Point cloud data registration method and electronic equipment

Abstract

The invention discloses a point cloud data registration method and electronic equipment. The method comprises the following steps: acquiring first point cloud data and second point cloud data; receiving a first selection instruction, and determining a plurality of first feature points in the first point cloud data according to the first selection instruction; receiving a second selection instruction, and determining a plurality of second feature points in the second point cloud data according to the second selection instruction; determining a plurality of first registration sub-domains based on the plurality of first feature points; determining a plurality of second registration sub-domains based on the plurality of second feature points; registering each second registration sub-domain with the corresponding first registration sub-domain to obtain a first registration matrix of each second registration sub-domain; and registering the second point cloud data with the first point cloud data based on the first registration matrixes of all the second registration sub-domains. According to the method, the second point cloud data and the first point cloud data are registered through the first registration matrix based on all the second registration sub-domains, and very high registration precision can still be realized without marking points.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and particularly to a point cloud data registration method and an electronic device. Background Art

[0002] In today's medical field, during neurosurgery, in order to improve the success rate of the surgery, doctors often need to perform surgical navigation. Surgical navigation refers to accurately matching the image data of the target object obtained before the surgery with the actual anatomical structure of the target object at the surgical site, so that doctors can perform the surgery accurately according to the surgical navigation information.

[0003] Currently, the point cloud data registration methods in neurosurgery mainly include the method using fiducial points and the global point cloud data registration method. The method using fiducial points requires pasting or implanting fiducial points on the face of the target object, and then tracking the fiducial points through optical devices or electromagnetic devices during the surgery for registration. The global point cloud data registration method uses a structured light camera to perform global point cloud data acquisition and registration on the face of the target object.

[0004] However, the method using fiducial points is susceptible to factors such as the displacement of the target object during the surgery, environmental humidity, and skin oil, resulting in displacement or shedding, thus leading to inaccurate registration and reducing the safety of the surgery. In addition, implanting fiducial points will also increase the risk of infection of the target object. The global point cloud data registration method is sensitive to noise and outliers, and various interference factors in the surgical environment may cause inaccurate registration when using the global point cloud data registration method, thereby reducing the safety of the surgery. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present application provides a point cloud data registration method and an electronic device. By registering the second point cloud data with the first point cloud data based on the first registration matrix of all second registration sub-domains, high registration accuracy can still be achieved without pasting or implanting fiducial points on the target object, thereby reducing the dependence on fiducial points during the surgery, reducing the risk of infection of the target object, and reducing the risk of displacement and shedding of fiducial points during the surgery, improving the safety of the surgery.

[0006] To solve the above problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present application provides a point cloud data registration method, including: obtaining first point cloud data and second point cloud data, where the first point cloud data is obtained by scanning the target object using a first scanning method, and the second point cloud data is obtained by scanning the target object using a second scanning method, and the first scanning method is different from the second scanning method;

[0008] Receive a first selection instruction, and determine a plurality of first feature points in the first point cloud data according to the first selection instruction;

[0009] Receive a second selection instruction, and determine a plurality of second feature points in the second point cloud data according to the second selection instruction, where the second feature points correspond to the first feature points one by one;

[0010] Determine a plurality of first registration sub-domains based on the plurality of first feature points;

[0011] Determine a plurality of second registration sub-domains based on the plurality of second feature points, where the second registration sub-domains correspond to the first registration sub-domains one by one;

[0012] Register each of the second registration sub-domains with the corresponding first registration sub-domain to obtain a first registration matrix for each second registration sub-domain;

[0013] Register the second point cloud data with the first point cloud data based on the first registration matrices of all the second registration sub-domains.

[0014] In some embodiments, the determining a plurality of first registration sub-domains based on the plurality of first feature points includes:

[0015] Receive a selection instruction, and select at least one target sub-domain determination method from multiple sub-domain determination methods according to the selection instruction;

[0016] For each of the target sub-domain determination methods, use the target sub-domain determination method to determine a plurality of the first registration sub-domains based on the plurality of first feature points, so as to obtain all the first registration sub-domains;

[0017] The determining a plurality of second registration sub-domains based on the plurality of second feature points includes:

[0018] For each of the target sub-domain determination methods, use the target sub-domain determination method to determine a plurality of the second registration sub-domains based on the plurality of second feature points, so as to obtain all the second registration sub-domains.

[0019] In some embodiments, the multiple sub-domain determination methods include a sphere sub-domain determination method, a cube sub-domain determination method, and an oblique prism sub-domain determination method. The plurality of first feature points include a plurality of first feature point pairs, and the plurality of second feature points include a plurality of second feature point pairs.

