A method, apparatus and device for coloring a scene point cloud map

Abstract

The application provides a coloring method, device and equipment for a scene point cloud map, which comprises the following steps: constructing a gridded scene point cloud map, and obtaining historical motion state parameters of each grid in the gridded scene point cloud map; obtaining current frame point cloud data and current motion state parameters corresponding to a current acquisition time; for each grid, judging whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters; if yes, coloring the point cloud located in the grid based on current frame image data acquired at the current acquisition time, so as to update the point cloud in the grid. The application ensures that each grid is ultimately colored based on the frame data with the most stable and highest data quality when being acquired, and improves the coloring accuracy and coloring efficiency of the point cloud.

Description

Technical Field

[0001] This invention relates to the field of scene mapping technology, specifically to a method, apparatus, and device for coloring scene point cloud maps. Background Technology

[0002] In safety inspections and digital modeling of power transmission, substation, and distribution operations, generating color 3D maps by fusing laser point clouds and camera images is a key technology. This technology can intuitively display equipment shape, spatial relationships, and texture information, providing a core data foundation for distance measurement, status recognition, and safety monitoring.

[0003] Currently, mainstream color point cloud map construction methods, whether based on hardware or software synchronization, focus on solving the instantaneous synchronization problem between laser point clouds, camera images, and device pose at the moment of a single data acquisition. For example, this can be achieved by adding a hardware trigger module to force sensor synchronization or by using software to align timestamps for data association. However, these methods have inherent limitations: During device scanning, its motion state is dynamically changing. Data acquired when the device is moving steadily yields more accurate pose estimation and clearer point clouds; however, data acquired during violent device movement suffers a significant quality degradation due to motion blur and increased pose estimation errors. Existing methods treat all these unevenly qualityed data equally and fuse them all into the final map, leading to color misalignment or distortion in the resulting scene map. Furthermore, they require substantial computational resources to process low-quality, high-noise data frames that should have been discarded, resulting in significant computational redundancy.

[0004] Therefore, there is an urgent need to provide a coloring method, apparatus, and device for scene point cloud maps that can differentiate the use of data generated during device scanning and movement to improve coloring accuracy, reduce computational redundancy, and improve coloring efficiency. Summary of the Invention

[0005] In view of this, it is necessary to provide a method, apparatus and device for coloring scene point cloud maps, so as to solve the technical problems existing in the prior art. In the process of coloring scene point cloud maps, the data is not distinguished, resulting in the use of low-quality data for point cloud coloring, which leads to low coloring accuracy and wastes resources on low-quality data processing, resulting in low computational efficiency.

[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a method for coloring a scene point cloud map, comprising: Construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map; Acquire the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time; For each grid cell, determine whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. If so, the point cloud located within the grid is colored based on the current frame image data acquired at the current acquisition time, so as to update the point cloud within the grid.

[0007] In one possible implementation, the historical motion state parameters include historical angular velocity values, and the current motion state parameters include current angular velocity values; then, determining whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters includes: When the current angular velocity value is less than the historical angular velocity value, it is determined whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters.

[0008] In one possible implementation, the method further includes: The current motion state parameters are used as the historical motion state parameters of the grid.

[0009] In one possible implementation, coloring the point cloud located within the grid based on the image data acquired at the current acquisition time includes: Obtain the previous basic pose output by the SLAM algorithm at the previous acquisition time and the current basic pose output at the current acquisition time; Based on the current motion state parameters, pose interpolation is performed between the previous base pose and the current base pose to generate an interpolated pose sequence containing at least one intermediate pose. The target interpolation pose with the highest matching degree to the current frame image data in the interpolation pose sequence is determined, and the point cloud in the grid is projected onto the current frame image data based on the target interpolation pose for coloring.

[0010] In one possible implementation, the pose interpolation based on the current motion state parameters between the previous base pose and the current base pose includes: The number of interpolations for the intermediate pose is determined based on the current motion state parameters; Based on the number of interpolations, pose interpolation is performed between the previous basic pose and the current basic pose using linear interpolation to obtain the interpolated pose sequence.

[0011] In one possible implementation, the number of interpolations is: N=k w(x,y,z) In the formula, N is the number of interpolations; k is the velocity coefficient; w(x,y,z) are the current motion state parameters; It is the modulo operator; This is the floor operator.

