Tunnel point cloud registration method, device, equipment, storage medium and program product
By mapping tunnel point clouds onto a two-dimensional plane to generate a flatness map and using image matching methods combined with the ICP algorithm for tunnel point cloud registration, the problems of excessive computation and inaccurate registration are solved, achieving efficient and accurate point cloud registration.
Patent Information
- Application Number
- CN202511472580.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-15
- Publication Date
- 2025-12-26
AI Technical Summary
Existing technologies require excessive computation for registering 3D point clouds of long tunnels, which is insufficient for conventional computing power. Furthermore, the convergence of tunnel geometric features leads to inaccurate registration results.
The three-dimensional tunnel point cloud is mapped onto a two-dimensional plane to generate a flatness map. Image matching methods are used for coarse matching, and algorithms such as ICP are combined for fine registration, which reduces computational complexity and improves accuracy.
It effectively avoids the bottleneck of massive point cloud computing, improves the accuracy and efficiency of point cloud registration, and solves the registration failure problem caused by similar geometric features of tunnels.
Smart Images
Figure CN121213626A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of point cloud registration technology, specifically to tunnel point cloud registration methods, apparatus, equipment, storage media, and program products. Background Technology
[0002] 3D laser scanning technology is widely used in tunnel engineering for twin modeling, quality inspection, and volume measurement. Tunnels are typically several kilometers long, and the point cloud models obtained from 3D scanning are enormous, with hundreds of thousands of points per meter and billions per kilometer.
[0003] Point cloud registration refers to the process of using algorithms to solve for the optimal spatial transformation relationship of 3D point clouds collected from different sources, at different times, or from different devices, and adjusting the spatial position of the source point cloud so that it is precisely aligned with the target point cloud in the same coordinate system.
[0004] However, when registering newly collected point clouds with large-scale known point clouds, the computational cost far exceeds conventional computing power, and due to the convergence of tunnel geometric features, it is difficult to obtain correct registration results. Summary of the Invention
[0005] In view of this, the present invention provides a method, apparatus, device, storage medium and program product for tunnel point cloud registration, so as to improve the accuracy of registration.
[0006] In a first aspect, the present invention provides a method for registering tunnel point clouds, the method comprising: acquiring an original tunnel point cloud and a tunnel point cloud to be registered for a target tunnel, wherein the original tunnel point cloud is a complete point cloud of the target tunnel, and the tunnel point cloud to be registered is a local point cloud in the target tunnel; mapping the original tunnel point cloud and the tunnel point cloud to be registered onto a plane respectively to obtain an original point cloud flatness map and a point cloud flatness map to be registered; matching the point cloud flatness map to be registered with the original point cloud flatness map to determine a target point cloud flatness sub-map from the original point cloud flatness map, wherein the target point cloud flatness sub-map is a local map in the original point cloud flatness map that matches the point cloud flatness map to be registered; determining a target tunnel point cloud that matches the tunnel point cloud to be registered from the original tunnel point cloud based on the target point cloud flatness sub-map; and performing fine registration between the tunnel point cloud to be registered and the target tunnel point cloud.
[0007] In this implementation, by mapping the original 3D tunnel point cloud and the tunnel point cloud to be registered onto a 2D plane respectively, while preserving the geometric features of the 3D data, the computational bottleneck of directly processing massive point clouds can be avoided. The original tunnel point cloud and the tunnel point cloud to be registered are converted into images, and image matching methods are used for coarse matching to reduce the computational complexity of matching. Furthermore, precise point cloud matching is performed, which can improve the accuracy of point cloud matching while reducing computational load.
[0008] In one optional implementation, the original tunnel point cloud and the tunnel point cloud to be registered are mapped onto a plane to obtain an original point cloud flatness map and a point cloud flatness map to be registered, respectively. This includes: obtaining the standard design contour of the target tunnel; mapping the original tunnel point cloud and the tunnel point cloud to be registered onto the standard design contour of the tunnel, and calculating the distance between each point cloud and the standard design contour of the tunnel, obtaining the original point cloud distance of each point cloud in the original tunnel point cloud and the point cloud distance to be registered of each point cloud in the tunnel point cloud to be registered; unfolding the original tunnel point cloud and the tunnel point cloud to be registered into a plane according to the standard design contour of the tunnel, respectively, to obtain an original planar point cloud and a planar point cloud to be registered; rendering the original point cloud distance of each point cloud in the original planar point cloud to generate an original point cloud flatness map; and rendering the point cloud distance to be registered of each point cloud in the planar point cloud to be registered to generate a point cloud flatness map to be registered.
[0009] In this implementation, the same surface fitting and mapping method is used to convert the original tunnel point cloud and the tunnel point cloud to be registered into a flatness map, which can ensure the comparability of the features of the original tunnel point cloud and the tunnel point cloud to be registered. The proposed method of converting point cloud flatness into an image can improve the difficulty of point cloud image conversion and reduce the amount of registration calculation.