[0020] The for each of the target sub-domain determination methods, use the target sub-domain determination method to determine a plurality of the first registration sub-domains based on the plurality of first feature points, so as to obtain all the first registration sub-domains, includes:

[0021] When the target sub-domain determination method is the sphere sub-domain determination method, based on the coordinates of each of the first feature points and a preset radius, a first registration sub-domain with the first feature point as the center of the sphere and the preset radius as the radius of the sphere is determined, so as to obtain a plurality of the first registration sub-domains determined by using the sphere sub-domain determination method;

[0022] When the target sub-domain determination method is the cube sub-domain determination method, based on the coordinates of each of the first feature points and a preset cube side length, a first registration sub-domain with the first feature point as the center of the cube and the preset cube side length as the side length of the cube is determined, so as to obtain a plurality of the first registration sub-domains determined by using the cube sub-domain determination method;

[0023] When the target sub-domain determination method is the oblique prism sub-domain determination method, based on the coordinates of two of the first feature points in each pair of the first feature points and a preset distance value, a first registration sub-domain including the pair of the first feature points is determined, so as to obtain a plurality of the first registration sub-domains determined by using the oblique prism sub-domain determination method;

[0024] For each of the target sub-domain determination methods, by using the target sub-domain determination method, a plurality of the second registration sub-domains are determined based on the plurality of second feature points, so as to obtain all the second registration sub-domains, including:

[0025] When the target sub-domain determination method is the sphere sub-domain determination method, based on the coordinates of each of the second feature points and the preset radius, a second registration sub-domain with the second feature point as the center of the sphere and the preset radius as the radius of the sphere is determined, so as to obtain a plurality of the second registration sub-domains determined by using the sphere sub-domain determination method;

[0026] When the target sub-domain determination method is the cube sub-domain determination method, based on the coordinates of each of the second feature points and the preset cube side length, a second registration sub-domain with the second feature point as the center of the cube and the preset cube side length as the side length of the cube is determined, so as to obtain a plurality of the second registration sub-domains determined by using the cube sub-domain determination method;

[0027] When the target sub-domain determination method is the oblique prism sub-domain determination method, based on the coordinates of two of the second feature points in each pair of the second feature points and the preset distance value, a second registration sub-domain including the pair of the second feature points is determined, so as to obtain a plurality of the second registration sub-domains determined by using the oblique prism sub-domain determination method.

[0028] In some embodiments, before determining the plurality of first registration sub-domains based on the plurality of first feature points, the method further includes:

[0029] Calculate a second registration matrix for the second point cloud data based on each of the first feature points and the corresponding second feature points;

[0030] Left-multiply the coordinates of each point in the second point cloud data by the second registration matrix, so as to perform a first registration of the second point cloud data with the first point cloud data;

[0031] The registering of the second point cloud data with the first point cloud data based on the first registration matrices of all the second registration sub-domains includes:

[0032] Perform weighted fusion on all the first registration matrices to obtain a third registration matrix;

[0033] Left-multiply the coordinates of each point in the second point cloud data by the third registration matrix, so as to perform a second registration of the second point cloud data with the first point cloud data.

[0034] In some embodiments, the performing weighted fusion on all the first registration matrices to obtain a third registration matrix includes:

[0035] Calculate the registration error between each second registration sub-domain and the corresponding first registration sub-domain;

[0036] Calculate the weight of each second registration sub-domain based on the registration error of each second registration sub-domain and the number of points in the second point cloud data included in the second registration sub-domain;

[0037] Perform weighted fusion on all the first registration matrices based on the weight of each second registration sub-domain and the first registration matrix of the second registration sub-domain to obtain the third registration matrix.

[0038] In some embodiments, the calculating the weight of each second registration sub-domain based on the registration error of each second registration sub-domain and the number of points in the second point cloud data included in the second registration sub-domain includes:

[0039] Calculate the initial weight of each second registration sub-domain based on the registration error of each second registration sub-domain and the number of points in the second point cloud data included in the second registration sub-domain;

[0040] Calculate the normalized weight of each second registration sub-domain based on the initial weight of each second registration sub-domain and the sum of the initial weights of all the second registration sub-domains.

[0041] In some embodiments, the calculating the initial weight of each second registration sub-domain based on the registration error of each second registration sub-domain and the number of points in the second point cloud data included in the second registration sub-domain includes:

[0042] For each of the second registration sub-domains, calculate a first weight calculation parameter as the exponent with the opposite of the value obtained by dividing the registration error of the second registration sub-domain by a preset first error weight as the preset base number;

[0043] Calculate a second weight calculation parameter as the exponent with the opposite of the value obtained by dividing the number of points in the second point cloud data included in the second registration sub-domain by a preset second error weight as the preset base number;

[0044] Multiply the first weight calculation parameter and the second weight calculation parameter to obtain the initial weight of the second registration sub-domain.

