WO2022143114A1

WO2022143114A1 - Static map generation method and apparatus, and computer device and storage medium

Info

Publication number: WO2022143114A1
Application number: PCT/CN2021/137379
Authority: WO
Inventors: 黄寅; 张涛
Original assignee: 深圳市普渡科技有限公司
Priority date: 2020-12-31
Filing date: 2021-12-13
Publication date: 2022-07-07
Also published as: CN112799095B; CN112799095A

Abstract

A static map generation method and apparatus, and a computer device and a storage medium. The method comprises: acquiring depth image data and pose data which are collected within the same time period (S10); determining, according to the pose data, pose information which is synchronous with the depth image data (S20); processing the depth image data on the basis of the pose information and an installation position corresponding to the depth image data, so as to generate point cloud data in a robot coordinate system (S30); intercepting the point cloud data according to a pre-set interception rule, so as to generate local point cloud data (S40); updating global obstacle map data according to the local point cloud data, wherein a coordinate system of the global obstacle map data is a world coordinate system (S50); and updating an initial static map according to the global obstacle map data, wherein the updated initial static map is a static map (S60).

Description

Static map generation method, device, computer equipment and storage medium

This application claims the priority of the Chinese patent application submitted to the China Patent Office on December 31, 2020, the application number is 202011640136.0, and the application name is "static map generation method, device, computer equipment and storage medium", the entire content of which is approved by Reference is incorporated in this application.

technical field

The present application relates to the field of robot positioning, and in particular, to a method, device, computer equipment and storage medium for generating a static map.

Background technique

Before the robot works normally, it generally needs to complete the work deployment through positioning and mapping. Localization and mapping can use single-line lidar, multi-line lidar, or depth cameras. The cost of single-line lidar is low, but its perception range is a two-dimensional plane, and the amount of map information generated is low. Multi-line lidar can perceive three-dimensional space, but its high price limits its application range. The depth camera is generally used to build a 3D map, and the generated 3D map will occupy a large storage space. Moreover, for some robots whose activity space is a 2D plane, the 3D map cannot further improve the positioning accuracy of the robot, but increases the storage space. and data processing requirements.

Application content

According to various embodiments of the present invention, a static map generation method, apparatus, computer device, and storage medium are provided.

A static map generation method, including:

Obtain the depth image data and pose data collected in the same time period;

Determine the pose information synchronized with the depth image data according to the pose data;

processing the depth image data based on the pose information and the installation position corresponding to the depth image data to generate point cloud data in the robot coordinate system;

intercepting the point cloud data according to a preset interception rule to generate local point cloud data;

Update global obstacle map data according to the local point cloud data, and the coordinate system of the global obstacle map data is the world coordinate system;

The initial static map is updated according to the global obstacle map data, and the updated initial static map is a static map.

A static map generating device, comprising:

The data acquisition module is used to acquire the depth image data and pose data collected in the same time period;

a synchronous pose information determining module, configured to determine pose information synchronized with the depth image data according to the pose data;

a point cloud generation module, configured to process the depth image data based on the pose information and the installation position corresponding to the depth image data, to generate point cloud data in the robot coordinate system;

a point cloud interception module, configured to intercept the point cloud data according to a preset interception rule to generate local point cloud data;

an obstacle map update module, configured to update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system;

The static map update module is used to update the initial static map according to the global obstacle map data, and the updated initial static map is the static map.

A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, the processor executing the computer-readable instructions to achieve the following static map generation method:

Obtain the depth image data and pose data collected in the same time period;

Update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system;

One or more readable storage media storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform a static map generation method as described above.

The details of one or more embodiments of the application are set forth in the accompanying drawings and the description below. Other features and advantages of the present application will be apparent from the description, drawings, and claims.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments of the present application. Obviously, the drawings in the following description are only some embodiments of the present application. , for those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative labor.

1 is a schematic flowchart of a method for generating a static map in an embodiment of the present application;

2 is an initial static map that has not been updated in an embodiment of the present application;

3 is a local obstacle map in a robot coordinate system in an embodiment of the present application;

4 is a global obstacle map in a world coordinate system in an embodiment of the present application;

5 is a schematic diagram of a depth camera perception area in an embodiment of the present application;

FIG. 6 is a comparison between four local obstacle maps at different times and an initial static map in an embodiment of the present application;

7 is a schematic structural diagram of a static map generating apparatus in an embodiment of the present application;

FIG. 8 is a schematic diagram of a computer device in an embodiment of the present application.