[0010] In one optional implementation, rendering the original point cloud distance of each point cloud in the original planar point cloud to generate an original point cloud flatness map includes: calculating a first flatness value of each point cloud in the original planar point cloud based on the original point cloud distance; rendering a cloud map based on the first flatness value of each point cloud in the original planar point cloud to generate an original point cloud flatness map; and rendering the distance of each point cloud to be registered in the planar point cloud to be registered to generate a point cloud flatness map to be registered includes: calculating a second flatness value of each point cloud in the planar point cloud to be registered based on the distance of the point cloud to be registered; and rendering a cloud map based on the second flatness value of each point cloud in the planar point cloud to be registered to generate a point cloud flatness map to be registered.
[0011] In one optional implementation, matching the point cloud flatness map to be registered with the original point cloud flatness map, and determining the target point cloud flatness sub-map from the original point cloud flatness map, includes: cutting the original point cloud flatness map into multiple original point cloud flatness sub-maps, and marking the tunnel information of each original point cloud flatness sub-map, the tunnel information including the tunnel name and tunnel station number; matching the point cloud flatness map to be registered with the multiple original point cloud flatness sub-maps, and determining the target point cloud flatness sub-map that matches the point cloud flatness map to be registered from the multiple original point cloud flatness sub-maps.
[0012] In one optional implementation, the original point cloud flatness map is cut into multiple original point cloud flatness sub-maps, including: obtaining a preset cutting specification and a preset overlap specification, wherein the preset cutting specification is the size of each original point cloud flatness sub-map, and the preset overlap specification is the overlap size of adjacent original point cloud flatness sub-maps; and cutting the original point cloud flatness map into multiple original point cloud flatness sub-maps according to the preset cutting specification and the preset overlap specification.
[0013] In this implementation, the smoothness map of the long tunnel is divided into blocks, the overlapping areas are preserved to ensure continuity, and segmented processing is supported, which can reduce the difficulty of image matching.
[0014] In one optional implementation, determining a target tunnel point cloud that matches the tunnel point cloud to be registered from the original tunnel point cloud based on the target point cloud flatness sub-map includes: acquiring tunnel information of the target point cloud flatness sub-map; and determining the local point cloud corresponding to the target point cloud flatness sub-map in the original tunnel point cloud based on the tunnel information, thereby obtaining the target tunnel point cloud that matches the tunnel point cloud to be registered.
[0015] Secondly, the present invention provides a tunnel point cloud registration device, which includes: an acquisition module for acquiring an original tunnel point cloud and a tunnel point cloud to be registered for a target tunnel, wherein the original tunnel point cloud is a complete point cloud of the target tunnel, and the tunnel point cloud to be registered is a local point cloud in the target tunnel; a mapping module for mapping the original tunnel point cloud and the tunnel point cloud to be registered onto a plane respectively to obtain a flatness map of the original point cloud and a flatness map of the point cloud to be registered; and a first matching module for... The system first matches the flatness map of the point cloud to be registered with the flatness map of the original point cloud, and determines the flatness sub-map of the target point cloud from the flatness map of the original point cloud. The flatness sub-map of the target point cloud is a local map in the flatness map of the original point cloud that matches the flatness map of the point cloud to be registered. The second matching module is used to determine the target tunnel point cloud that matches the tunnel point cloud to be registered from the original tunnel point cloud based on the flatness sub-map of the target point cloud. The registration module is used to perform fine registration between the tunnel point cloud to be registered and the target tunnel point cloud.
[0016] Thirdly, the present invention provides a computer device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the tunnel point cloud registration method of the first aspect or any corresponding embodiment described above.
[0017] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the tunnel point cloud registration method of the first aspect or any corresponding embodiment thereof.
