A device positioning method, device, system and medium based on point cloud matching

Abstract

The present invention discloses a device positioning method, device, system and medium based on point cloud matching, which is applied to the field of tunnel engineering technology. Aiming at the problem of low positioning accuracy, it collects a reference point cloud at a first position and a matching point cloud at a second position; obtains a corresponding reference normal map according to the reference point cloud, and obtains a corresponding matching normal map according to the matching point cloud; matches the reference normal map with the matching normal map to obtain an image matching relationship; obtains a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship; uses the point cloud matching relationship as the initial value of the ICP algorithm, and performs point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position; the present invention can make the obtained initial value more accurate, and perform point cloud registration based on this initial value, making the obtained relative pose relationship more accurate and improving the positioning accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel engineering, and particularly to a device positioning method, device, system and computer-readable storage medium based on point cloud matching. Background Art

[0002] During the construction of domestic tunnel equipment, it is necessary to position the engineering equipment to establish its relationship with the tunnel to meet the requirements of informatization and intelligentization of engineering equipment construction. For the traditional positioning technology used, surveyors need to use the positioning laser of a total station or a scanner to mark points for positioning. This method has a long positioning time, low work efficiency, and low automation level, and is not suitable for frequent mobile positioning during the construction process. At the same time, the positioning accuracy is easily affected by the subjective judgment and relevant experience of the surveyors. The surveyors are exposed to certain operation risks in the tunnel.

[0003] With the development of laser measurement and target positioning technologies, the most commonly used is the positioning technology based on point cloud matching. This technology usually collects adjacent front and rear point cloud data, and obtains the pose of point cloud matching positioning through front and rear point cloud matching, with few condition constraints and high flexibility. The current positioning technology based on point cloud matching mainly uses the point cloud information collected by a scanner to estimate the pose of an engineering equipment through point cloud matching. Among them, in order to obtain high-precision positioning information, it is first necessary to determine the pose relationship of point cloud matching, and the method is mainly to obtain the global optimal solution of point cloud matching through the iterative closest point (ICP) algorithm for point cloud matching.

[0004] However, the ICP algorithm is prone to falling into local maxima, and its initial value is highly dependent on the initial registration position. In the prior art, the initial value of the ICP algorithm is directly obtained after matching two point clouds, which is greatly affected by noise points and outliers, and the obtained initial value has low accuracy, affecting the accuracy of the point cloud matching pose and further affecting the positioning accuracy.

[0005] In view of this, how to provide a device positioning method, device, system and computer-readable storage medium based on point cloud matching to solve the above technical problems has become a problem that needs to be solved by those skilled in the art. Summary of the Invention

[0006] The purpose of the embodiments of the present invention is to provide a device positioning method, device, system and computer-readable storage medium based on point cloud matching, which can make the obtained relative pose relationship more accurate and improve the positioning accuracy during use.

[0007] To solve the above technical problems, the embodiments of the present invention provide a device positioning method based on point cloud matching, including:

[0008] Collecting the reference point cloud at the first position and the matching point cloud at the second position;

[0009] Obtain the corresponding reference normal map based on the reference point cloud, and obtain the corresponding matching normal map based on the matching point cloud;

[0010] Match the reference normal map with the matching normal map to obtain an image matching relationship;

[0011] Obtain the point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship;

[0012] Use the point cloud matching relationship as the initial value of the ICP algorithm, and perform point cloud registration using the ICP algorithm to obtain the relative pose relationship between the second position and the first position of the device.

[0013] Optionally, the obtaining the corresponding reference normal map based on the reference point cloud and obtaining the corresponding matching normal map based on the matching point cloud includes:

[0014] Perform horizontal calibration on the reference point cloud to obtain a horizontally calibrated reference point cloud, and perform horizontal calibration on the matching point cloud data to obtain a horizontally calibrated matching point cloud;

[0015] Process the horizontally calibrated reference point cloud to obtain a reference normal map, and process the horizontally calibrated matching point cloud to obtain a matching normal map.