Detailed ways

In order to facilitate understanding of the present application, the present application will be described more fully below with reference to the related drawings. The preferred embodiments of the present application are shown in the accompanying drawings. However, the application may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that a thorough and complete understanding of the disclosure of this application is provided.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field of the invention. The terms used herein in the description of the invention are for the purpose of describing particular embodiments only and are not intended to limit the application. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

In one embodiment, as shown in FIG. 1, a static map generation method is provided, which includes the following steps:

S10. Acquire depth image data and pose data collected in the same time period.

Understandably, the depth image data refers to image data with depth information collected by a depth camera installed on the robot. In some cases, depth image data may also be referred to as RGBD data. The depth image data includes several frames of depth images. Each frame of depth image is marked with a timestamp corresponding to its acquisition time.

The pose data is the position and attitude of the robot (usually represented by an angle) collected by the positioning sensor on the robot. The pose data includes several pieces of pose information. The pose information can include plane coordinates (values on the X, Y axis), angles. Each pose information has its corresponding acquisition time.

The same time period can refer to the acquisition time of the depth image data and the pose data both in the mapping stage of the robot. Since the depth camera and the positioning sensor are independent acquisition devices, their sampling frequencies can be the same or different. Therefore, although the depth image data and the pose data are collected in the same time period, for a certain frame of depth image at a specific time, there may not necessarily have the pose information at the same time, but in the time interval ( If the total length is 10 seconds), there is one or more pose information.

S20. Determine, according to the pose data, pose information synchronized with the depth image data.

Understandably, synchronous refers to being at the same time. For the depth image whose acquisition time is a specified time in the depth image data, if the pose data contains the pose information at the specified time, the pose information at the specified time is directly obtained, which is the pose information synchronized with the depth image data; If the pose data does not contain the pose information at the specified time, the closest first pose information before the specified time and the second closest after the specified time can be obtained. According to the first pose information and the second The pose information estimates the pose information at a specified time, that is, the pose information synchronized with the depth image data.

S30. Process the depth image data based on the pose information and the installation position corresponding to the depth image data to generate point cloud data in the robot coordinate system.

Understandably, the installation position corresponding to the depth image data refers to the installation position of the depth camera on the robot. Usually, the installation position of the depth image data can be converted into a coordinate point in the robot coordinate system. The pose information can determine the orientation of the robot coordinate system in the world coordinate system (for example, it can be the position of the origin of the robot coordinate system in the world coordinate system, and the angle between the X axis of the robot coordinate system and the X axis of the world coordinate system). One point cloud can be generated per frame of depth image. By traversing all depth images (ie depth image data) in chronological order, multiple point clouds can be obtained to form point cloud data. That is, the point cloud data includes several point clouds. In an example, the depth image can be converted into a point cloud in camera coordinates based on the internal parameters of the depth camera, and then converted into a point cloud in the robot coordinate system according to the point cloud data in the camera coordinates of the installation location.

S40. Intercept the point cloud data according to a preset interception rule to generate local point cloud data.

Understandably, in order to ensure the accuracy of the static map, the preset interception rule may stipulate that a point cloud within a specified spatial range in front of the depth camera (that is, in front of the robot) is selected. The specified spatial extent is related to the imaging capabilities of the depth camera. In an example, the preset interception rule may be set to: intercept the point cloud within 2 meters in front of the depth camera, which is higher than the ground and lower than the visual height of the robot. By intercepting the point cloud, a local point cloud can be generated. Local point cloud can greatly reduce the processing volume of point cloud data and improve the quality of point cloud at the same time. The local point cloud data includes several local point clouds.

S50. Update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system.

Understandably, the conversion of local point cloud data to global obstacle map data requires operations of projection and coordinate system switching. You can project first and then switch the coordinate system, or you can switch the coordinate system first and then project. In an example, the local point cloud is first projected on the robot coordinate system to generate a local obstacle map, and then based on the orientation of the robot coordinate system in the world coordinate system determined by the pose information, the local obstacle map is converted into Global obstacle map. When projecting, the pixel value of the grid with point projection is set to 255, and the pixel value of the grid without point projection is set to 0. Multiple local point cloud data can generate multiple global obstacle maps. The global obstacle map data includes several global obstacle maps.

S60. Update the initial static map according to the global obstacle map data, and the updated initial static map is the static map.