[0018] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the tunnel point cloud registration method of the first aspect or any corresponding embodiment thereof. Attached Figure Description
[0019] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0020] Figure 1 This is a flowchart of a tunnel point cloud registration method according to an embodiment of the present invention; Figure 2 This is a flowchart of another tunnel point cloud registration method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a primitive tunnel point cloud according to an embodiment of the present invention; Figure 4 This is a schematic diagram of an original planar point cloud according to an embodiment of the present invention; Figure 5 This is a schematic diagram of an original point cloud flatness map according to an embodiment of the present invention; Figure 6 This is a schematic diagram of a method for registering an original tunnel point cloud and a tunnel point cloud to be registered according to an embodiment of the present invention; Figure 7 This is a schematic diagram of an original point cloud flatness sub-map according to an embodiment of the present invention; Figure 8 This is a structural block diagram of a tunnel point cloud registration device according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation
[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] According to an embodiment of the present invention, a method for tunnel point cloud registration is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0023] When registering scanned point clouds of a local tunnel construction section with large-scale known point clouds (such as complete point clouds of the entire tunnel), a dual core dilemma arises: Firstly, conventional computing power is completely exceeded, meaning the data volume of the large-scale known point cloud far exceeds the memory and computing power of ordinary computers, and directly loading the full data will cause program lag or even crash; secondly, the convergence of tunnel geometric features leads to registration inaccuracies, i.e., tunnels are mostly long cylindrical or horseshoe-shaped structures, with inner walls composed of numerous repetitive arc surfaces and textures parallel to the axis. Local features in the new point cloud are highly similar to features at multiple non-overlapping locations in the known point cloud, making the registration algorithm... The high risk of mismatching points from a new point cloud to non-corresponding regions of a known point cloud is a core technical bottleneck in large-scale tunnel point cloud registration. Therefore, this application proposes a tunnel point cloud registration method that maps the original 3D tunnel point cloud and the tunnel point cloud to be registered onto a 2D plane while preserving the geometric features of the 3D data, thus avoiding the computational bottleneck of directly processing massive point clouds. The original tunnel point cloud and the tunnel point cloud to be registered are converted into images, and image matching methods are used for coarse matching to reduce computational complexity. Furthermore, precise point cloud matching is performed, which improves the accuracy of point cloud matching while reducing computational load.
[0024] This embodiment provides a method for tunnel point cloud registration. Figure 1 This is a flowchart of a tunnel point cloud registration method according to an embodiment of the present invention. It should be noted that if substantially the same result is obtained, this embodiment does not necessarily reflect that result. Figure 1 The illustrated process sequence is limited. For example... Figure 1 As shown, the process includes the following steps: Step S101: Obtain the original tunnel point cloud and the tunnel point cloud to be registered for the target tunnel.
[0025] Among them, the original tunnel point cloud is the complete point cloud of the target tunnel.
[0026] Specifically, a complete high-precision scan of the target tunnel is performed to obtain a complete point cloud covering the entire length and cross-section of the target tunnel, resulting in the original tunnel point cloud. The original tunnel point cloud includes all point cloud features of the target tunnel from the entrance to the exit, such as the arch and sidewalls.
[0027] In one possible implementation, the corresponding laser scanning equipment is selected according to the tunnel type, and laser scanning equipment is deployed at the tunnel entrance, exit and key intermediate locations. Multiple laser scanning equipment are used to perform laser scanning, and the tunnel number and location are recorded. The point cloud data obtained from the scanning of multiple laser scanning equipment are stitched together to obtain the original tunnel point cloud of the target tunnel.
[0028] Among them, the point cloud of the tunnel to be registered is a local point cloud in the target tunnel.
[0029] Specifically, by focusing on a local area of the tunnel, the point cloud features of the target tunnel that need to be registered are obtained, resulting in the point cloud of the target tunnel to be registered. The point cloud of the tunnel to be registered can be the point cloud of areas such as the damaged section, the construction section, or the equipment installation area.
[0030] In one possible implementation, noise reduction processing is further performed on the original tunnel point cloud and the tunnel point cloud to be registered.
[0031] Step S102: Map the original tunnel point cloud and the tunnel point cloud to be registered onto the plane respectively to obtain the original point cloud flatness map and the point cloud flatness map to be registered.
[0032] Both the original tunnel point cloud and the tunnel point cloud to be registered retain the outline shape of the original tunnel. The original tunnel point cloud is unfolded according to its outline shape, mapped onto a plane, and its concavity and convexity features are preserved to generate a corresponding original point cloud flatness map. Similarly, the tunnel point cloud to be registered is unfolded according to its outline shape, mapped onto a plane, and its concavity and convexity features are preserved to generate a corresponding point cloud flatness map.
[0033] Among them, the original point cloud flatness map and the point cloud flatness map to be registered can characterize the concavity and convexity features of the original tunnel point cloud and the tunnel point cloud to be registered, that is, reflect the flatness of the original tunnel point cloud and the tunnel point cloud to be registered.
[0034] Specifically, in one possible implementation, the original 3D tunnel point cloud and the tunnel point cloud to be registered are projected onto a 2D plane along an axis to obtain a corresponding planar image. In another possible implementation, each point cloud in the original 3D tunnel point cloud and the tunnel point cloud to be registered is projected onto a 2D plane to obtain a corresponding planar image. Further, by combining the corresponding point cloud planar images and the flatness of each point cloud, corresponding original point cloud flatness images and point cloud flatness images to be registered are generated.
[0035] Step S103: Match the point cloud flatness map to be registered with the original point cloud flatness map, and determine the target point cloud flatness sub-map from the original point cloud flatness map.
[0036] Generally, the original point cloud flatness map is larger than the point cloud flatness map to be registered.