[0016] Optionally, the processing the horizontally calibrated reference point cloud to obtain a reference normal map and processing the horizontally calibrated matching point cloud to obtain a matching normal map includes:

[0017] For each of the horizontally calibrated reference point cloud and the horizontally calibrated matching point cloud, calculate the center point and bounding box of the point cloud;

[0018] Perform grid division on the point cloud according to the center point in the xy plane, and project the point cloud into a grid two-dimensional map through the bounding box and the grid;

[0019] Obtain the normal vector of each grid in the grid two-dimensional map;

[0020] Calculate the angle between each of the normal vectors and the Z axis;

[0021] Convert the respective angles corresponding to the horizontally calibrated reference point cloud into gray values to obtain a reference normal map;

[0022] Convert the respective angles corresponding to the horizontally calibrated matching point cloud into gray values to obtain a matching normal map.

[0023] Optionally, the converting the respective angles corresponding to the horizontally calibrated reference point cloud into gray values to obtain a reference normal map includes:

[0024] Convert each included angle corresponding to the horizontal reference point cloud into a gray value to obtain an initial reference normal map;

[0025] Perform filtering processing on the initial reference normal map to obtain a final reference normal map;

[0026] The step of converting each included angle corresponding to the horizontal matching point cloud into a gray value to obtain a matching normal map includes:

[0027] Convert each included angle corresponding to the horizontal matching point cloud into a gray value to obtain an initial matching normal map;

[0028] Perform filtering processing on the initial matching normal map to obtain a final matching normal map.

[0029] Optionally, the step of obtaining the point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship includes:

[0030] According to the image matching relationship, obtain a first translation transformation from the center point of the reference normal map to the center point of the reference point cloud, a second translation transformation from the target point of the matching normal map to the target point of the matching point cloud, and a rotation transformation of the reference normal map relative to the matching normal map; wherein, the target point of the matching normal map is determined according to the center point of the reference normal map;

[0031] According to the target point position and the rotation angle, obtain the point cloud matching relationship between the reference point cloud and the matching point cloud.

[0032] Optionally, the step of performing point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position includes:

[0033] Calculate the closest point corresponding to each point in the current matching point cloud in the reference point cloud;

[0034] According to the current matching point cloud and each of the closest points, calculate a rigid body transformation that minimizes the average distance;

[0035] Based on the rigid body transformation, obtain a current translation matrix and a current rotation matrix;

[0036] Obtain a historical translation matrix and a historical rotation matrix according to the initial value, and update the matching point cloud through the historical translation matrix and the historical rotation matrix to obtain a new matching point cloud;

[0037] Based on the current translation matrix and the current rotation matrix, determine whether the new matching point cloud meets the requirements of the objective function. If so, end the iteration and obtain the relative pose relationship between the device at the second position and the first position according to the current translation matrix and the current rotation matrix; if not, use the current translation matrix as the historical translation matrix, use the current rotation matrix as the historical rotation matrix, use the new matching point cloud as the current matching point cloud, and return to execute the step of calculating the closest point corresponding to each point in the current matching point cloud in the reference point cloud.

[0038] An embodiment of the present invention further provides a device positioning device based on point cloud matching, including:

[0039] An acquisition module, configured to acquire a reference point cloud at a first position and a matching point cloud at a second position;

[0040] A processing module, configured to obtain a corresponding reference normal map according to the reference point cloud, and obtain a corresponding matching normal map according to the matching point cloud;

[0041] A matching module, configured to match the reference normal map with the matching normal map to obtain an image matching relationship;

[0042] A conversion module, configured to obtain a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship;

[0043] A calculation module, configured to use the point cloud matching relationship as the initial value of the ICP algorithm, and perform point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position.

[0044] Optionally, the processing module includes:

[0045] A calibration unit, configured to horizontally calibrate the reference point cloud to obtain a horizontally calibrated reference point cloud, and horizontally calibrate the matching point cloud data to obtain a horizontally calibrated matching point cloud;

[0046] A processing unit, configured to process the horizontally calibrated reference point cloud to obtain a reference normal map, and process the horizontally calibrated matching point cloud to obtain a matching normal map.