[0012] In one possible implementation, determining the target interpolation pose in the interpolation pose sequence that has the highest matching degree with the current frame image data includes: Extract the image geometric features from the current frame image data and the point cloud geometric features from the current frame point cloud data; Based on each of the intermediate poses, the geometric features of the point cloud are projected and transformed to obtain geometric projection features; Determine the similarity between the image geometric features and the geometric projection features, and use the intermediate pose corresponding to the highest similarity value as the target interpolation pose.

[0013] In one possible implementation, both the image geometric features and the point cloud geometric features are structured straight line features.

[0014] Secondly, the present invention also provides a coloring device for a scene point cloud map, comprising: A rasterized map construction unit is used to construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map. The current motion state parameter acquisition unit is used to acquire the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time. A data quality comparison unit is used to determine, for each of the grid cells, whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. The map coloring unit is used to color the point cloud located in the grid based on the current frame image data acquired at the current acquisition time when the data quality represented by the current motion state parameter is better than the data quality represented by the historical motion state parameter, so as to update the point cloud in the grid.

[0015] Thirdly, the present invention also provides a coloring device for a scene point cloud map, comprising: a data acquisition sub-device and a data processing sub-device, wherein the data acquisition sub-device includes a camera and a lidar, and the data processing sub-device includes a memory and a processor; The camera is used to acquire image data of the current frame; The lidar is used to collect point cloud data in the current frame; The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the scene point cloud map coloring method described in any of the above possible implementations.

[0016] The beneficial effects of this invention are as follows: The scene point cloud map coloring method provided by this invention, for each grid cell, determines whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. If so, the point cloud within the grid cell is colored based on the current frame image data acquired at the current acquisition time. Compared to the prior art, which indiscriminately uses all acquired data to color the point cloud, potentially causing early high-quality data to be covered or contaminated by later low-quality data, this invention utilizes the principle of selective updating to ensure that each grid cell is ultimately colored based on the most stable and highest-quality frame data acquired at the time of acquisition. This improves the accuracy of point cloud coloring and meets the needs of refined scene modeling and ranging.

[0017] Furthermore, compared to traditional methods that process all data collected in each frame, this invention only processes data with higher quality than historical data, which significantly improves data processing efficiency, reduces computational resource consumption, and thus improves coloring efficiency.

[0018] Furthermore, the coloring process of this invention is implemented through software definition rather than hardware adjustment, achieving high-precision and high-efficiency point cloud coloring without increasing additional hardware costs, increasing device size, or sacrificing battery life. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic flowchart of an embodiment of the scene point cloud map coloring method provided by the present invention; Figure 2 For the present invention Figure 1 A schematic diagram of an embodiment of S104; Figure 3 For the present invention Figure 2 Another embodiment of S202 is illustrated in the flowchart. Figure 4 A schematic flowchart of an embodiment of the present invention for determining the target frame interpolation pose; Figure 5A schematic diagram of an embodiment of the scene point cloud map coloring device provided by the present invention; Figure 6 This is a schematic diagram of an embodiment of the scene point cloud map coloring device provided by the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this invention illustrate operations implemented according to some embodiments of the invention. It should be understood that the operations in the flowcharts may be implemented out of order, and steps without logical contextual relationships may be reversed or performed simultaneously. Furthermore, those skilled in the art, guided by the content of this invention, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0023] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0024] This invention provides a method, apparatus, and device for coloring scene point cloud maps, which will be described below.

[0025] Figure 1 A schematic flowchart of an embodiment of the scene point cloud map coloring method provided by the present invention is shown below. Figure 1 As shown, the coloring methods for scene point cloud maps include: S101. Construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map.

[0026] Specifically, taking the starting point as the origin, the entire scene is rasterized based on a preset grid size to obtain a rasterized scene point cloud map.

[0027] It should be noted that the initial values ​​of the historical motion state parameters are set values, specifically, the initial values ​​of the historical motion state parameters are +∞.

[0028] S102. Obtain the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time.