[0037] Specifically, an image matching method is used to match the point cloud flatness map to be registered with the original point cloud flatness map. The local map that best matches the point cloud flatness map to be registered is determined from the original point cloud flatness map, and the target point cloud flatness sub-map is extracted.
[0038] Step S104: Based on the flatness sub-map of the target point cloud, determine the target tunnel point cloud that matches the tunnel point cloud to be registered from the original tunnel point cloud.
[0039] Based on the position of the target point cloud flatness sub-image in the original point cloud flatness map, the target tunnel point cloud is extracted from the corresponding position in the original tunnel point cloud. Here, the target tunnel point cloud is the point cloud corresponding to the target point cloud flatness sub-image.
[0040] Step S105: Perform fine registration between the point cloud of the tunnel to be registered and the point cloud of the target tunnel.
[0041] The point cloud of the tunnel to be registered is precisely registered with the point cloud of the target tunnel according to the point cloud registration method.
[0042] Point cloud registration methods include ICP (Iterative Closest Point) algorithm, NDT (Normal Distribution) algorithm, 3DSC algorithm, PFH (Point Feature Histogram) and FPFH (Fast Point Feature Histogram) algorithms.
[0043] The tunnel point cloud registration method provided in this embodiment maps the original 3D tunnel point cloud and the tunnel point cloud to be registered onto a 2D plane respectively, while preserving the geometric features of the 3D data, thus avoiding the computational bottleneck of directly processing massive point clouds. The original tunnel point cloud and the tunnel point cloud to be registered are converted into images, and image matching methods are used for coarse matching to reduce the computational complexity of matching. Furthermore, precise point cloud matching is performed, which can improve the accuracy of point cloud matching while reducing computational load.
[0044] This embodiment provides a method for tunnel point cloud registration. Figure 2 This is a flowchart of another tunnel point cloud registration method according to an embodiment of the present invention. It should be noted that if substantially the same result is obtained, this embodiment does not necessarily use it. Figure 2 The illustrated process sequence is limited. For example... Figure 2 As shown, the process includes the following steps: Step S201: Obtain the original tunnel point cloud and the tunnel point cloud to be registered for the target tunnel.
[0045] Please see details Figure 1 Step S101 of the illustrated embodiment will not be described again here.
[0046] Step S202: Map the original tunnel point cloud and the tunnel point cloud to be registered onto the plane respectively to obtain the original point cloud flatness map and the point cloud flatness map to be registered.
[0047] Specifically, step S202 includes: Step S2021: Obtain the standard design profile of the target tunnel.
[0048] Among them, the standard design profile of the tunnel is the target design shape of the target tunnel during construction.
[0049] In one possible implementation, construction drawings or completion reports of the target tunnel are extracted from construction documents to obtain the standard design outline of the target tunnel.
[0050] Step S2022: Map the original tunnel point cloud and the tunnel point cloud to be registered onto the standard design profile of the tunnel, respectively, and calculate the distance between each point cloud and the standard design profile of the tunnel to obtain the original point cloud distance of each point cloud in the original tunnel point cloud and the point cloud distance to be registered of each point cloud in the tunnel point cloud to be registered.
[0051] In one possible implementation, the original tunnel point cloud is mapped onto the standard design profile of the tunnel using a normal mapping method.
[0052] Specifically, using the normal direction of the original tunnel point cloud surface as the projection path, each point cloud in the original tunnel point cloud is projected onto the standard design profile of the tunnel along the normal direction, thus obtaining the reference point corresponding to each point cloud in the original tunnel point cloud on the standard design profile of the tunnel. Further, the distance between each point cloud in the original tunnel point cloud and the corresponding reference point on the standard design profile of the tunnel is calculated, thus obtaining the original point cloud distance of each point cloud in the original tunnel point cloud.
[0053] Similarly, the above method is used to map the tunnel point cloud to be registered onto the standard design profile of the tunnel, so as to obtain the reference point corresponding to each point cloud in the tunnel point cloud on the standard design profile of the tunnel, and calculate the distance between each point cloud in the tunnel point cloud and the corresponding reference point on the standard design profile of the tunnel, so as to obtain the distance of each point cloud in the tunnel point cloud to be registered.
[0054] Step S2023: Unfold the original tunnel point cloud and the tunnel point cloud to be registered into a plane according to the standard tunnel design outline, to obtain the original planar point cloud and the planar point cloud to be registered.
[0055] Specifically, the original tunnel point cloud is unfolded onto a plane along with the standard design outline of the tunnel, and the original point cloud distance of each point cloud in the original tunnel point cloud is retained to generate an original planar point cloud. The original tunnel point cloud retains the concave and convex features of the original tunnel point cloud.
[0056] For example, please refer to Figure 3 and Figure 4 , Figure 3 This is a schematic diagram of a raw tunnel point cloud according to an embodiment of the present invention. Figure 4 This is a schematic diagram of a primitive planar point cloud according to an embodiment of the present invention. Figure 3 The original tunnel point cloud in the image is unfolded into a plane according to the standard tunnel design outline, resulting in... Figure 4 The original planar point cloud is shown.