[0047] An embodiment of the present invention further provides a device positioning system based on point cloud matching, including:

[0048] A memory, configured to store a computer program;

[0049] A processor, configured to implement the steps of the device positioning method based on point cloud matching as described above when executing the computer program.

[0050] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the device positioning method based on point cloud matching as described above are implemented.

[0051] An embodiment of the present invention provides a device positioning method, device, system and computer-readable storage medium based on point cloud matching. The method includes: collecting a reference point cloud at a first position and a matching point cloud at a second position; obtaining a corresponding reference normal map according to the reference point cloud, and obtaining a corresponding matching normal map according to the matching point cloud; matching the reference normal map with the matching normal map to obtain an image matching relationship; obtaining a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship; using the point cloud matching relationship as the initial value of the ICP algorithm, and performing point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position.

[0052] It can be seen that the present invention collects a reference point cloud at a first position and a matching point cloud at a second position, obtains a reference normal map and a matching normal map, then obtains an image matching relationship according to the reference normal map and the matching normal map, and then obtains a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship, and uses it as the initial value of the ICP algorithm, which can make the obtained initial value more accurate, and perform point cloud registration based on this initial value, so that the obtained relative pose relationship is more accurate and the positioning accuracy is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the prior art and the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0054] Figure 1 It is a schematic flowchart of a device positioning method based on point cloud matching provided by an embodiment of the present invention;

[0055] Figure 2 It is a schematic diagram of a tunnel engineering device provided by an embodiment of the present invention;

[0056] Figure 3 It is a schematic diagram of point cloud scanning at two positions of a trolley provided by an embodiment of the present invention;

[0057] Figure 4 It is an effect diagram of point cloud horizontal calibration provided by an embodiment of the present invention;

[0058] Figure 5 It is a schematic diagram of point cloud data provided by an embodiment of the present invention;

[0059] Figure 6 A schematic diagram of the maximum Z value of point cloud data provided by an embodiment of the present invention;

[0060] Figure 7 A schematic diagram of the median Z value of point cloud data provided by an embodiment of the present invention;

[0061] Figure 8 A point cloud matching graph optimized by the ICP algorithm provided by an embodiment of the present invention;

[0062] Figure 9 A schematic structural diagram of a device positioning device based on point cloud matching provided by an embodiment of the present invention. Detailed implementation manners

[0063] The embodiments of the present invention provide a device positioning method, device, system and computer-readable storage medium based on point cloud matching, which can make the obtained relative pose relationship more accurate and improve the positioning accuracy during use.

[0064] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0065] Please refer to Figure 1 , Figure 1 A flowchart of a device positioning method based on point cloud matching provided by an embodiment of the present invention. The method includes:

[0066] S110: Collect the reference point cloud at the first position and the matching point cloud at the second position;

[0067] It should be noted that the first position in the embodiments of the present invention is a position that has been located during construction. The construction equipment moves from the first position to the second position, and the method provided by the present invention is used to determine the relative pose relationship between the equipment at the second position and the first position, so as to realize the positioning of the second position.

[0068] Specifically, the method in the embodiments of the present invention can be applied to tunnel engineering equipment, such as Figure 2The tunnel engineering equipment shown, where 1 represents the tunnel wall, 2 represents the trolley, 3 represents the on-vehicle computer, and 4 represents the scanner. After the trolley enters the working space, the position of the first position is determined by performing an initial positioning on the trolley. And after the trolley moves, the moved position is called the second position. The embodiments of the present invention take this scenario as an example for detailed description. Specifically, the trolley can be initially positioned after it enters the working space, and then the inner wall of the tunnel is scanned to obtain point cloud data. These point cloud data are used as the reference point cloud. And after the trolley moves, the moved position is determined as the second position, and the inner wall of the tunnel at the second position is scanned, and the obtained point cloud data is used as the matching point cloud. The schematic diagram of the point cloud scanning at the two positions of the trolley is as shown in Figure 3 shown, where t0 represents the moment at the first position and t1 represents the moment at the second position.

