in the case of obstructions, the required point cloud data can be quickly and accurately selected by adjusting the viewing angle, reducing the requirements for the skills of the operator. In addition, the present application allows the operator to flexibly switch between the bird's-eye view operation window and the laser point cloud window, and select the most suitable view to work according to different situations. This flexibility enables the operator to better process data in complex scenes, improves the adaptability of work and the ability to cope with complex situations.

In some embodiments, the first determination module 420 is configured to: convert the point cloud data by a preset conversion method to obtain pixel position data corresponding to the point cloud data; and generate a top view corresponding to the point cloud data based on the pixel position data corresponding to the point cloud data.

In some embodiments, before determining the overhead view corresponding to the point cloud data based on the point cloud data, the vector data generating device 400 is also used to: denoise the point cloud data to obtain denoised point cloud data; and remove clutter from the denoised point cloud data to obtain target point cloud data.

In some embodiments, the second determination module 440 is configured to: respond to user operations, record the operation start point and operation end point in the overhead view, and generate a target area image in the overhead view operation window based on the operation start point and operation end point in the overhead view.

In some embodiments, the second determination module 440 is further configured to: in response to user operations, record the operation start point and operation end point in the point cloud data, and generate a target area image in the laser point cloud window based on the operation start point and operation end point in the point cloud data.

In some embodiments, the generation module 450 is configured to: call the recognition strategy corresponding to the button corresponding to the traffic identifier selected by the user; identify and process the target area image in the overhead view operation window through the recognition strategy corresponding to the button to obtain vector data corresponding to the traffic identifier; generate vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and display the vector data in the laser point cloud window, including: call the recognition strategy corresponding to the button corresponding to the traffic identifier selected by the user; identify and process the target area image in the laser point cloud window through the recognition strategy corresponding to the button to obtain vector data displayed in the laser point cloud window.

In some embodiments, the target area image in the overhead view operation window is identified and processed through the recognition strategy corresponding to the button, and the vector data corresponding to the traffic identifier is obtained, including: matching the target area image in the overhead view operation window with a preset traffic identifier template to determine multiple key pixel points in the target area image; assigning elevations to the initial coordinates of each key pixel point to obtain the target coordinates of each key pixel point; generating vector data corresponding to the traffic identifier based on the target coordinates of each key pixel point; identifying and processing the target area image in the laser point cloud window through the recognition strategy corresponding to the button, and obtaining the vector data displayed in the laser point cloud window, including: matching the target area image in the laser point cloud window with a preset traffic identifier template to determine multiple key pixel points in the target area image; assigning elevations to the initial coordinates of each key pixel point to obtain the target coordinates of each key pixel point; generating vector data corresponding to the traffic identifier based on the target coordinates of each key pixel point.

FIG5 is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. As shown in FIG5 , the electronic device 500 according to the embodiment includes: a processor 510, a memory 520, and a computer program 530 stored in the memory 520 and executable on the processor 510. When the processor 510 executes the computer program 530, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 510 executes the computer program 530, the functions of the modules in the above-mentioned device embodiments are implemented.

The electronic device 500 may be an electronic device installed in a vehicle or a background server. The electronic device 500 may include, but is not limited to, a processor 510 and a memory 520. Those skilled in the art will appreciate that FIG. 5 is merely an example of the electronic device 500 and does not constitute a limitation on the electronic device 500, and may include more or fewer components than shown in the figure, or different components.

The processor 510 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

The memory 520 may be an internal storage unit of the electronic device 500, for example, a hard disk or memory of the electronic device 500. The memory 520 may also be an external storage device of the electronic device 500, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 500. The memory 520 may also include both an internal storage unit of the electronic device 500 and an external storage device. The memory 520 is used to store computer programs and other programs and data required by the electronic device.

Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.