Understandably, the global obstacle map data includes several global obstacle maps. The initial static map can be updated using the frame-by-frame global obstacle map in chronological order. In an example, as shown in FIG. 2 , the unupdated initial static map is a grid map with a pixel value of 0. The pixel value of the corresponding grid on the initial static map can be updated based on the size of the pixel value of the grid on the global obstacle map. For example, if the pixel value of grid (100,100) on the global obstacle map is 255, after the update, the pixel value of grid (100,100) on the initial static map is V ₀ +V ₁ ; if the grid on the global obstacle map has a pixel value of V 0 +V 1 ; The pixel value of (100,100) is 0, then after the update, the pixel value of grid (100,100) on the initial static map is V ₀ -V ₂ , where V ₀ is the value of grid (100,100) on the static map before the update. Pixel value, V ₁ and V ₂ are all self-defined amplitudes (positive numbers), which can be set according to actual needs. The pixel value of the grid on the initial static map ranges from 0 to 255. If the updated pixel value exceeds this range, the updated pixel value is set to 0 or 255.

V ₁ and V ₂ respectively control the increase and attenuation of the pixel value of the static map grid. For dynamic obstacles, when the dynamic obstacle does not leave, the pixel value of the grid corresponding to its location will increase. When the dynamic obstacle leaves, the pixel value of the grid corresponding to its location will gradually decay, and finally becomes 0. . Therefore, when the robot is building a map, as long as it is ensured that after the dynamic obstacle leaves, it continues to collect data at the original stop position of the dynamic obstacle, and the interference of the dynamic obstacle can generally be eliminated. However, since the position of the static obstacle does not change, the pixel value of the grid corresponding to the position of the static obstacle will not be attenuated. Therefore, the pixel value of the grid corresponding to the position of the static obstacle is higher.

In steps S10-S60, the depth image data and pose data collected in the same time period are acquired to obtain the original data for generating the static map. The pose information synchronized with the depth image data is determined according to the pose data, so as to determine the association relationship (through time association) between the pose data and the depth image data. The depth image data is processed based on the pose information and the installation position corresponding to the depth image data to generate point cloud data in the robot coordinate system, so as to convert the depth image data into three-dimensional data that is easy to process (ie point cloud data), there is no need to build a 3D model, which greatly saves computing resources. The point cloud data is intercepted according to preset interception rules, and local point cloud data is generated to screen the point cloud. On the one hand, the accuracy of the point cloud is improved, and on the other hand, the amount of data processing is reduced. The global obstacle map data is updated according to the local point cloud data, and the coordinate system of the global obstacle map data is the world coordinate system, so as to process the three-dimensional data into two-dimensional data. The initial static map is updated according to the global obstacle map data, and the updated initial static map is a static map, so as to generate a static map.

Optionally, step S50, that is, updating the global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system, including:

S501. Update local obstacle map data according to the local point cloud data, where the coordinate system of the local obstacle map data is a robot coordinate system;

S502. Convert the updated local obstacle map data into the global obstacle map data.

Understandably, updating the local obstacle map data according to the local point cloud data is to project each point in the point cloud to the ground plane of the robot coordinates (the plane formed by the X axis and the Y axis). When all points in the point cloud are projected, the update of the local obstacle map data is completed. During projection, on the local obstacle map, the corresponding pixel value of the grid projected by the point cloud is set to 255, and the corresponding pixel value of the grid not projected by the point cloud remains the original value of 0. As shown in FIG. 3 , FIG. 3 is a local obstacle map in the robot coordinate system.

The updated local obstacle map data can be converted into global obstacle map data according to the orientation of the robot coordinate system in the world coordinate system determined by the pose information. The coordinate system of the global obstacle map data is the world coordinate system. As shown in FIG. 4 , FIG. 4 is a global obstacle map in the world coordinate system.

Optionally, before step S10, that is, before the acquiring the depth image data and pose data that are collected synchronously, the method further includes:

S11. Acquire a plurality of initial pose data collected at preset condition intervals during the mapping process;

S12. Correct the plurality of initial pose data by using a preset loop closure detection algorithm to obtain the pose data.

Understandably, the initial pose data is uncorrected data. The initial pose data is the data collected when the robot performs the mapping process. The methods that can be used for robot mapping include, but are not limited to, lidar-based mapping methods and two-dimensional code-based mapping methods.

Optionally, in order to reduce the amount of calculation, the initial pose data is collected at preset condition intervals during the mapping process. Optionally, the preset condition intervals can be time intervals, for example, collected every 0.5S. , it can also be a frame interval, for example, it is collected every 10 frames, it can also be an angle interval, for example, it can be collected once every time the change amplitude is greater than 30 degrees compared with the last collection, or it can be a distance interval, for example, every time the run exceeds 0.5 m can be collected once, etc., which is not limited here. Therefore, during the whole mapping process, multiple initial pose data will be collected, such as 10 or 20, and so on.

