CN113639745B - Point cloud map construction method, device and storage medium - Google Patents

Abstract

The application discloses a method and a device for constructing a point cloud map and a storage medium, which are used for improving the accuracy of constructing the point cloud map. The method for constructing the point cloud map comprises the following steps: collecting first information of a first position; after the first information is acquired, moving to a second position, and acquiring second information of the second position; locally matching the second information with the first information to obtain a local point cloud map; deleting the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map; the first information comprises a first environment point cloud image and a first environment image; the second information includes a second ambient point cloud image and a second ambient image. The application also provides a device for constructing the point cloud map and a storage medium.

Description

A method, device and storage medium for constructing a point cloud map

Technical field

The present application relates to the field of unmanned vehicles, and in particular, to a method, device and storage medium for constructing a point cloud map.

Background technique

Currently, mapping technology is receiving more and more attention as an important technology in the field of unmanned vehicles. In existing mapping technology, removal of dynamic targets in point cloud maps is not considered, resulting in inaccurate mapping results.

Contents of the invention

In response to the above technical problems, embodiments of the present application provide a method, device and storage medium for constructing a point cloud map, which are used to perceive dynamic targets in the environment and delete the point clouds of dynamic targets in the point cloud map, thereby obtaining a more accurate Environmental point cloud image.

In the first aspect, an embodiment of the present application provides a method for constructing a point cloud map, including:

Collect the first information from the first position;

After completing the collection of the first information, move to the second position and collect the second information at the second position;

Locally match the second information with the first information to obtain a local point cloud map;

Delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map;

Wherein, the first information includes a first environment point cloud map and a first environment image;

The second information includes a second environment point cloud map and a second environment image.

Further, when the point cloud map construction starts, the first position is the starting position;

When the next position is collected after the first position collection is completed, the second position is the current collection position, and the first position is the last collected position.

Preferably, the collection device is turned on, and the collection device includes one or a combination of the following:

lidar;

binocular camera;

depth camera;

Inertial Measurement Unit IMU;

Wheel speed odometer.

Preferably, after completing the collection of the entire mapping area, the collection is ended, and the laser point cloud maps of all collection points are spliced to obtain an environmental point cloud map of the entire mapping area.

Preferably, collecting the first information of the first location includes:

Establish a first environmental point cloud map based on the lidar data collected by the lidar;

A first environment image is established based on the image data collected by the binocular camera and the depth camera.

The collecting the second information of the second location includes:

Establish a second environmental point cloud map based on the lidar data collected by the lidar;

A second environment image is established based on the image data collected by the binocular camera and the depth camera.

Furthermore, it also includes:

Upload the lidar data collected by the lidar to the host computer system;

Upload the image data collected by the binocular camera and the depth camera to the host computer system.

Preferably, deleting the dynamic target point cloud in the local point cloud map includes:

Match the image at the first position or the second position with the point cloud;

Delete the point cloud corresponding to the dynamic target identified in the image from the point cloud map.

Preferably, the point cloud characterizing the dynamic target includes:

Point clouds representing pedestrians or non-motor vehicles or motor vehicles.

Preferably, matching the image at the first position or the second position with the point cloud includes:

Match image information with point cloud information according to the transformation matrix.

Using the point cloud map construction method provided by the present invention, local matching is performed based on the environmental image information and environmental point cloud map information of the two previous and subsequent collection points, and then the dynamic target is identified, and then the point cloud of the dynamic target is extracted from the local point cloud map. Delete, thus improving the accuracy of mapping. Maps with dynamic targets will cause serious interference to subsequent use of maps for positioning and navigation, while maps established using the method of the present invention do not contain dynamic targets and are more accurate.

In a second aspect, embodiments of the present application also provide a device for constructing a point cloud map, including:

The collection module is configured to collect the first information at the first position, and after completing the collection of the first information, moves to the second position to collect the second information at the second position;

a local matching module configured to locally match the second information with the first information to obtain a local point cloud map;

A dynamic target deletion module configured to delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map;

Wherein, the first information includes a first environment point cloud map and a first environment image;

The second information includes a second environment point cloud map and a second environment image.

In a third aspect, embodiments of the present application also provide a device for constructing a point cloud map, including: a memory, a processor, and a user interface;

The memory is used to store computer programs;

The user interface is used to interact with users;

The processor is configured to read the computer program in the memory. When the processor executes the computer program, it implements the point cloud map construction method provided by the present invention.