[0057] Similarly, the tunnel point cloud to be registered is unfolded onto a plane along with the standard design contour of the tunnel, and the distance of each point cloud to be registered is retained in the tunnel point cloud to be registered, generating a planar point cloud to be registered, on which the concave and convex features of the tunnel point cloud to be registered are retained.
[0058] Step S2024: Render the original point cloud distance of each point cloud in the original planar point cloud to generate an original point cloud flatness map; render the point cloud distance to be registered of each point cloud in the planar point cloud to be registered to generate a point cloud flatness map to be registered.
[0059] In one possible implementation, a first flatness value is calculated for each point cloud in the original planar point cloud based on the original point cloud distance. Specifically, the original point cloud distance of each point cloud in the original planar point cloud is used as the flatness value. When a point cloud is below the plane, the flatness value is negative; when a point cloud is above the plane, the flatness value is positive. Further, cloud rendering is performed on the first flatness value of each point cloud in the original planar point cloud to generate an original point cloud flatness map.
[0060] In one possible implementation, the original planar point cloud is divided into multiple preset grid regions. For each preset grid region, the highest point is selected based on the original point cloud distance of each point cloud, denoted as point A. The connection between point A and other point clouds in the preset grid region is obtained, and the point with the lowest slope is selected, denoted as point B. The distance L1 between line segments AB is calculated, the distance L2 between points between line segments AB and between line segments AB is calculated, and the first flatness value of the preset grid region is calculated as P = L2 / L1.
[0061] In one possible implementation, the first flatness value obtained by quantization is correlated with the cloud map color gradient value of the image, the cloud map color gradient value is mapped onto the plane, and cloud map rendering is performed to generate a complete and continuous original point cloud flatness map.
[0062] For example, please refer to Figure 5 , Figure 5 This is a schematic diagram of an original point cloud flatness map according to an embodiment of the present invention. Figure 4 The original planar point cloud rendering shown Figure 5 The original point cloud flatness map is shown.
[0063] Similarly, using the above method, each point cloud in the plane point cloud to be registered is processed, and the second flatness value of each point cloud in the plane point cloud to be registered is calculated based on the distance of the point cloud to be registered; cloud map rendering is performed on the second flatness value of each point cloud in the plane point cloud to be registered to generate a flatness map of the point cloud to be registered.
[0064] In this implementation, the same surface fitting and mapping method is used to convert the original tunnel point cloud and the tunnel point cloud to be registered into a flatness map, which can ensure the comparability of the features of the original tunnel point cloud and the tunnel point cloud to be registered. The proposed method of converting point cloud flatness into an image can improve the difficulty of point cloud image conversion and reduce the amount of registration calculation.
[0065] Step S203: Match the point cloud flatness map to be registered with the original point cloud flatness map, and determine the target point cloud flatness sub-map from the original point cloud flatness map.
[0066] The target point cloud flatness sub-map is a local map in the original point cloud flatness map that matches the flatness map of the point cloud to be registered. Please refer to [link / reference]. Figure 6 , Figure 6 This is a schematic diagram of a method for registering an original tunnel point cloud and a tunnel point cloud to be registered according to an embodiment of the present invention.
[0067] Specifically, step S203 includes: Step S2031: Cut the original point cloud flatness map into multiple original point cloud flatness sub-maps and mark the tunnel information of each original point cloud flatness sub-map.
[0068] In one possible implementation, the original point cloud flatness map is divided into multiple original point cloud flatness sub-maps according to certain specifications, and each sub-map has a certain overlap area.
[0069] Specifically, preset cutting specifications and preset overlap specifications are obtained, where the preset cutting specifications are the size of each original point cloud flatness sub-image, and the preset overlap specifications are the overlap size of adjacent original point cloud flatness sub-images. The original point cloud flatness image is then cut into multiple original point cloud flatness sub-images according to the preset cutting specifications and preset overlap specifications.
[0070] For example, please refer to Figure 7 , Figure 7 This is a schematic diagram of an original point cloud flatness sub-map according to an embodiment of the present invention. The sub-map is divided according to preset cutting and overlapping specifications. Figure 5 The original point cloud flatness map shown is horizontally divided into multiple original point cloud flatness sub-maps of the same size, and there are overlapping areas between adjacent original point cloud flatness sub-maps.
[0071] Furthermore, tunnel information is marked in each original point cloud flatness sub-map. This tunnel information includes the tunnel name and tunnel station number.
[0072] Step S2032: Match the point cloud flatness map to be registered with multiple original point cloud flatness sub-maps, and determine the target point cloud flatness sub-map that matches the point cloud flatness map to be registered from the multiple original point cloud flatness sub-maps.