Generally, in the process of building a robot, there is a process of loopback detection and optimization. In the process of implementing the robot, the robot can obtain its current low-accuracy initial pose data according to the odometer. The initial pose data is greatly influenced by sensors or processing algorithms. When the robot pushes back to the origin, due to the existence of errors, the robot may think that there is still a large deviation from the origin. Through the loopback detection function, the robot can make clear the position information of the end point (that is, the origin), and correct multiple initial pose data by calculating the deviation between the measured value and the real value, and obtain the pose data with higher accuracy. Optional, Correction of each initial pose data will form corrected pose data, and multiple corrected pose data form the pose data.

Optionally, step S20, that is, determining the pose information synchronized with the depth image data according to the pose data, including:

S201, acquiring the acquisition time of the depth image data;

S202, judging whether the pose data has pose information at the collection time;

S203. If there is no pose information at the collection time in the pose data, process the pose data through an interpolation algorithm to generate pose information at the collection time, and the position at the collection time The pose information and the pose data are the pose information synchronized with the depth image data.

Understandably, the acquisition time of the depth image data, that is, the generation time of each depth image, can be represented by a timestamp. That is, the acquisition time of the depth image data includes timestamps of several depth images. The pose data includes several pose information. Each pose information has its corresponding acquisition time. The collection time of pose information can also be represented by a timestamp.

Whether there is corresponding pose information in the pose data can be searched according to the timestamp of the depth image. If there is, the pose information is the pose information synchronized with the depth image data. If it does not exist, the pose information corresponding to the timestamp of the depth image can be determined through an interpolation algorithm. Specifically, the pose data at the collection time may be generated by using the pose data whose time relationship is most adjacent to the collection time.

For example, it is possible to obtain the nearest corrected pose data before the specified time stamp, the nearest corrected pose data after the specified time stamp, and infer the pose information of the specified time stamp according to the two corrected pose data, which is the same as the depth image data. Synchronized pose information.

Optionally, the pose information and pose data at the acquisition time are the pose information synchronized with the depth image data. Among them, the pose data corresponding to the acquisition time is selected as the pose information synchronized with the depth image data.

Optionally, step S40, that is, intercepting the point cloud data according to a preset interception rule to generate local point cloud data, including:

S401. Determine the facing position of the robot according to the pose information;

S402. Determine a target area according to the facing position, and extract the local point cloud data from the point cloud data according to the target area, where the target area includes within a specified distance in front of the robot, higher than the ground and low The perception area at the height of the robot's vision.

Understandably, the pose information includes the plane coordinates and angle information of the robot, and the facing position of the robot can be determined through the angle information in the pose information. In some cases, the angle information may include plane rotation angle (horizontal direction) and/or elevation angle (vertical direction). In an example, it can be defined that the angle is 0°, the facing position is rightward, and the angle is 180°, and the facing position is leftward.

The sensing area (ie sensing range) of the depth camera can be determined by the facing position of the robot. The perception area of the depth camera may be the imaging area of the camera. And the target area is less than or equal to the perception area of the depth camera. As shown in FIG. 5 , FIG. 5 is a schematic diagram of the perception area of the depth camera. In Figure 5, since the point cloud data is only distributed in the perceptual range (sector area) of RGBD (representing the depth camera here), when using the local point cloud data to update the local obstacle map data, only the perceptual range of the RGBD is required. The grid is updated, and the RGBD perceptual blind area is not updated.

The target area includes the perception area within the specified distance in front of the robot, higher than the ground and lower than the robot's visual height. The specified distance can be set according to actual needs. In an example, the target area can be set as an area within 2 meters in front of the depth camera, higher than the ground and lower than the vision height of the robot. By intercepting the point cloud, a local point cloud can be generated. Local point cloud can greatly reduce the processing volume of point cloud data and improve the quality of point cloud at the same time. Here, the local point cloud data includes several local point clouds.

Optionally, step S501, that is, updating the local obstacle map data according to the local point cloud data, where the coordinate system of the local obstacle map data is a robot coordinate system, including:

S5011, project the local point cloud data to the local obstacle map on the ground plane, and set the pixel value of the grid where the point projection exists as the specified pixel value;

S5012. After all points in the local point cloud data are projected, complete the update of the local obstacle map data.

Understandably, each point in the point cloud can be projected to the ground plane of the robot coordinates (the plane formed by the X axis and the Y axis). When all points in the point cloud are projected, the update of the local obstacle map data is completed. During projection, on the local obstacle map, the corresponding pixel value of the grid projected by the point cloud is set to 255, and the corresponding pixel value of the grid not projected by the point cloud remains the original value of 0. As shown in FIG. 3 , FIG. 3 is a local obstacle map in the robot coordinate system.