[0069] S120: Obtain the corresponding reference normal map according to the reference point cloud, and obtain the corresponding matching normal map according to the matching point cloud;

[0070] It should be noted that the reference point cloud in the embodiments of the present invention is point cloud data based on the tunnel coordinate system, and the matching point cloud is point cloud data based on the scanner. And since the Z-axis of the scanner coordinate system is generally not perpendicular to the actual ground, in order to project the point cloud onto the unified X-Y coordinate plane when generating the normal map subsequently, the matching point cloud data can be horizontally calibrated to obtain the horizontally matched point cloud, or the reference point cloud can be horizontally calibrated to obtain the horizontally reference point cloud, and the horizontally reference point cloud is processed to obtain the reference normal map, and the horizontally matched point cloud is processed to obtain the matching normal map. Among them, the effect diagram of the point cloud horizontal calibration is as shown in Figure 4 shown.

[0071] Specifically, in the embodiments of the present invention, the normal vector is determined according to the reference point cloud, and further the reference normal map corresponding to the reference point cloud is obtained according to the normal vector. The normal vector is determined according to the matching point cloud, and further the matching normal map is obtained according to the normal vector.

[0072] Further, the above-mentioned processing of the horizontally reference point cloud to obtain the reference normal map and the processing of the horizontally matched point cloud to obtain the matching normal map include:

[0073] For each point cloud in the horizontally reference point cloud and the horizontally matched point cloud, calculate the center point and the bounding box of the point cloud;

[0074] According to the center point, the point cloud is meshed in the xy plane, and the point cloud is projected onto the grid two-dimensional map through the bounding box and the grid;

[0075] Obtain the normal vector of each grid in the grid two-dimensional map;

[0076] Calculate the angle between each normal vector and the Z-axis respectively;

[0077] Convert each included angle corresponding to the horizontal reference point cloud into a grayscale value to obtain a reference normal map;

[0078] Convert each included angle corresponding to the horizontal matching point cloud into a grayscale value to obtain a matching normal map.

[0079] It should be noted that in the embodiments of the present invention

[0080] It should be noted that for each of the horizontal reference point cloud and the horizontal matching point cloud, calculate the center point and the bounding box of the point cloud, and perform grid division on the point cloud in the X-Y plane. Through the bounding box and the grid, the point cloud can be projected onto a grid two-dimensional map, and each grid corresponds to a sub-point cloud. For the sub-point cloud in each grid, the normal vector of each point in the sub-point cloud can be obtained, and each normal vector can be sorted. Based on the respective normal vectors, the normal vector of the network can be obtained. Specifically, the maximum normal vector can be selected as the normal vector of the grid, or the median value can be selected, or the average value of each normal vector can be calculated as the normal vector of the grid, so as to obtain the normal vector of each grid. Then, calculate the included angle between each normal vector and the Z-axis respectively, and convert each included angle into a grayscale value, so as to obtain the grayscale value corresponding to each grid respectively, and thus obtain the corresponding normal map. That is, through the above method, a reference normal map corresponding to the horizontal reference point cloud and a matching normal map corresponding to the horizontal matching point cloud can be obtained. Among them, the point cloud data is as Figure 5 shown, the maximum Z value is as Figure 6 shown, and the Z median value is as Figure 7 shown.

[0081] Furthermore, in order to make the obtained normal map more accurate, in the embodiments of the present invention, each included angle corresponding to the horizontal reference point cloud can be converted into a grayscale value to obtain an initial reference normal map, and then the initial reference normal map can be filtered to obtain a final reference normal map; each included angle corresponding to the horizontal matching point cloud can be converted into a grayscale value to obtain an initial matching normal map, and then the initial matching normal map can be filtered to obtain a final matching normal map.

[0082] S130: Match the reference normal map and the matching normal map to obtain an image matching relationship;

[0083] Specifically, in the embodiments of the present invention, after obtaining the reference normal map and the matching normal map, the reference normal map and the matching normal map are matched. Specifically, the optimal matching parameters of the reference normal map and the matching normal map can be found by calculating NCC (normalized cross correlation), so as to obtain the image matching relationship between the two.

[0084] Among them, L(u, v) represents the pixel value of the reference normal map at the pixel point (u, v). represents the average value of the pixels within the window of the reference normal map. R(u + r, v + c) represents the pixel value of the reference normal map after the position offset corresponding to the point on the matching normal map. represents the pixel value of the matching normal map at the pixel point (r, c).