[0073] In one implementation, feature points are extracted from the point cloud flatness map to be registered and multiple original point cloud flatness sub-maps using image recognition technology. The feature point similarity of the point cloud flatness map to be registered and multiple original point cloud flatness sub-maps is calculated using image recognition technology. The feature similarity corresponding to each original point cloud flatness sub-map is calculated, and the original point cloud flatness sub-map with the highest feature similarity is selected as the target point cloud flatness sub-map to be matched with the point cloud flatness map to be registered.
[0074] For example, image recognition technologies include SIFT (Scale-Invariant Feature Transform), SURF (Speeded Up Robust Features), ORB (Oriented Fast and Rotated BRIEF), etc.
[0075] like Figure 6 As shown, the existing tunnel point cloud smoothness map set and the smoothness map of the newly acquired point cloud are image matched to obtain the matching target point cloud smoothness sub-map.
[0076] Step S204: Based on the flatness sub-map of the target point cloud, determine the target tunnel point cloud that matches the tunnel point cloud to be registered from the original tunnel point cloud.
[0077] Specifically, the tunnel information of the target point cloud flatness sub-map is obtained, the tunnel name and tunnel station number corresponding to the tunnel point cloud to be registered are determined, and the position of the tunnel point cloud to be registered in the tunnel is determined according to the tunnel name and tunnel station number. The local point cloud corresponding to the target point cloud flatness sub-map is determined, and the target tunnel point cloud matching the tunnel point cloud to be registered is obtained.
[0078] like Figure 6 As shown, the local point cloud to be matched is determined in the existing tunnel point cloud based on the flatness sub-image of the target point cloud obtained by image matching.
[0079] Step S205: Perform fine registration between the point cloud of the tunnel to be registered and the point cloud of the target tunnel.
[0080] In one possible implementation, the ICP algorithm is used to perform fine registration between the point cloud of the tunnel to be registered and the point cloud of the target tunnel.
[0081] Specifically, an initial transformation matrix is constructed between the point cloud of the tunnel to be registered and the point cloud of the target tunnel. The initial transformation matrix includes an initial translation matrix and an initial rotation matrix. For each point cloud in the point cloud of the tunnel to be registered, the nearest point cloud in the point cloud of the target tunnel is found to form a point cloud pair. Using the least squares method, the distance error between the two point clouds in the point cloud pair is minimized, and the optimal transformation matrix between each point cloud in the point cloud of the tunnel to be registered and each point cloud in the target tunnel is obtained. The optimal transformation matrix includes the optimal translation matrix and the optimal rotation matrix.
[0082] Furthermore, the optimal transformation matrix is applied to the point cloud of the tunnel to be registered, and the spatial position of the point cloud is transformed. For each point cloud in the updated point cloud of the tunnel to be registered, the nearest point cloud in the target tunnel point cloud is found to form a point cloud pair. The distance error between the two point clouds in the point cloud pair is calculated. If the distance error between the two point clouds in the point cloud pair is less than or equal to the error threshold or the number of iterations reaches the upper limit, the iteration stops; otherwise, the search for point cloud pairs continues until convergence.
[0083] like Figure 6 As shown, the local point cloud that has been determined to match in the existing tunnel point cloud is finely registered with the newly acquired electrical cloud to achieve registration between the newly acquired electrical cloud and the local point cloud.
[0084] The tunnel point cloud registration method provided in this embodiment converts the massive 3D tunnel point clouds in the original tunnel point cloud and the tunnel point cloud to be registered into a 2D flatness map that preserves geometric features, avoiding the computational bottleneck of directly processing massive point clouds. An image recognition-based coarse registration method replaces traditional direct point cloud registration calculations, significantly reducing computational complexity and solving the registration failure problem caused by tunnel geometric convergence. The long tunnel flatness map is segmented, preserving overlapping areas to ensure continuity and supporting segmented processing. The new and old point clouds use identical surface fitting parameters and segmentation methods to ensure feature comparability. Furthermore, this application combines image coarse registration and point cloud fine registration methods to further improve registration accuracy.
[0085] This embodiment also provides a tunnel point cloud registration device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.
[0086] This embodiment provides a tunnel point cloud registration device, such as... Figure 8 As shown, it includes: The acquisition module 801 is used to acquire the original tunnel point cloud and the tunnel point cloud to be registered for the target tunnel. The original tunnel point cloud is the complete point cloud of the target tunnel, and the tunnel point cloud to be registered is a local point cloud in the target tunnel.
[0087] The mapping module 802 is used to map the original tunnel point cloud and the tunnel point cloud to be registered onto a plane respectively, so as to obtain the original point cloud flatness map and the point cloud flatness map to be registered.
[0088] The first matching module 803 is used to match the point cloud flatness map to be registered with the original point cloud flatness map, and to determine the target point cloud flatness sub-map from the original point cloud flatness map. The target point cloud flatness sub-map is a local map in the original point cloud flatness map that matches the point cloud flatness map to be registered.