[0029] The current motion state parameters can be obtained based on the inertial measurement unit (IMU), that is, the inertial measurement unit is bound to the lidar, and the current motion state parameters are obtained simultaneously based on the inertial measurement unit when the lidar collects the current point cloud data.

[0030] S103. For each grid cell, determine whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. S104. If so, then the point cloud located in the grid is colored based on the current frame image data acquired at the current acquisition time, so as to update the point cloud in the grid.

[0031] It should be understood that if not, the point cloud within the raster will not be updated.

[0032] It should also be understood that the scene point cloud map coloring method in this embodiment of the invention can be implemented in any device based on the scene point cloud map coloring method, such as an inspection device or a distance measurement device based on the scene point cloud map coloring method. Specifically, the scene point cloud map coloring method is stored in the aforementioned device as a pre-programmed program. When the device is started, the program is called, and the scene point cloud map coloring method is implemented.

[0033] Compared with existing technologies, the scene point cloud map coloring method provided in this embodiment of the invention, for each grid cell, determines whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. If so, the point cloud within the grid cell is colored based on the current frame image data acquired at the current acquisition time. Compared with the existing technology that indiscriminately uses all acquired data to color the point cloud, which may lead to early high-quality data being covered or contaminated by later low-quality data, this embodiment of the invention utilizes the principle of selective updating to ensure that each grid cell is ultimately colored based on the most stable and highest-quality frame data acquired at the time of acquisition, thereby improving the point cloud coloring accuracy and meeting the needs of refined scene modeling and ranging.

[0034] Furthermore, compared to traditional methods that process all data collected in each frame, this embodiment of the invention only processes data with higher quality than historical data, which significantly improves data processing efficiency, reduces computational resource consumption, and thus improves coloring efficiency.

[0035] Furthermore, the coloring process in this embodiment of the invention is implemented through software definition rather than hardware adjustment, achieving high-precision and high-efficiency point cloud coloring without increasing additional hardware costs, increasing device size, or sacrificing battery life.

[0036] Since a higher motion speed indicates lower data quality due to motion blur and other factors in point cloud scanning, and vice versa, a lower motion speed results in better data quality. Therefore, in some embodiments of this invention, the current motion state parameters and historical motion state parameters should be parameters that characterize motion speed.

[0037] Specifically, the current motion state parameters and the historical motion state can be at least one of angular velocity, linear velocity, or acceleration.

[0038] In a specific embodiment of the present invention, the current motion state parameters include historical angular velocity values, and the historical motion state parameters include historical angular velocity values. Therefore, step S103 specifically comprises: When the current angular velocity value is less than the historical angular velocity value, determine whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters.

[0039] This invention determines whether the data quality of the current frame point cloud data is better based on the current angular velocity value and the historical angular velocity value, which can ensure that point cloud data with more stable and accurate motion is used for coloring, thereby ensuring coloring accuracy.

[0040] Since the coloring process is an iterative optimization process that occurs as data is acquired, and coloring depends on accurate judgment of data quality, in a specific embodiment of the present invention, after step S104, the following step is further included: Use the current motion state parameters as the historical motion state parameters of the grid.

[0041] This invention, by setting and updating historical motion state parameters, ensures that each data quality judgment is compared with the best data quality from the previous multiple judgments, thus ensuring that the final colored point cloud data uses the highest quality point cloud data and maximizing the coloring accuracy.

[0042] In practical use, the current frame point cloud data is acquired by a LiDAR scanner, while the current frame image data is acquired by a camera. However, the LiDAR and camera have different acquisition frequencies. For example, the LiDAR acquisition frequency is 10Hz, while the camera acquisition frequency is 30Hz. Due to the difference in acquisition frequencies, the point cloud data and image data will be out of sync, resulting in point cloud distortion.

[0043] To address this technical problem, in some embodiments of the present invention, such as... Figure 2 As shown, step S104 includes: S201. Obtain the previous basic pose output by the SLAM algorithm at the previous acquisition time and the current basic pose output at the current acquisition time.

[0044] Among them, the SLAM algorithm is a mature algorithm, and the process of generating basic poses based on the SLAM algorithm will not be described in detail here.

[0045] S202. Based on the current motion state parameters, perform pose interpolation between the previous basic pose and the current basic pose to generate an interpolated pose sequence containing at least one intermediate pose.