[0085] S140: Obtain the point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship.

[0086] Specifically, in the embodiments of the present invention, according to the image matching relationship, the first translation transformation T1 from the center point of the reference normal map to the center point of the reference point cloud, the second translation transformation T2 from the target point of the matching normal map to the target point of the matching point cloud, and the rotation transformation R' of the reference normal map relative to the matching normal map can be obtained. Among them, the target point of the matching normal map is determined according to the center point of the reference normal map. Then, according to the target point position and the rotation angle, the point cloud matching relationship between the reference point cloud and the matching point cloud is obtained. Specifically, the point cloud matching relationship can be obtained according to T1×R'×T2, that is, the initial value of the point cloud change is obtained, so that the obtained initial value can be more accurate.

[0087] S150: Use the point cloud matching relationship as the initial value of the ICP algorithm, and perform point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position.

[0088] It can be understood that using the obtained point cloud matching relationship as the initial value of the ICP algorithm, and further performing point cloud registration using the ICP algorithm to obtain a more accurate relative pose relationship between the device at the second position and the first position. In practical applications, during the movement of the trolley in the tunnel, the relative pose can be obtained through point cloud matching to achieve the dynamic positioning of the tunnel equipment point cloud matching trolley positioning, solving problems such as long repeated positioning time and low work efficiency of the tunnel equipment.

[0089] Specifically, in S150, the relative position relationship can be obtained through the following method:

[0090] For each point in the reference point cloud, determine the point with the closest distance corresponding to this point from the current matching point cloud. Specifically, determine the point with the smallest Euclidean distance as the corresponding point with the closest distance.

[0091] According to the current matching point cloud and each point with the closest distance, calculate the rigid body transformation that minimizes the average distance.

[0092] Based on the rigid body transformation, obtain the current translation matrix and the current rotation matrix.

[0093] Obtain the historical translation matrix and the historical rotation matrix according to the initial values, and update the matching point cloud through the historical translation matrix and the historical rotation matrix to obtain a new matching point cloud;

[0094] Based on the current translation matrix and the current rotation matrix, determine whether the new matching point cloud meets the requirements of the objective function. If so, end the iteration, and obtain the relative pose relationship between the device at the second position and the first position according to the current translation matrix and the current rotation matrix; if not, use the current translation matrix as the historical translation matrix, the current rotation matrix as the historical rotation matrix, and the new matching point cloud as the current matching point cloud, and return to execute the step of calculating the closest point corresponding to each point in the current matching point cloud in the reference point cloud until the requirements of the objective function are met, thereby obtaining an accurate pose relationship. Figure 8 It is the point cloud matching graph optimized by the ICP algorithm.

[0095] It should be noted that the Iterative Closest Point algorithm ICP is based on the idea of iterative optimization, uses the spatial distance as the basis for selecting matching points, and continuously adjusts the pose of the point cloud to minimize the cumulative distance between the matching points. Assume that Q i (i = 1, 2, 3,...) represents the reference point cloud, and P i (i = 1, 2, 3,...) represents the matching point cloud. The alignment registration of the two point sets is converted to minimize the following objective function:

[0096] R represents the rotation matrix, T represents the translation matrix, Q i represents the i-th point in the reference point cloud, and P i represents the i-th point in the matching point cloud, and n represents the number of points in the reference point cloud.

[0097] It can be seen that the present invention collects the reference point cloud at the first position and the matching point cloud at the second position, obtains the reference normal map and the matching normal map, then obtains the image matching relationship according to the reference normal map and the matching normal map, and then obtains the point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship, and uses it as the initial value of the ICP algorithm, which can make the obtained initial value more accurate, and perform point cloud registration based on this initial value, making the obtained relative pose relationship more accurate and improving the positioning accuracy.