If the integrated module is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium (for example, a computer-readable storage medium). Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. The computer program may include computer program code, which may be in source code form, object code form, executable file or some intermediate form. Computer-readable media may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims (10)

1. A vector data generation method, characterized in that the method comprises: Acquiring point cloud data, wherein the point cloud data is point cloud data obtained by acquiring the external environment of the vehicle through a vehicle-mounted laser scanner; Determine, according to the point cloud data, a top view corresponding to the point cloud data; Loading the top view and the point cloud data into the corresponding top view operation window and laser point cloud window respectively; Determine a target area image in the top view operation window, or determine a target area image in the laser point cloud window, wherein the target area image includes a traffic identifier; Generate vector data corresponding to the traffic identifier based on the target area image in the overhead view operation window, and display the vector data in the overhead view operation window; or generate vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and display the vector data in the laser point cloud window. 2. The method according to claim 1, wherein determining the top view corresponding to the point cloud data according to the point cloud data comprises: The point cloud data is converted by a preset conversion method to obtain pixel position data corresponding to the point cloud data; Based on the pixel position data corresponding to the point cloud data, a top view corresponding to the point cloud data is generated. 3. The method according to claim 1, characterized in that before determining the top view corresponding to the point cloud data according to the point cloud data, the method further comprises: Performing denoising processing on the point cloud data to obtain denoised point cloud data; The denoised point cloud data is subjected to clutter removal processing to obtain target point cloud data. 4. The method according to claim 1, characterized in that the determining the target area image in the top view operation window comprises: responding to a user operation, recording an operation start point and an operation end point in the top view, and generating the target area image in the top view operation window based on the operation start point and the operation end point in the top view; Determining the target area image in the laser point cloud window includes: in response to a user operation, recording an operation start point and an operation end point in the point cloud data, and generating the target area image in the laser point cloud window based on the operation start point and the operation end point in the point cloud data. 5. The method according to claim 1, characterized in that the generating the vector data corresponding to the traffic identifier based on the target area image in the overhead view operation window comprises: calling the recognition strategy corresponding to the button according to the button corresponding to the traffic identifier selected by the user; performing recognition processing on the target area image in the overhead view operation window through the recognition strategy corresponding to the button to obtain the vector data corresponding to the traffic identifier; The generating of the vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window and displaying the vector data in the laser point cloud window comprises: calling the recognition strategy corresponding to the button according to the button corresponding to the traffic identifier selected by the user; performing recognition processing on the target area image in the laser point cloud window through the recognition strategy corresponding to the button, so as to obtain the vector data displayed in the laser point cloud window. 6. The method according to claim 5, characterized in that the step of performing recognition processing on the target area image in the overhead view operation window by using the recognition strategy corresponding to the button to obtain the vector data corresponding to the traffic identifier comprises: Based on the target area image in the overhead view operation window, the preset traffic identifier template is matched to determine a plurality of key pixel points in the target area image; the initial coordinates of each of the key pixel points are assigned elevations to obtain the target coordinates of each of the key pixel points; and the vector data corresponding to the traffic identifier is generated based on the target coordinates of each of the key pixel points; By using the recognition strategy corresponding to the button, the target area image in the laser point cloud window is recognized and processed, and the vector data displayed in the laser point cloud window includes: The target area image in the laser point cloud window is matched with a preset traffic identifier template to determine multiple key pixel points in the target area image; the initial coordinates of each key pixel point are assigned elevation to obtain the target coordinates of each key pixel point; and the vector data corresponding to the traffic identifier is generated based on the target coordinates of each key pixel point. 7. The method according to claim 5 is characterized in that the button corresponding to the traffic identifier includes any one of the following: a traffic sign button, a road marking button, a parking space button, a zebra crossing button, a no-parking zone button, and a traffic light button. 8. A vector data generating device, characterized in that the device comprises: An acquisition module is used to acquire point cloud data, wherein the point cloud data is point cloud data obtained by acquiring the external environment of the vehicle through a vehicle-mounted laser scanner; A first determining module, configured to determine a top view corresponding to the point cloud data according to the point cloud data; A loading module, used for loading the top view and the point cloud data into the corresponding top view operation window and laser point cloud window respectively; A second determination module is used to determine a target area image in the top view operation window, or to determine a target area image in the laser point cloud window, wherein the target area image includes a traffic identifier; A generation module is used to generate vector data corresponding to the traffic identifier based on the target area image in the overhead view operation window, and display the vector data in the overhead view operation window; or to generate vector data corresponding to the traffic identifier based on the target area image in the laser point cloud window, and display the vector data in the laser point cloud window. 9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 7 when executing the computer program. 10. A readable storage medium storing a computer program, wherein the computer program implements the steps of the method according to any one of claims 1 to 7 when executed by a processor.