[0046] The interpolated pose sequence consists of the previous basic pose, at least one intermediate pose, and the current basic pose.

[0047] S203. Determine the target interpolation pose with the highest matching degree with the current frame image data in the interpolation pose sequence, and project the point cloud in the raster onto the current frame image data based on the target interpolation pose for coloring.

[0048] It should be understood that coloring is essentially the process of accurately mapping the pixel colors in image data to the corresponding geometric positions of points in a 3D point cloud. Specifically, it involves constructing a mapping relationship between point cloud data and image data, and then assigning colors from the image data to the point cloud data based on this mapping relationship to generate a scene point cloud map.

[0049] The embodiments of the present invention dynamically insert multiple intermediate poses between the previous base pose and the current base pose to compensate for the problem of excessive interval between base poses, so as to make the poses of image data and point cloud data more matched, eliminate the problem of asynchrony, and thus achieve the purpose of improving the coloring accuracy.

[0050] Furthermore, by setting pose interpolation based on the current motion state parameters, the embodiments of the present invention can improve the rationality of the inserted intermediate poses. For example, when the motion speed represented by the current motion state parameters is relatively fast, more intermediate poses can be inserted to ensure the synchronization of point cloud and image; when the motion speed represented by the current motion state parameters is relatively slow, fewer intermediate poses can be inserted to reduce the consumption of computing resources.

[0051] In specific embodiments of the present invention, such as Figure 3 As shown, step S202 includes: S301. Determine the number of interpolations for the intermediate pose based on the current motion state parameters.

[0052] Specifically, the number of interpolations is:

[0053] In the formula, This represents the number of interpolations. This is the speed coefficient, which can be set or adjusted according to the hardware performance of the acquisition device and the required accuracy. These are the parameters of the current motion state; It is the modulo operator; This is the floor operator.

[0054] S302. Based on the number of interpolations, perform pose interpolation between the previous basic pose and the current basic pose using linear interpolation to obtain the interpolated pose sequence.

[0055] Specifically, the pose interval of the interpolated frames is determined based on the number of interpolations, i.e.:

[0056] In the formula, This is the previous basic pose; This is the current base pose; Posture interval.

[0057] Linear interpolation is performed based on the pose interval to obtain at least one intermediate pose. The intermediate pose, together with the previous base pose and the current base pose, constitutes the interpolated pose sequence.

[0058] Specifically, the intermediate pose can be represented as:

[0059]

[0060] In the formula, This is the j-th intermediate pose.

[0061] The embodiments of the present invention can dynamically improve the basic pose that was originally acquired at a low frequency through the above interpolation process, ensuring that each frame of image data can be matched with the corresponding pose data, thus solving the synchronization problem between the lidar and the camera.

[0062] In some embodiments of the present invention, such as Figure 4 As shown, the step S203 of determining the target interpolation pose with the highest matching degree to the current frame image data in the interpolation pose sequence includes: S401. Extract the image geometric features from the current frame image data and the point cloud geometric features from the current frame point cloud data; S402. Based on each intermediate pose, perform projection transformation on the geometric features of the point cloud to obtain geometric projection features.

[0063] The projection transformation utilizes the transformation matrix calibrated using the extrinsic parameters of the camera and lidar. S403. Determine the similarity between the geometric features and geometric projection features of the image, and take the intermediate pose corresponding to the highest similarity value as the target interpolation pose.

[0064] The embodiments of the present invention determine the target interpolation pose by the similarity between image geometric features and geometric projection features, which can improve the matching degree between the selected target interpolation pose and image data and point cloud data, thereby further ensuring the accuracy of point cloud coloring.

[0065] Because substation environments are highly structured, with equipment such as supports, beams, and insulators containing numerous regular horizontal and vertical edges, these structured features remain relatively stable across different viewpoints, making them ideal for accurate feature matching. Therefore, in some embodiments of this invention, both image geometric features and point cloud geometric features are structured linear features.

[0066] This invention, by setting both image geometric features and point cloud geometric features to be structured straight line features, can achieve accurate pose matching in substation scenarios through simple straight line features, balancing shading accuracy and shading efficiency.