[0089] The second matching module 804 is used to determine the target tunnel point cloud that matches the tunnel point cloud to be registered from the original tunnel point cloud based on the target point cloud flatness sub-map.
[0090] The registration module 805 is used to perform fine registration between the point cloud of the tunnel to be registered and the point cloud of the target tunnel.
[0091] In some alternative implementations, the mapping module 802 includes: The acquisition unit is used to acquire the standard design profile of the target tunnel. The mapping unit is used to map the original tunnel point cloud and the tunnel point cloud to be registered onto the standard design profile of the tunnel, respectively, and calculate the distance between each point cloud and the standard design profile of the tunnel to obtain the original point cloud distance of each point cloud in the original tunnel point cloud and the point cloud distance to be registered of each point cloud in the tunnel point cloud to be registered. The unfolding unit is used to unfold the original tunnel point cloud and the tunnel point cloud to be registered into a plane according to the standard tunnel design outline, respectively, to obtain the original planar point cloud and the planar point cloud to be registered. The rendering unit is used to render the original point cloud distance of each point cloud in the original planar point cloud to generate an original point cloud flatness map; and to render the point cloud distance to be registered of each point cloud in the planar point cloud to be registered to generate a point cloud flatness map to be registered.
[0092] In some alternative implementations, the rendering unit includes: The first rendering subunit is used to calculate the first flatness value of each point cloud in the original planar point cloud based on the distance of the original point cloud; and to perform cloud map rendering on the first flatness value of each point cloud in the original planar point cloud to generate the original point cloud flatness map.
[0093] The second rendering subunit is used to calculate the second flatness value of each point cloud in the plane point cloud to be registered based on the distance of the point cloud to be registered; and to perform cloud map rendering on the second flatness value of each point cloud in the plane point cloud to be registered to generate a flatness map of the point cloud to be registered.
[0094] In some alternative implementations, the first matching module 803 includes: The cutting unit is used to cut the original point cloud flatness map into multiple original point cloud flatness sub-maps and mark the tunnel information of each original point cloud flatness sub-map. The tunnel information includes the tunnel name and tunnel station number.
[0095] The matching unit is used to match the point cloud flatness map to be registered with multiple original point cloud flatness sub-maps, and to determine the target point cloud flatness sub-map that matches the point cloud flatness map to be registered from the multiple original point cloud flatness sub-maps.
[0096] In some alternative implementations, the cutting unit includes: The first acquisition subunit is used to acquire preset cutting specifications and preset overlap specifications. The preset cutting specifications are the dimensions of each original point cloud flatness sub-image, and the preset overlap specifications are the overlap dimensions of adjacent original point cloud flatness sub-images.
[0097] The cutting sub-unit is used to cut the original point cloud flatness map into multiple original point cloud flatness sub-maps according to preset cutting specifications and preset overlap specifications.
[0098] In some alternative implementations, the matching unit includes: The second acquisition subunit is used to acquire tunnel information from the target point cloud flatness sub-map.
[0099] A sub-unit is determined to identify the local point cloud corresponding to the flatness sub-map of the target point cloud in the original tunnel point cloud based on tunnel information, thereby obtaining the target tunnel point cloud that matches the tunnel point cloud to be registered.
[0100] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.
[0101] In this embodiment, the tunnel point cloud registration device is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.
[0102] This invention also provides a computer device having the above-described features. Figure 8 The tunnel point cloud registration device shown.
[0103] Please see Figure 9 , Figure 9 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 9 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 9 Take a processor 10 as an example.
[0104] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GPA), or any combination thereof.
[0105] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.
[0106] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
[0107] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.
[0108] The computer device also includes an input device 30 and an output device 40. The processor 10, memory 20, input device 30, and output device 40 can be connected via a bus or other means. Figure 9 Taking the example of a connection between China and Israel via a bus.
[0109] Input device 30 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the computer device, such as a touchscreen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. Output device 40 may include display devices, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The aforementioned display devices include, but are not limited to, liquid crystal displays, light-emitting diodes, displays, and plasma displays. In some alternative embodiments, the display device may be a touchscreen.
[0110] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.
[0111] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.
[0112] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.
Claims
1. A method of tunnel point cloud registration, the method comprising: The method comprises: obtaining an original tunnel point cloud of a target tunnel and a to-be-registered tunnel point cloud, the original tunnel point cloud being a complete point cloud of the target tunnel, and the to-be-registered tunnel point cloud being a local point cloud in the target tunnel; mapping the original tunnel point cloud and the to-be-registered tunnel point cloud onto a plane respectively to obtain an original point cloud flatness graph and a to-be-registered point cloud flatness graph; matching the to-be-registered point cloud flatness graph with the original point cloud flatness graph, determining a target point cloud flatness subgraph from the original point cloud flatness graph, the target point cloud flatness subgraph being a local graph in the original point cloud flatness graph that matches the to-be-registered point cloud flatness graph; determining a target tunnel point cloud that matches the to-be-registered tunnel point cloud from the original tunnel point cloud based on the target point cloud flatness subgraph; precisely registering the to-be-registered tunnel point cloud with the target tunnel point cloud.