In a fourth aspect, embodiments of the present application also provide a processor-readable storage medium, the processor-readable storage medium stores a computer program, and when the processor executes the computer program, the point cloud map provided by the present invention is implemented. construction method.

Description of the drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, a brief introduction will be given below to the drawings required to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. Those of ordinary skill in the art can also obtain other drawings based on these drawings without exerting creative efforts.

Figure 1 is a schematic diagram of the point cloud map construction provided in this embodiment;

Figure 2 is a schematic diagram of environmental image and environmental point cloud image collection provided by the embodiment of the present application;

Figure 3 is a schematic diagram of first location and second location information collection provided by the embodiment of the present application;

Figure 4 is a schematic diagram of dynamic target recognition provided by an embodiment of the present application;

Figure 5 is a schematic diagram after deletion of the dynamic target point cloud provided by the embodiment of the present application;

Figure 6 is a schematic structural diagram of a point cloud map construction device provided by an embodiment of the present application;

Figure 7 is a schematic structural diagram of another point cloud map construction device provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some of the embodiments of the present invention, not all of them. . Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

Here are some explanations of some words that appear in the text:

1. In the embodiment of the present invention, the term "and/or" describes the association relationship of associated objects, indicating that there can be three relationships, for example, A and/or B, which can mean: A exists alone, A and B exist simultaneously, alone There are three situations B. The character "/" generally indicates that the related objects are in an "or" relationship.

2. In the embodiments of this application, the term "multiple" refers to two or more, and other quantifiers are similar to it.

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

It should be noted that the display order of the embodiments of this application only represents the order of the embodiments, and does not represent the quality of the technical solutions provided by the embodiments.

Embodiment 1

Referring to Figure 1, a schematic diagram of a point cloud map construction method provided by an embodiment of the present application is shown. As shown in Figure 1, the method includes steps S101 to S104:

S101. Collect the first information of the first position;

S102. After completing the collection of the first information, move to the second position and collect the second information at the second position;

S103. Locally match the second information with the first information to obtain a local point cloud map;

S104. Delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map;

Wherein, the first information includes a first environment point cloud map and a first environment image;

The second information includes a second environment point cloud map and a second environment image.

Specifically, the first environment point cloud image and the second environment point cloud image are established through lidar data collected by lidar, and the first environment image and the second environment image are images collected through binocular cameras and depth cameras. Data is created.

It should be noted that this method is applied to unmanned vehicles that construct point cloud maps. The unmanned vehicle carries at least one of a lidar, a binocular camera, a depth camera, an inertial measurement unit (IMU), and a wheel speed odometer. Lidar is used to collect lidar data, binocular cameras and depth cameras are used to collect image data, IMU is used to collect inertial measurement data, and wheel speed odometer is used to collect mileage data, as shown in Figure 2.

When the point cloud map begins to be constructed, the unmanned vehicle is at the starting position, which is the first position. After the collection of the first information at the first position is completed, it moves to the second position and collects the second information at the second position. That is, when the point cloud map construction starts, the first position is the starting position; when the next position is collected after the first position collection is completed, the second position is the current collection position, and the first position is the starting position. The location is the last collected location.

As a preferred example, before starting the construction, turn on the lidar, binocular camera, depth camera, inertial measurement unit (IMU) and wheel speed odometer carried on the unmanned vehicle.

It should be noted that the second position in this embodiment is the next collection position after the first position is collected. The second position moves within the entire mapping area. After completing the collection of the entire mapping area, the collection ends and all The laser point cloud maps of the collection points are stitched together to obtain an environmental point cloud map of the entire mapping area.

As a preferred example, the second position moves within the entire mapping area, and can be moved manually, or the unmanned vehicle can move automatically according to preset rules, which is not limited in this embodiment.

As a preferred example, in this embodiment S101, collecting the first information of the first location includes:

Establish a first environmental point cloud map based on the lidar data collected by the lidar;

A first environment image is established based on the image data collected by the binocular camera and the depth camera.

As a preferred example, in S102 of this embodiment, collecting the second information of the second location includes:

Establish a second environmental point cloud map based on the lidar data collected by the lidar;

A second environment image is established based on the image data collected by the binocular camera and the depth camera.

As a preferred example, after the unmanned vehicle collects the first information or the second information, the lidar data collected by the lidar is uploaded to the host computer system; the images collected by the binocular camera and the depth camera are The data is uploaded to the host computer system, and data collected by other devices can also be uploaded to the host computer system.