Optionally, in step S60, the initial static map is updated according to the global obstacle map data, and the updated initial static map is a static map, including:

S601. If the pixel value of the first specified grid in the global obstacle map data is greater than 0, increase the pixel value of the first static grid by a first preset amplitude, and increase the first preset amplitude The pixel value of the subsequent first static grid is not greater than 255, and the first static grid is the grid corresponding to the first specified grid in the initial static map;

S602. If the pixel value of the second specified grid in the global obstacle map data is 0, reduce the pixel value of the second static grid by a second preset amplitude, and reduce the second preset amplitude The pixel value of the subsequent second static grid is not less than 0, and the second static grid is a grid corresponding to the second specified grid in the initial static map.

Understandably, the initial static map may be updated frame by frame among multiple global obstacle maps in the global obstacle map data in chronological order. A static map is available when all global obstacle maps are updated. In the static map, the pixel value of the grid without obstacles is 0, which is black, and the pixel value of the grid with obstacles is greater than 0. The larger the pixel value, the higher the probability that the obstacle exists. Thus, grids with obstacles appear gray or white on the static map.

When the initial static map is updated according to the global obstacle map, if the pixel value of the first specified grid in the global obstacle map is greater than 0 (usually 255), the pixel value of the first static grid is increased by the first preset value. and the pixel value of the first static grid after increasing the first preset amplitude value is not greater than 255, and the first static grid is the grid corresponding to the first specified grid in the initial static map. The first preset amplitude is a positive number, which can be set according to actual needs. The size of the first preset amplitude is inversely proportional to the number of frames of the global obstacle map per second, and inversely proportional to the brightening time (the time for the pixel value to continuously increase from 0 to 255).

If the pixel value of the second specified grid in the global obstacle map data is 0, reduce the pixel value of the second static grid by a second preset amplitude, and reduce the second static grid by the second preset amplitude. The pixel value of the grid is not less than 0, and the second static grid is the grid corresponding to the second specified grid in the initial static map. The second preset amplitude is a positive number, which can be set according to actual needs. The size of the second preset amplitude is inversely proportional to the number of frames of the global obstacle map per second, and inversely proportional to the decay time (the time for the pixel value to continuously decrease from 255 to 0).

In an example, as shown in FIG. 6 , the first row of images in FIG. 6 is a local obstacle map at four different times, and the second row of images is an initial static map at four different times. Among them, the position pointed by the white arrow is the actual position of the robot at the corresponding moment (T=0-3). The perception range of the robot is ±90 degrees in front, and it drives forward along the X-axis direction. There is an L-shaped static obstacle and a point-shaped dynamic obstacle in front of the robot, and the dynamic obstacle also travels forward along the X-axis direction. The following is the update process of the initial static map at 4 moments.

At time T=0, the robot observes the L-shaped and point-shaped obstacles ahead (the upper left in Figure 6), and transfers it to the world coordinate system to generate the corresponding global obstacle map (not shown in Figure 6, for reference Figure 3 and Figure 4), the corresponding position grid value begins to increase V ₁ (Figure 6 bottom left).

At the moment of T=1, the robot moves forward by a grid, and the dynamic obstacle also moves forward by a grid (the second upper left of Figure 6). Since a dynamic obstacle is observed at the new position, the grid value corresponding to this position is increased by V ₁ , but no object was observed at the position where the dynamic obstacle was located in the previous frame, and the grid value corresponding to this position is attenuated by V ₂ , namely darken. For an L-shaped static obstacle, since its location is the same as the previous frame, the grid value of its location continues to increase by V ₁ , that is, it becomes brighter (the second lower left in Figure 6 ).

At time T=2, the robot continues to move forward by a grid, and the dynamic obstacle also continues to move forward by a grid (the second upper right in Figure 6). Since a dynamic obstacle is observed at the new position, the grid value corresponding to this position is increased by V ₁ , but no object is observed at the position where the dynamic obstacle is located in the previous two frames, and the corresponding grid value is attenuated by V ₂ , that is, it becomes darker . For an L-shaped static obstacle, since its location is the same as the previous frame, the grid value of its location continues to increase by V ₁ , that is, it becomes brighter (the second bottom right of Figure 6 ).