[0098] On the basis of the above embodiments, the embodiments of the present invention further provide a device positioning device based on point cloud matching. Please refer to Figure 9 This device includes:

[0099] An acquisition module 21, configured to acquire a reference point cloud at the first position and a matching point cloud at the second position;

[0100] A processing module 22, configured to obtain a corresponding reference normal map based on the reference point cloud, and obtain a corresponding matching normal map based on the matching point cloud;

[0101] A matching module 23, configured to match the reference normal map with the matching normal map to obtain an image matching relationship;

[0102] A conversion module 24, configured to obtain a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship;

[0103] A calculation module 25, configured to use the point cloud matching relationship as the initial value of the ICP algorithm, and perform point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position.

[0104] Further, the above-mentioned processing module 22 includes:

[0105] A calibration unit, configured to horizontally calibrate the reference point cloud to obtain a horizontal reference point cloud, and horizontally calibrate the matching point cloud data to obtain a horizontal matching point cloud;

[0106] A processing unit, configured to process the horizontal reference point cloud to obtain a reference normal map, and process the horizontal matching point cloud to obtain a matching normal map.

[0107] It should be noted that the device positioning device based on point cloud matching provided in the embodiments of the present invention has the same beneficial effects as the device positioning method based on point cloud matching provided in the above embodiments. For the specific introduction of the device positioning method based on point cloud matching involved in the embodiments of the present invention, please refer to the above embodiments, and the present invention will not be elaborated herein.

[0108] Based on the above embodiments, an embodiment of the present invention further provides a device positioning system based on point cloud matching, and the system includes:

[0109] A memory, configured to store a computer program;

[0110] A processor, configured to implement the steps of the device positioning method based on point cloud matching as described above when executing the computer program.

[0111] For example, the processor in the embodiments of the present invention can specifically be configured to implement collecting a reference point cloud at the first position and a matching point cloud at the second position; obtaining a corresponding reference normal map based on the reference point cloud, and obtaining a corresponding matching normal map based on the matching point cloud; matching the reference normal map with the matching normal map to obtain an image matching relationship; obtaining a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship; using the point cloud matching relationship as the initial value of the ICP algorithm, and performing point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position.

[0112] Based on the above embodiments, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the device positioning method based on point cloud matching as described above are implemented.

[0113] The computer-readable storage medium may include: various media such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs that can store program codes.

[0114] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0115] It should also be noted that in this specification, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0116] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A device positioning method based on point cloud matching, characterized in that Including: Collecting the reference point cloud at the first position and the matching point cloud at the second position; Obtaining the corresponding reference normal map according to the reference point cloud, and obtaining the corresponding matching normal map according to the matching point cloud; Matching the reference normal map with the matching normal map to obtain an image matching relationship; Obtaining a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship; Taking the point cloud matching relationship as the initial value of the ICP algorithm, and performing point cloud registration using the ICP algorithm to obtain the relative pose relationship between the device at the second position and the first position; where: The obtaining the point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship includes: According to the image matching relationship, obtaining a first translation transformation from the center point of the reference normal map to the center point of the reference point cloud, a second translation transformation from the target point of the matching normal map to the target point of the matching point cloud, and a rotation transformation of the reference normal map relative to the matching normal map; wherein, the target point of the matching normal map is determined according to the center point of the reference normal map; Obtaining a point cloud matching relationship between the reference point cloud and the matching point cloud according to the first translation transformation, the second translation transformation and the rotation transformation.

2. The device positioning method based on point cloud matching according to claim 1, wherein The obtaining the corresponding reference normal map according to the reference point cloud, and obtaining the corresponding matching normal map according to the matching point cloud includes: Horizontally calibrating the reference point cloud to obtain a horizontally reference point cloud, and horizontally calibrating the matching point cloud data to obtain a horizontally matching point cloud; Processing the horizontally reference point cloud to obtain a reference normal map, and processing the horizontally matching point cloud to obtain a matching normal map.