As a preferred example, in S103 of this embodiment, locally matching the second information with the first information to obtain a local point cloud map includes:

The above-mentioned local matching process is to register the point clouds collected by the lidar at the two positions, and transform the point clouds at different positions to the same position through the information of the overlapping parts. As a preferred example, the ICP algorithm (Iterative Closest Point, nearest neighbor iterative algorithm) is used to unify the two sets of point clouds into a unified coordinate system and perform point cloud registration. The point cloud map obtained after registration becomes the local environment map.

As a preferred example, in S104 of this embodiment, deleting the dynamic target point cloud in the local point cloud map and obtaining the laser point cloud map includes:

Match the image at the first position or the second position with the point cloud;

Delete the point cloud corresponding to the dynamic target identified in the image from the point cloud map. As shown in Figure 4, two dynamic targets are identified, as shown in Figure 5, the point cloud map after deleting the dynamic targets identified in Figure 4.

As a preferred example, matching the image at the first position or the second position with the point cloud includes:

Match image information with point cloud information according to the transformation matrix. That is to say, the first environment image is matched with the first environment point cloud image according to the transformation matrix, or the second environment image is matched with the second environment point cloud image according to the transformation matrix.

Preferably, the transformation matrix is determined by the following steps:

A1: Calibration.

Multi-sensor calibration is performed before the point cloud map is constructed, that is, the camera image information and radar point cloud information are calibrated, that is, in an environment with clear feature points, the camera is turned on to collect image information, the radar is turned on to collect point cloud information, and then the staff selects Feature locations on the image and draw corresponding locations on the point cloud information. As a preferred example, the extraction of each feature point on the image can be obtained by SIFT image feature point extraction algorithm or the like.

A2: Determine the transformation matrix.

The conversion relationship between the drawn point cloud position and the selected image position is the conversion matrix. The conversion matrix represents the direct mapping relationship between the three-dimensional point cloud and the visual image pixels. That is, the point cloud can be found at any position on the image through the conversion matrix. Point cloud locations on the information.

Preferably, the point cloud characterizing the dynamic target includes:

Point clouds representing pedestrians or non-motor vehicles or motor vehicles. That is, dynamic targets include one or a combination of the following: pedestrians, motor vehicles, and non-motor vehicles. You can also include other objects that move within the mapping area, such as crawlers on the ground, flying objects in the sky, and other moving objects. As shown in Figure 3, the dynamic target is a vehicle.

A specific example of this embodiment is given below:

S1. The unmanned vehicle is at the starting position, turns on the lidar, binocular camera, depth camera, inertial measurement unit (IMU) and wheel speed odometer carried on the unmanned vehicle, and starts collecting lidar data, image data, and inertial measurement Unit (IMU) data and wheel speed odometer data;

S2. The unmanned vehicle uploads the lidar data collected by the lidar to the host computer system to establish an environmental point cloud image. The unmanned vehicle uploads the image data collected by the binocular camera and depth camera to the host computer system to establish an environmental image;

S3. After the environmental point cloud map and environmental image of the unmanned vehicle at the initial position are established, the staff will observe the surrounding environment through the camera carried by the unmanned vehicle and remotely control the unmanned vehicle to move to the next position;

S4. The lidar, binocular camera, depth camera, inertial measurement unit (IMU) and wheel speed odometer carried on the unmanned vehicle continue to collect lidar data, image data, inertial measurement unit (IMU) data and wheel speed odometer. data and upload it to the host computer;

S5. The host computer creates an environmental point cloud image of the new location based on the newly collected lidar data, and creates an environmental image of the new location based on the newly collected image data;

S6. Locally match the environment point cloud map and environment image at the new location with the environment point cloud map and environment image at the previous location to form a local point cloud map;

S7. Use the image information to delete the point clouds represented by dynamic targets such as pedestrians, non-motorized vehicles, and motor vehicles on the locally matched local point cloud map, and use the point cloud information of the new position and the previous position to compare the dynamic objects. Point cloud is deleted;

S8. Save the laser point cloud map after deleting the dynamic point cloud;

Through the method of this embodiment, local matching is performed based on the environmental image information and environmental point cloud map information of the two previous collection points, and then the dynamic target is identified, and then the point cloud of the dynamic target is deleted from the local point cloud map, thereby improving Accuracy of mapping. Maps with dynamic targets will cause serious interference to subsequent use of maps for positioning and navigation, while maps established using the method of the present invention do not contain dynamic targets and are more accurate.