2. The tunnel point cloud registration method of claim 1, wherein, The method comprises: obtaining a tunnel standard design contour of the target tunnel; mapping the original tunnel point cloud and the to-be-registered tunnel point cloud onto the tunnel standard design contour respectively, and calculating the distance of each point cloud from the tunnel standard design contour to obtain an original point cloud distance of each point cloud in the original tunnel point cloud and a to-be-registered point cloud distance of each point cloud in the to-be-registered tunnel point cloud; unfolding the original tunnel point cloud and the to-be-registered tunnel point cloud according to the tunnel standard design contour into a plane to obtain an original plane point cloud and a to-be-registered plane point cloud; rendering the original point cloud distance of each point cloud in the original plane point cloud to generate the original point cloud flatness graph, and rendering the to-be-registered point cloud distance of each point cloud in the to-be-registered plane point cloud to generate the to-be-registered point cloud flatness graph.
3. The tunnel point cloud registration method of claim 2, wherein, The method comprises: calculating a first flatness value of each point cloud in the original plane point cloud based on the original point cloud distance; cloud map rendering the first flatness value of each point cloud in the original plane point cloud to generate the original point cloud flatness graph; The method comprises: calculating a second flatness value of each point cloud in the to-be-registered plane point cloud based on the to-be-registered point cloud distance; cloud map rendering the second flatness value of each point cloud in the to-be-registered plane point cloud to generate the to-be-registered point cloud flatness graph.
4. The tunnel point cloud registration method of claim 1, wherein, The method comprises: cutting the original point cloud flatness graph into a plurality of original point cloud flatness subgraphs, and marking tunnel information of each original point cloud flatness subgraph, the tunnel information comprising a tunnel name and a tunnel stake number. The to-be-registered point cloud flatness graph is matched with a plurality of the original point cloud flatness subgraphs, and a target point cloud flatness subgraph that matches the to-be-registered point cloud flatness graph is determined from the plurality of the original point cloud flatness subgraphs.
5. The tunnel point cloud registration method of claim 4, wherein, The original point cloud flatness graph is cut into a plurality of original point cloud flatness subgraphs, including: A preset cutting specification and a preset overlap specification are obtained, the preset cutting specification being a size of each of the original point cloud flatness subgraphs, and the preset overlap specification being an overlap size of adjacent original point cloud flatness subgraphs; The original point cloud flatness graph is cut into a plurality of original point cloud flatness subgraphs according to the preset cutting specification and the preset overlap specification.
6. The tunnel point cloud registration method of claim 4, wherein, The target tunnel point cloud that matches the to-be-registered tunnel point cloud is determined from the original tunnel point cloud based on the target point cloud flatness subgraph, including: The tunnel information of the target point cloud flatness subgraph is obtained; Based on the tunnel information, a local point cloud corresponding to the target point cloud flatness subgraph is determined in the original tunnel point cloud, to obtain the target tunnel point cloud that matches the to-be-registered tunnel point cloud.
7. A tunnel point cloud registration apparatus, characterized by, The device includes: An obtaining module is configured to obtain an original tunnel point cloud of a target tunnel and a to-be-registered tunnel point cloud, the original tunnel point cloud being a complete point cloud of the target tunnel, and the to-be-registered tunnel point cloud being a local point cloud in the target tunnel; A mapping module is configured to map the original tunnel point cloud and the to-be-registered tunnel point cloud to a plane respectively to obtain an original point cloud flatness graph and a to-be-registered point cloud flatness graph; A first matching module is configured to match the to-be-registered point cloud flatness graph with the original point cloud flatness graph to determine a target point cloud flatness subgraph from the original point cloud flatness graph, the target point cloud flatness subgraph being a local graph in the original point cloud flatness graph that matches the to-be-registered point cloud flatness graph; A second matching module is configured to determine a target tunnel point cloud that matches the to-be-registered tunnel point cloud from the original tunnel point cloud based on the target point cloud flatness subgraph; A registration module is configured to perform fine registration of the to-be-registered tunnel point cloud and the target tunnel point cloud.
8. A computer device, comprising: It includes: A memory and a processor are communicatively connected with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the tunnel point cloud registration method in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing a computer to perform the tunnel point cloud registration method in any one of claims 1 to 6.
10. A computer program product, characterised in that, It includes computer instructions for causing a computer to perform the tunnel point cloud registration method in any one of claims 1 to 6.