At time T=3, the robot continues to move forward by a grid, and the dynamic obstacle also continues to move forward by a grid (the upper right of Figure 6). Since a dynamic obstacle is observed at the new position, the grid value corresponding to this position is increased by V ₁ , but no object is observed at the position where the dynamic obstacle is located in the previous three frames, and the corresponding grid value is attenuated by V ₂ , that is, it becomes darker , and the grid value of the position of the dynamic obstacle at T=0 has decayed to 0, that is, it is the same as the background. For the L-shaped static obstacle, its leftmost part has entered the blind spot of the robot, the grid value is no longer updated, and the rest is still within the robot's perception range, and its position is the same as the previous frame, so its location grid The value continues to increase V ₁ , i.e. it becomes brighter (lower right in Figure 6).

Optionally, after step S60, that is, after the initial static map is updated according to the global obstacle map data, the updated initial static map is the static map, and further includes:

S61. Obtain code map data;

S62. Generate a topology path of the robot according to the code map data and the static map.

Understandably, the code map data may refer to the visual sign recognition data collected during the process of the robot's map building. Before the robot builds a map, a number of visual markers (generally pasted on the ceiling) are pasted on the robot's driving path. The code map data can determine the position of the robot. The static map contains obstacle information. According to the code map data and static map, the feasible path of the robot can be planned, that is, the topology path.

Optionally, the code map data is the map formed by the robot based on the two-dimensional code, etc. The map does not consider the obstacle information, so if the topological path is only based on the map, the robot will not be able to run the map. The impact of obstacles is considered in advance (and the immediate avoidance will only be performed when it is detected soon, which will affect the entire avoidance time and path), thereby affecting the operation effect. The present application constructs a static map, and uses the obstacle information in the static map to form a topological path together with the code map data, that is, when planning the topological path, the obstacle information is taken into account, so that the robot can be Obstacles will be considered in advance, that is, the avoidance method will be provided in advance, so as to produce a complete avoidance path, thereby effectively improving the operation effect.

It should be understood that the size of the sequence numbers of the steps in the above embodiments does not mean the sequence of execution, and the execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

In one embodiment, an apparatus for generating a static map is provided, and the apparatus for generating a static map is in one-to-one correspondence with the method for generating a static map in the above-mentioned embodiment. As shown in FIG. 7 , the static map generation device includes a data acquisition module 10 , a synchronized pose information determination module 20 , a point cloud generation module 30 , a point cloud interception module 40 , an obstacle map update module 50 and a static map update module 60 . The detailed description of each functional module is as follows:

an acquisition data module 10 for acquiring depth image data and pose data collected in the same time period;

a synchronous pose information determining module 20, configured to determine pose information synchronized with the depth image data according to the pose data;

A point cloud generating module 30 is configured to process the depth image data based on the pose information and the installation position corresponding to the depth image data, and generate point cloud data in the robot coordinate system;

A point cloud interception module 40, configured to intercept the point cloud data according to a preset interception rule to generate local point cloud data;

an obstacle map update module 50, configured to update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system;

The static map updating module 60 is configured to update the initial static map according to the global obstacle map data, and the updated initial static map is the static map.

Optionally, the obstacle map update module 50 includes:

a local map update unit, configured to update local obstacle map data according to the local point cloud data, where the coordinate system of the local obstacle map data is a robot coordinate system;

A global map conversion unit, configured to convert the updated local obstacle map data into the global obstacle map data.

Optionally, the static map generating apparatus further includes:

an initial data acquisition module for acquiring initial pose data including several corrected pose data;

A data correction module, configured to correct the initial pose data according to the corrected pose data to generate the pose data.

Optionally, the determining synchronization pose information module 20 includes:

an acquisition time unit for acquiring the acquisition time of the depth image data;

a pose information unit for determining whether the pose data has pose information at the collection time;

A pose information generating unit is used for, if the pose data does not have pose information at the collection time, process the pose data through an interpolation algorithm to generate pose information at the collection time, and the pose data at the collection time is generated. The pose information at the acquisition time is the pose information synchronized with the depth image data.

Optionally, the point cloud interception module 40 includes:

determining a facing unit for determining the facing position of the robot according to the pose information;

a point cloud intercepting unit, configured to determine a target area according to the facing position, and extract the local point cloud data from the point cloud data according to the target area, and the target area includes within a specified distance in front of the robot, The perception area that is above the ground and below the robot's visual height.

Optionally, the local map update unit includes:

a projection unit, used to project the local point cloud data to the local obstacle map on the ground plane, and set the pixel value of the grid where the point projection exists as the specified pixel value;

The completion of the local map update unit is used to complete the update of the local obstacle map data after all points in the local point cloud data are projected.