Embodiment 2

Based on the same inventive concept, embodiments of the present invention also provide a device for constructing a point cloud map. As shown in Figure 6, the device includes:

The collection module 601 is configured to collect the first information at the first position. After completing the first information collection, move to the second position and collect the second information at the second position;

The local matching module 602 is configured to locally match the second information with the first information to obtain a local point cloud map;

The dynamic target deletion module 603 is configured to delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map;

Wherein, the first information includes a first environment point cloud map and a first environment image;

The second information includes a second environment point cloud map and a second environment image.

In this embodiment, when the point cloud map construction starts, the first position is the starting position;

When the next position is collected after the first position collection is completed, the second position is the current collection position, and the first position is the last collected position.

As a preferred example, the collection module 601 is also used to turn on the collection device before starting the point cloud map construction. The collection device includes one or a combination of the following:

lidar;

binocular camera;

depth camera;

Inertial Measurement Unit IMU;

Wheel speed odometer.

As a preferred example, the acquisition module 601 is also configured to establish a first environment point cloud map based on the lidar data collected by the lidar; and establish the first environment based on the image data collected by the binocular camera and the depth camera. image. A second environment point cloud image is established based on the lidar data collected by the lidar; a second environment image is established based on the image data collected by the binocular camera and the depth camera.

As a preferred example, the acquisition module 601 is also used to upload the lidar data collected by the lidar to the host computer system; and upload the image data collected by the binocular camera and the depth camera to the host computer system.

As a preferred example, the local matching module 602 is also configured to locally match the second information with the first information according to the following method:

The point clouds collected by the lidar at the two positions before and after (i.e. the first position and the second position) are registered, and the point clouds at different positions are transformed to the same position through the information of the overlapping parts. As a preferred example, the ICP algorithm (Iterative Closest Point, nearest neighbor iterative algorithm) is used to unify two groups or point clouds into a unified coordinate system, and perform point cloud registration. The point cloud map obtained after registration becomes the environment Local map.

As a preferred example, the dynamic target deletion module 603 is also used to:

Match the image at the first position or the second position with the point cloud;

Delete the point cloud corresponding to the dynamic target identified in the image from the point cloud map.

Among them, matching the image at the first position or the second position with the point cloud includes:

Match image information with point cloud information according to the transformation matrix. That is to say, the first environment image is matched with the first environment point cloud image according to the transformation matrix, or the second environment image is matched with the second environment point cloud image according to the transformation matrix.

Preferably, the dynamic target deletion module 603 is also used to determine the transformation matrix according to the following steps:

A1: Calibration.

Multi-sensor calibration is performed before the point cloud map is constructed, that is, the camera image information and radar point cloud information are calibrated, that is, in an environment with clear feature points, the camera is turned on to collect image information, the radar is turned on to collect point cloud information, and then the staff selects Feature locations on the image and draw corresponding locations on the point cloud information. As a preferred example, the extraction of each feature point on the image can be obtained by SIFT image feature point extraction algorithm or the like.

A2: Determine the transformation matrix.

The conversion relationship between the drawn point cloud position and the selected image position is the conversion matrix. The conversion matrix represents the direct mapping relationship between the three-dimensional point cloud and the visual image pixels. That is, the point cloud can be found at any position on the image through the conversion matrix. Point cloud locations on the information.

Among them, point clouds that represent dynamic targets include: point clouds that represent pedestrians, non-motor vehicles, or motor vehicles.

As a preferred example, the dynamic target deletion module 603 is also used to:

After completing the collection of the entire mapping area, end the collection, and stitch the laser point cloud maps of all collection points to obtain an environmental point cloud map of the entire mapping area.

It should be noted that the collection module 601 provided in this embodiment can realize all the functions included in steps S101 and S102 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be described again here;

It should be noted that the local matching module 602 provided in this embodiment can realize all the functions included in step S103 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be described again here;

It should be noted that the dynamic target deletion module 603 provided in this embodiment can realize all the functions included in step S104 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be described again here;

It should be noted that the device provided in Embodiment 2 and the method provided in Embodiment 1 belong to the same inventive concept, solve the same technical problem, and achieve the same technical effect. The device provided in Embodiment 2 can implement all the methods in Embodiment 1. , the similarities will not be repeated.