[0067] Specifically, the process of extracting image geometric features is as follows: the edge line features of the current frame image data are extracted using the Canny operator, and straight lines are extracted from the edge line features using the Hough transform.

[0068] The point cloud geometric feature extraction process is as follows: the horizontal and vertical point clouds in the current frame point cloud data are extracted using the normal vector estimation algorithm, and the straight line features are accurately extracted according to the Hough transform method.

[0069] In this embodiment of the invention, after coloring the scene point cloud map, each grid cell in the map carries unique coordinate and color information; that is, a grid cell can be represented as... , , This refers to coordinate information; Color information is used. Based on coordinate information, the distance between devices in the scene can be calculated to achieve distance measurement. Based on color information, the area between equipment and personnel can be distinguished. Based on color changes, equipment abnormalities can be identified, such as color differences caused by insulator damage, so as to realize the safety monitoring of operators and the inspection of equipment status.

[0070] In summary, the scene point cloud map coloring method proposed in this embodiment of the invention has the following beneficial effects: (1) Higher synchronization accuracy: Compared with the existing software synchronization scheme, this invention generates multiple frames of pose data through 10Hz SLAM basic pose interpolation, so that each frame of point cloud and image can be matched with the accurate target pose, which meets the high-precision modeling and ranging requirements of power transmission, substation and distribution scenarios. (2) Lower hardware cost: Unlike hardware synchronization schemes, there is no need to integrate an additional trigger module. Data synchronization is achieved only through software algorithms, which can reduce equipment cost, reduce power consumption, and extend the equipment's battery life. This is more in line with the portability and usage requirements of handheld power transmission, substation and distribution scenario modeling and ranging equipment. (3) Stronger adaptability: The number of interpolated frames is dynamically adjusted based on the device's movement speed. When the device moves at high speed, the number of interpolated frames is automatically increased to ensure synchronization; when the device moves at low speed, the number of interpolated frames is reduced to reduce computing resource consumption and adapt to the speed changes of different operating scenarios. (4) Better map quality: By using rasterized iterative coloring, only the raster with more accurate data is updated, avoiding redundant calculations while improving the color uniformity and detail clarity of the color point cloud map, making the status of distribution network equipment (such as insulator damage and line corrosion) easier to identify, and providing a more reliable three-dimensional data foundation for subsequent distance measurement and safety monitoring.

[0071] On the other hand, embodiments of the present invention also provide a coloring device for scene point cloud maps, such as... Figure 5 As shown, the scene point cloud map coloring device 500 includes: The rasterized map construction unit 501 is used to construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map. The current motion state parameter acquisition unit 502 is used to acquire the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time. The data quality comparison unit 503 is used to determine, for each grid cell, whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. The map coloring unit 504 is used to color the point cloud located in the grid based on the current frame image data acquired at the current acquisition time when the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters, so as to update the point cloud in the grid.

[0072] The scene point cloud map coloring device 500 provided in the above embodiments can implement the technical solutions described in the above scene point cloud map coloring method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above scene point cloud map coloring method embodiments, and will not be repeated here.

[0073] like Figure 6As shown, the present invention also provides a scene point cloud map coloring device 600. The scene point cloud map coloring device 600 includes a data acquisition sub-device 610 and a data processing sub-device 620. The data acquisition sub-device 610 includes a camera 611 and a lidar 612, and the data processing sub-device 620 includes a processor 621, a memory 622 and a display 623. Figure 6 Only a portion of the components of the scene point cloud map shading device 600 are shown; however, it should be understood that implementation of all shown components is not required, and more or fewer components may be implemented instead.

[0074] Camera 611 is used to acquire image data of the current frame; The LiDAR 612 is used to collect point cloud data in the current frame; In some embodiments, processor 621 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in memory 622 or process data, such as the scene point cloud map coloring method in this invention.

[0075] In some embodiments, the memory 622 may be an internal storage unit of the scene point cloud map shading device 600, such as a hard disk or memory of the scene point cloud map shading device 600. In other embodiments, the memory 622 may also be an external storage device of the scene point cloud map shading device 600, such as a pluggable hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the scene point cloud map shading device 600.

[0076] In some embodiments, display 623 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 623 is used to display the shaded scene point cloud map on a display interface. Components 610-620 of the scene point cloud map shading device 600 communicate with each other via a device bus.