Optionally, the static map update module 60 includes:

A first update unit, configured to increase the pixel value of the first static grid by a first preset amplitude if the pixel value of the first specified grid in the global obstacle map data is greater than 0, and increase the pixel value of the first specified grid The pixel value of the first static grid after a preset amplitude is not greater than 255, and the first static grid is the grid corresponding to the first designated grid in the initial static map;

The second updating unit is configured to reduce the pixel value of the second static grid by a second preset magnitude if the pixel value of the second specified grid in the global obstacle map data is 0, and decrease the pixel value of the first grid. The pixel value of the second static grid after two preset amplitudes is not less than 0, and the second static grid is a grid corresponding to the second specified grid in the initial static map.

Optionally, the static map generating apparatus further includes:

Get the code map data module, which is used to obtain the code map data;

A topological path generating module is configured to generate a topological path of the robot according to the code map data and the static map.

For specific limitations on the static map generating apparatus, please refer to the limitations on the static map generating method above, which will not be repeated here. Each module in the above-mentioned static map generating apparatus may be implemented in whole or in part by software, hardware and combinations thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

In one embodiment, a computer device is provided, and the computer device may be a terminal, and its internal structure diagram may be as shown in FIG. 8 . The computer equipment includes a processor, memory, a network interface, a display screen, and an input device connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium, an internal memory. The non-volatile storage medium stores an operating system and computer-readable instructions. The internal memory provides an environment for the execution of the operating system and computer-readable instructions in the readable storage medium. The network interface of the computer device is used to communicate with an external server over a network connection. The computer readable instructions, when executed by a processor, implement a static map generation method. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

In one embodiment, a computer device is provided, comprising a memory, a processor, and computer-readable instructions stored on the memory and executable on the processor, and the processor implements the following steps when executing the computer-readable instructions:

Obtain the depth image data and pose data collected in the same time period;

In one embodiment, one or more computer-readable storage media storing computer-readable instructions are provided, and the readable storage media provided in this embodiment include non-volatile readable storage media and volatile readable storage media storage medium. Computer-readable instructions are stored on the readable storage medium, and when the computer-readable instructions are executed by one or more processors, implement the following steps:

Obtain the depth image data and pose data collected in the same time period;

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a non-volatile computer. In the read storage medium or the volatile readable storage medium, the computer-readable instructions, when executed, may include the processes of the foregoing method embodiments. Wherein, any reference to memory, storage, database or other medium used in the various embodiments provided in this application may include non-volatile and/or volatile memory. Nonvolatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Road (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

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 practical applications, the above-mentioned functions can be allocated to different functional units, Module completion, that is, dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above.

The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity, all possible combinations of the technical features in the above-described embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, All should be regarded as the scope described in this specification.

The above-mentioned embodiments only represent several embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but should not be construed as a limitation on the scope of the patent application. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the application, some modifications and improvements can also be made, which all belong to the protection scope of the application. Therefore, the scope of protection of the patent of the present application shall be subject to the appended claims.

Claims

A method for generating a static map, comprising:

Obtain the depth image data and pose data collected in the same time period;

Determine the pose information synchronized with the depth image data according to the pose data;

processing the depth image data based on the pose information and the installation position corresponding to the depth image data to generate point cloud data in the robot coordinate system;

intercepting the point cloud data according to a preset interception rule to generate local point cloud data;

Update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system;

The initial static map is updated according to the global obstacle map data, and the updated initial static map is a static map.
The method for generating a static map according to claim 1, wherein the updating of the global obstacle map data according to the local point cloud data, wherein the coordinate system of the global obstacle map data is a world coordinate system, comprising:

Update local obstacle map data according to the local point cloud data, and the coordinate system of the local obstacle map data is the robot coordinate system;

Convert the updated local obstacle map data into the global obstacle map data.
The method for generating a static map according to claim 1, wherein before acquiring the depth image data and pose data collected synchronously, the method further comprises:

Acquire multiple initial pose data collected at preset condition intervals during the mapping process;

The plurality of initial pose data are corrected by using a preset loop closure detection algorithm to obtain the pose data.
The method for generating a static map according to claim 3, wherein the preset condition interval is a time interval, a frame number interval or an angle interval.
The method for generating a static map according to claim 1, wherein the determining the pose information synchronized with the depth image data according to the pose data comprises:

obtaining the acquisition time of the depth image data;

judging whether the pose data has pose information at the collection time;