Embodiment 3

Based on the same inventive concept, embodiments of the present invention also provide a device for constructing a point cloud map. As shown in Figure 7, the device includes:

Includes memory 702, processor 701 and user interface 703;

The memory 702 is used to store computer programs;

The user interface 703 is used to interact with users;

The processor 701 is used to read the computer program in the memory 702. When the processor 701 executes the computer program, it implements:

Collect the first information from the first position;

After completing the collection of the first information, move to the second position and collect the second information at the second position;

Locally match the second information with the first information to obtain a local point cloud map;

Delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map;

Wherein, the first information includes a first environment point cloud map and a first environment image;

The second information includes a second environment point cloud map and a second environment image.

Preferably, when the point cloud map construction starts, the first position is the starting position;

When the next position is collected after the first position collection is completed, the second position is the current collection position, and the first position is the last collected position.

As a preferred example, when the processor 701 executes the computer program, it implements:

Turn on the collection device before starting the construction of the point cloud map. The collection device includes one or a combination of the following:

lidar;

binocular camera;

depth camera;

Inertial Measurement Unit IMU;

Wheel speed odometer.

As a preferred example, when the processor 701 executes the computer program, it implements:

After completing the collection of the entire mapping area, end the collection, and stitch the laser point cloud maps of all collection points to obtain an environmental point cloud map of the entire mapping area.

As a preferred example, when the processor 701 executes the computer program, it implements:

According to the lidar data collected by the lidar, a first environmental point cloud image is established; according to the image data collected by the binocular camera and the depth camera, a first environment image is established. A second environment point cloud image is established based on the lidar data collected by the lidar; a second environment image is established based on the image data collected by the binocular camera and the depth camera.

Further, the lidar data collected by the lidar is uploaded to the host computer system; the image data collected by the binocular camera and the depth camera are uploaded to the host computer system.

As a preferred example, when the processor 701 executes the computer program, it implements:

According to the second environment image and the first environment image, a point cloud representing a dynamic target on the local point cloud map is identified; according to the second environment point cloud map and the first environment point cloud map, delete Point clouds representing dynamic targets in the local point cloud map.

Among them, point clouds that represent dynamic targets include: point clouds that represent pedestrians, non-motor vehicles, or motor vehicles.

As a preferred example, when the processor 701 executes the computer program, it implements:

The second information is partially matched with the first information according to the following method:

The point clouds collected by the lidar at the two positions before and after (i.e. the first position and the second position) are registered, and the point clouds at different positions are transformed to the same position through the information of the overlapping parts. As a preferred example, the ICP algorithm (Iterative Closest Point, nearest neighbor iterative algorithm) is used to unify two groups or point clouds into a unified coordinate system, and perform point cloud registration. The point cloud map obtained after registration becomes the environment Local map.

As a preferred example, when the processor 701 executes the computer program, it implements: matching the image at the first position or the second position with the point cloud;

Delete the point cloud corresponding to the dynamic target identified in the image from the point cloud map.

Among them, matching the image at the first position or the second position with the point cloud includes:

Match image information with point cloud information according to the transformation matrix. That is to say, the first environment image is matched with the first environment point cloud image according to the transformation matrix, or the second environment image is matched with the second environment point cloud image according to the transformation matrix.

When the processor 701 executes the computer program, it implements: determining the transformation matrix according to the following steps:

A1: Calibration.

Multi-sensor calibration is performed before the point cloud map is constructed, that is, the camera image information and radar point cloud information are calibrated, that is, in an environment with clear feature points, the camera is turned on to collect image information, the radar is turned on to collect point cloud information, and then the staff selects Feature locations on the image and draw corresponding locations on the point cloud information. As a preferred example, the extraction of each feature point on the image can be obtained by SIFT image feature point extraction algorithm or the like.

A2: Determine the transformation matrix.

The conversion relationship between the drawn point cloud position and the selected image position is the conversion matrix. The conversion matrix represents the direct mapping relationship between the three-dimensional point cloud and the visual image pixels. That is, the point cloud can be found at any position on the image through the conversion matrix. Point cloud locations on the information.

In FIG. 7 , the bus architecture may include any number of interconnected buses and bridges, specifically one or more processors represented by processor 701 and various circuits of the memory represented by memory 702 are linked together. The bus architecture can also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are all well known in the art and therefore will not be described further herein. The bus interface provides the interface. The processor 701 is responsible for managing the bus architecture and general processing, and the memory 702 can store data used by the processor 701 when performing operations.

The processor 701 can be a CPU, ASIC, FPGA or CPLD, and the processor 701 can also adopt a multi-core architecture.

When the processor 701 executes the computer program stored in the memory 702, the method for constructing any point cloud map in Embodiment 1 is implemented.