3. The device positioning method based on point cloud matching according to claim 2, wherein The processing the horizontally reference point cloud to obtain a reference normal map, and processing the horizontally matching point cloud to obtain a matching normal map includes: For each point cloud in the horizontally reference point cloud and the horizontally matching point cloud, calculating the center point and the bounding box of the point cloud; According to the center point, dividing the point cloud in the xy plane into grids, and projecting the point cloud into a grid two-dimensional map through the bounding box and the grids; Obtaining the normal vector of each grid in the grid two-dimensional map; Calculating the angle between each normal vector and the Z axis; Converting the respective angles corresponding to the horizontally reference point cloud into gray values to obtain a reference normal map; Converting the respective angles corresponding to the horizontally matching point cloud into gray values to obtain a matching normal map.

4. The device positioning method based on point cloud matching according to claim 3, characterized in that The converting the respective angles corresponding to the horizontally reference point cloud into gray values to obtain a reference normal map includes: Converting the respective angles corresponding to the horizontally reference point cloud into gray values to obtain an initial reference normal map; Performing filtering processing on the initial reference normal map to obtain a final reference normal map; The converting the respective angles corresponding to the horizontally matching point cloud into gray values to obtain a matching normal map includes: Converting the respective angles corresponding to the horizontally matching point cloud into gray values to obtain an initial matching normal map; Performing filtering processing on the initial matching normal map to obtain a final matching normal map.

5. The device positioning method based on point cloud matching according to claim 3, characterized in that, Performing point cloud registration using the ICP algorithm to obtain the relative pose relationship between the second position and the first position of the device includes: Calculating the point with the closest distance corresponding to each point in the current matching point cloud in the reference point cloud; Calculating a rigid body transformation that minimizes the average distance based on the current matching point cloud and each of the closest distance points; Based on the rigid body transformation, obtaining the current translation matrix and the current rotation matrix; Obtaining the historical translation matrix and the historical rotation matrix according to the initial value, and updating the matching point cloud through the historical translation matrix and the historical rotation matrix to obtain a new matching point cloud; Based on the current translation matrix and the current rotation matrix, determining whether the new matching point cloud meets the requirements of the objective function. If so, ending the iteration, and obtaining the relative pose relationship between the second position and the first position of the device according to the current translation matrix and the current rotation matrix; if not, using the current translation matrix as the historical translation matrix, using the current rotation matrix as the historical rotation matrix, using the new matching point cloud as the current matching point cloud, and returning to execute the step of calculating the point with the closest distance corresponding to each point in the current matching point cloud in the reference point cloud.

6. A device positioning device based on point cloud matching, characterized in that, Including: An acquisition module for acquiring a reference point cloud at a first position and a matching point cloud at a second position; A processing module for obtaining a corresponding reference normal map according to the reference point cloud and obtaining a corresponding matching normal map according to the matching point cloud; A matching module for matching the reference normal map with the matching normal map to obtain an image matching relationship; A conversion module for obtaining a point cloud matching relationship between the reference point cloud and the matching point cloud according to the image matching relationship; A calculation module for using the point cloud matching relationship as the initial value of the ICP algorithm and performing point cloud registration using the ICP algorithm to obtain the relative pose relationship between the second position and the first position of the device; wherein: The matching module is used for: According to the image matching relationship, obtaining a first translation transformation from the center point of the reference normal map to the center point of the reference point cloud, a second translation transformation from the target point of the matching normal map to the target point of the matching point cloud, and a rotation transformation of the reference normal map relative to the matching normal map; wherein, the target point of the matching normal map is determined according to the center point of the reference normal map; according to the first translation transformation, the second translation transformation and the rotation transformation, obtaining the point cloud matching relationship between the reference point cloud and the matching point cloud.

7. The device positioning device based on point cloud matching according to claim 6, characterized in that, The processing module includes: A calibration unit for horizontally calibrating the reference point cloud to obtain a horizontally calibrated reference point cloud and horizontally calibrating the matching point cloud data to obtain a horizontally calibrated matching point cloud; A processing unit for processing the horizontally calibrated reference point cloud to obtain a reference normal map and processing the horizontally calibrated matching point cloud to obtain a matching normal map.

8. A device positioning system based on point cloud matching, characterized in that Including: A memory for storing computer programs; A processor for implementing the steps of the device positioning method based on point cloud matching according to any one of claims 1 to 5 when executing the computer program.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the device positioning method based on point cloud matching according to any one of claims 1 to 5 are implemented.

Images