[0077] In some embodiments of the present invention, when the processor 621 executes the scene point cloud map coloring program in the memory 622, the following steps can be implemented: Construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map; Acquire the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time; For each grid cell, determine whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. If so, the point cloud within the raster is colored based on the current frame image data acquired at the current acquisition time in order to update the point cloud within the raster.

[0078] It should be understood that when the processor 621 executes the scene point cloud map coloring program in the memory 622, in addition to the functions mentioned above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0079] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0080] The above provides a detailed description of the coloring method, apparatus, and device for scene point cloud maps provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for coloring a scene point cloud map, characterized in that, include: Construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map; Acquire the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time; For each grid cell, determine whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. If so, the point cloud located within the grid is colored based on the current frame image data acquired at the current acquisition time, so as to update the point cloud within the grid.

2. The scene point cloud map coloring method according to claim 1, characterized in that, The historical motion state parameters include historical angular velocity values, and the current motion state parameters include current angular velocity values; then, determining whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters includes: When the current angular velocity value is less than the historical angular velocity value, it is determined whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters.

3. The scene point cloud map coloring method according to claim 1, characterized in that, The method further includes: The current motion state parameters are used as the historical motion state parameters of the grid.

4. The scene point cloud map coloring method according to claim 1, characterized in that, Coloring the point cloud located within the raster based on the image data acquired at the current acquisition time includes: Obtain the previous basic pose output by the SLAM algorithm at the previous acquisition time and the current basic pose output at the current acquisition time; Based on the current motion state parameters, pose interpolation is performed between the previous base pose and the current base pose to generate an interpolated pose sequence containing at least one intermediate pose. The target interpolation pose with the highest matching degree to the current frame image data in the interpolation pose sequence is determined, and the point cloud in the grid is projected onto the current frame image data based on the target interpolation pose for coloring.

5. The scene point cloud map coloring method according to claim 4, characterized in that, The step of performing pose interpolation between the previous base pose and the current base pose based on the current motion state parameters includes: The number of interpolations for the intermediate pose is determined based on the current motion state parameters; Based on the number of interpolations, pose interpolation is performed between the previous basic pose and the current basic pose using linear interpolation to obtain the interpolated pose sequence.

6. The scene point cloud map coloring method according to claim 5, characterized in that, The number of interpolations is: In the formula, This represents the number of interpolations. The velocity coefficient; These are the parameters of the current motion state; It is the modulo operator; This is the floor operator.

7. The scene point cloud map coloring method according to claim 4, characterized in that, Determining the target interpolation pose with the highest matching degree to the current frame image data in the interpolation pose sequence includes: Extract the image geometric features from the current frame image data and the point cloud geometric features from the current frame point cloud data; Based on each of the intermediate poses, the geometric features of the point cloud are projected and transformed to obtain geometric projection features; Determine the similarity between the image geometric features and the geometric projection features, and use the intermediate pose corresponding to the highest similarity value as the target interpolation pose.

8. The scene point cloud map coloring method according to claim 7, characterized in that, Both the image geometric features and the point cloud geometric features are structured straight line features.

9. A coloring device for a scene point cloud map, characterized in that, include: A rasterized map construction unit is used to construct a rasterized scene point cloud map and obtain the historical motion state parameters of each grid in the rasterized scene point cloud map. The current motion state parameter acquisition unit is used to acquire the current frame point cloud data and the current motion state parameters corresponding to the current acquisition time. A data quality comparison unit is used to determine, for each of the grid cells, whether the data quality represented by the current motion state parameters is better than the data quality represented by the historical motion state parameters. The map coloring unit is used to color the point cloud located in the grid based on the current frame image data acquired at the current acquisition time when the data quality represented by the current motion state parameter is better than the data quality represented by the historical motion state parameter, so as to update the point cloud in the grid.

10. A coloring device for a scene point cloud map, characterized in that, include: The data acquisition sub-device includes a camera and a lidar, and the data processing sub-device includes a memory and a processor. The camera is used to acquire image data of the current frame; The lidar is used to collect point cloud data in the current frame; The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the scene point cloud map coloring method according to any one of claims 1 to 8.

Images