It should be noted that the device provided in Embodiment 3 and the method provided in Embodiment 1 belong to the same inventive concept, solve the same technical problem, and achieve the same technical effect. The device provided in Embodiment 3 can implement all the methods in Embodiment 1. , the similarities will not be repeated.

This application also proposes a processor-readable storage medium. Wherein, the processor-readable storage medium stores a computer program, and when the processor executes the computer program, the method for constructing any point cloud map in Embodiment 1 is implemented.

It should be noted that the division of units in the embodiment of the present application is schematic and is only a logical function division. In actual implementation, there may be other division methods. In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or software functional units.

Those skilled in the art will understand that embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment that combines software and hardware aspects. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, magnetic disk storage and optical storage, etc.) embodying computer-usable program code therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each process and/or block in the flowchart illustrations and/or block diagrams, and combinations of processes and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a use A device for realizing the functions specified in one process or multiple processes of the flowchart and/or one block or multiple blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory that causes a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction means, the instructions The device implements the functions specified in a process or processes of the flowchart and/or a block or blocks of the block diagram.

Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. In this way, if these modifications and variations of the present application fall within the scope of the claims of the present application and equivalent technologies, the present application is also intended to include these modifications and variations.

Claims (12)

1. A method for constructing a point cloud map, which is characterized by including: Collect the first information from the first position; After completing the collection of the first information, move to the second position and collect the second information at the second position; Locally match the second information with the first information to obtain a local point cloud map; Delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map; Wherein, the first information includes a first environment point cloud map and a first environment image; The second information includes a second environment point cloud map and a second environment image; Deleting the dynamic target point cloud in the local point cloud map includes: Match the image at the first position or the second position with the point cloud; Delete the point cloud corresponding to the dynamic target identified in the image from the point cloud map. 2. The method according to claim 1, characterized in that, comprising: When the point cloud map construction starts, the first position is the starting position; When the next position is collected after the first position collection is completed, the second position is the current collection position, and the first position is the last collected position. 3. The method according to claim 2, characterized in that, before starting the point cloud map construction, it also includes: Turn on the collection device, which includes one or a combination of the following: lidar; binocular camera; depth camera; Inertial Measurement Unit IMU; Wheel speed odometer. 4. The method of claim 2, further comprising: After completing the collection of the entire mapping area, end the collection, and stitch the laser point cloud maps of all collection points to obtain an environmental point cloud map of the entire mapping area. 5. The method of claim 3, wherein collecting the first information of the first location includes: Establish a first environmental point cloud map based on the lidar data collected by the lidar; A first environment image is established based on the image data collected by the binocular camera and the depth camera. 6. The method of claim 3, wherein collecting the second information of the second location includes: Establish a second environmental point cloud map based on the lidar data collected by the lidar; A second environment image is established based on the image data collected by the binocular camera and the depth camera. 7. The method according to claim 5 or 6, characterized in that, further comprising: Upload the lidar data collected by the lidar to the host computer system; Upload the image data collected by the binocular camera and the depth camera to the host computer system. 8. The method according to claim 1, characterized in that the point cloud characterizing the dynamic target includes: Point clouds representing pedestrians or non-motor vehicles or motor vehicles. 9. The method of claim 1, wherein matching the image at the first position or the second position with the point cloud includes: Match image information with point cloud information according to the transformation matrix. 10. A point cloud map construction device, characterized by including: The collection module is configured to collect the first information at the first position, and after completing the collection of the first information, moves to the second position to collect the second information at the second position; a local matching module configured to locally match the second information with the first information to obtain a local point cloud map; A dynamic target deletion module configured to delete the dynamic target point cloud in the local point cloud map to obtain a laser point cloud map; Wherein, the first information includes a first environment point cloud map and a first environment image; The second information includes a second environment point cloud map and a second environment image; Deleting the dynamic target point cloud in the local point cloud map includes: Match the image at the first position or the second position with the point cloud; Delete the point cloud corresponding to the dynamic target identified in the image from the point cloud map. 11. A point cloud map construction device, characterized by including a memory, a processor and a user interface; The memory is used to store computer programs; The user interface is used to interact with users; The processor is configured to read the computer program in the memory. When the processor executes the computer program, it implements the point cloud map construction method according to one of claims 1 to 9. 12. A processor-readable storage medium, characterized in that the processor-readable storage medium stores a computer program, and when the processor executes the computer program, the method of one of claims 1 to 9 is implemented. Method of constructing point cloud map.