If the pose data does not have the pose information at the collection time, the pose data is processed through an interpolation algorithm to generate the pose information at the collection time, and the pose information at the collection time is generated. And the pose data is the pose information synchronized with the depth image data.
The method for generating a static map according to claim 5, wherein the processing of the pose data through an interpolation algorithm to generate the pose information at the collection time comprises:

The pose information at the collection time is generated using the pose data whose temporal relationship is most adjacent to the collection time.
The method for generating a static map according to claim 6, wherein the generating the pose information at the collection time by using the pose data whose time relationship is most adjacent to the collection time comprises:

Obtain the nearest corrected pose data before the specified time stamp and the nearest corrected pose data after the specified time stamp, and infer the pose information of the specified time stamp according to the two corrected pose data.
The method for generating a static map according to claim 1, wherein the intercepting the point cloud data according to a preset interception rule to generate local point cloud data, comprising:

Determine the facing position of the robot according to the pose information;

Determine the target area according to the facing position, and extract the local point cloud data from the point cloud data according to the target area. The target area includes within a specified distance in front of the robot, higher than the ground and lower than the robot Perceptual area of visual height.
The method for generating a static map according to claim 2, wherein the local obstacle map data is updated according to the local point cloud data, and the coordinate system of the local obstacle map data is a robot coordinate system, include:

Projecting the local point cloud data to the local obstacle map on the ground plane, and setting the pixel value of the grid with the point projection to the specified pixel value;

After all points in the local point cloud data are projected, the update of the local obstacle map data is completed.
The method for generating a static map according to claim 9, wherein the projecting the local point cloud data to the local obstacle map on the ground plane, and setting the pixel value of the grid where the point projection exists. The process for specifying pixel values is:

During projection, the corresponding pixel value of the grid projected by the point cloud on the local obstacle map is set to 255; the corresponding pixel value of the grid not projected by the point cloud remains the original value of 0.
The method for generating a static map according to claim 1, wherein the initial static map is updated according to the global obstacle map data, and the updated initial static map is a static map, comprising:

If the pixel value of the first specified grid in the global obstacle map data is greater than 0, increase the pixel value of the first static grid by a first preset amplitude, and increase the pixel value of the first preset amplitude. The pixel value of the first static grid is not greater than 255, and the first static grid is the grid corresponding to the first specified grid in the initial static map;

If the pixel value of the second specified grid in the global obstacle map data is 0, reduce the pixel value of the second static grid by a second preset amplitude, and reduce the pixel value of the second preset amplitude. The pixel value of the second static grid is not less than 0, and the second static grid is a grid corresponding to the second specified grid in the initial static map.
The method for generating a static map according to claim 1, wherein the updating of the initial static map according to the global obstacle map data, and the updated initial static map is the static map, further comprising:

Get code map data;

A topological path of the robot is generated according to the code map data and the static map.
The method for generating a static map according to claim 12, wherein the code map data is visual sign recognition data collected during the process of building a map by the robot.
The method for generating a static map according to claim 12, wherein the static map includes obstacle information.
The method for generating a static map according to claim 1, wherein the depth image data comprises several frames of depth images, and each frame of the depth images is marked with a timestamp corresponding to the acquisition time.
The method for generating a static map according to claim 1, wherein the position data includes several pieces of pose information, the pose information includes plane coordinates and angles, and each of the pose information has a corresponding collection time.
A static map generating device, comprising:

The data acquisition module is used to acquire the depth image data and pose data collected in the same time period;

a synchronous pose information determining module, configured to determine pose information synchronized with the depth image data according to the pose data;

a point cloud generation module, configured to process the depth image data based on the pose information and the installation position corresponding to the depth image data, to generate point cloud data in the robot coordinate system;

a point cloud interception module, configured to intercept the point cloud data according to a preset interception rule to generate local point cloud data;

an obstacle map update module, configured to update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system;

The static map update module is used to update the initial static map according to the global obstacle map data, and the updated initial static map is the static map.
A computer device, comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein the processor implements the following when executing the computer-readable instructions Static map generation method:

Obtain the depth image data and pose data collected in the same time period;

Determine the pose information synchronized with the depth image data according to the pose data;

processing the depth image data based on the pose information and the installation position corresponding to the depth image data to generate point cloud data in the robot coordinate system;

intercepting the point cloud data according to a preset interception rule to generate local point cloud data;

Update global obstacle map data according to the local point cloud data, where the coordinate system of the global obstacle map data is the world coordinate system;

The initial static map is updated according to the global obstacle map data, and the updated initial static map is a static map.
One or more readable storage media storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform any of the methods of claims 1 to 16. One of the static map generation methods.