[0045] In some embodiments, when the multiple first registration sub-domains and the multiple second registration sub-domains are determined by multiple target sub-domain determination methods, each target sub-domain determination method corresponds to a preset sub-domain weight.

[0046] The calculating of the normalized weight of each second registration sub-domain based on the initial weight of each second registration sub-domain and the sum of the initial weights of all second registration sub-domains includes:

[0047] For each second registration sub-domain, divide the initial weight of the second registration sub-domain by the sum of the initial weights of all second registration sub-domains to obtain a first calculated value;

[0048] Divide the sub-domain weight of the target sub-domain determination method corresponding to the second registration sub-domain by the sum of the sub-domain weights of all target sub-domain determination methods to obtain a second calculated value;

[0049] Multiply the first calculated value by the second calculated value to obtain the normalized weight of the second registration sub-domain.

[0050] In some embodiments, the weighted fusion of all the first registration matrices based on the weight of each second registration sub-domain and the first registration matrix of the second registration sub-domain to obtain the third registration matrix includes:

[0051] Convert each first registration matrix into the Lie algebra space to obtain the corresponding first registration matrix element;

[0052] Based on the weight of each second registration sub-domain and the first registration matrix element corresponding to the second registration sub-domain, perform weighted fusion on all the first registration matrix elements to obtain the weighted-fused first registration matrix element;

[0053] Convert the weighted-fused first registration matrix element into the third registration matrix in the Euclidean space.

[0054] Second aspect, an embodiment of the present application provides an electronic device, and the electronic device includes:

[0055] At least one processor; and,

[0056] A memory communicatively connected to the at least one processor; wherein,

[0057] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the point cloud data registration method as described in the first aspect.

[0058] Third aspect, an embodiment of the present application provides a computer-readable storage medium, and the computer-readable storage medium stores an executable program, and the executable program is executed by a processor to implement the point cloud data registration method as described in the first aspect.

[0059] The present application provides a point cloud data registration method and an electronic device. The present application registers the second point cloud data with the first point cloud data based on the first registration matrix of all the second registration sub-domains, and can still achieve a very high registration accuracy without pasting or implanting marker points on the target object, thereby reducing the dependence on marker points during the operation, reducing the risk of infection of the target object, and reducing the risk of displacement and detachment of the marker points during the operation, improving the safety of the operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a schematic flowchart of the first implementation manner of the point cloud data registration method provided by the embodiment of the present application.

[0061] Figure 2A is a schematic diagram of a plurality of first feature points provided by the embodiment of the present application.

[0062] Figure 2B is a schematic diagram of a plurality of first registration sub-domains determined by using the sphere sub-domain determination method provided by the embodiment of the present application.

[0063] Figure 2C is a schematic diagram of a plurality of first registration sub-domains determined by using the cube sub-domain determination method provided by the embodiment of the present application.

[0064] Figure 2D is a schematic diagram of a plurality of first registration sub-domains determined by using the oblique prism sub-domain determination method provided by the embodiment of the present application.

[0065] Figure 3 is a schematic flowchart of the second implementation manner of the point cloud data registration method provided by the embodiment of the present application.

[0066] Figure 4AIt is a schematic diagram of the first reconstruction model provided by an embodiment of the present application.

[0067] Figure 4B It is a schematic diagram of the second point cloud data provided by an embodiment of the present application.

[0068] Figure 4C It is a schematic diagram of the result after the first registration of the second point cloud data and the first point cloud data.

[0069] Figure 4D It is a schematic diagram of the result after the second registration of the second point cloud data and the first point cloud data.

[0070] Figure 5 It is a schematic diagram of the structure of the point cloud data registration device provided by an embodiment of the present application.

[0071] Figure 6 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application.

[0072] Figure 7 It is a structural block diagram of a computer-readable storage medium provided by an embodiment of the present application. Detailed implementation manners

[0073] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0074] In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0075] The present application provides a point cloud data registration method and an electronic device. By registering the second point cloud data and the first point cloud data based on the first registration matrix of all the second registration sub-domains, high registration accuracy can still be achieved without pasting or implanting marker points on the target object, thereby reducing the dependence on marker points during the operation, reducing the risk of infection of the target object, and reducing the risk of displacement and falling off of the marker points during the operation, improving the safety of the operation.

[0076] Next, the point cloud data registration method provided by the present application will be specifically described in conjunction with the accompanying drawings.

[0077] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of the first implementation manner of the point cloud data registration method provided by an embodiment of the present application. As Figure 1 shown, the point cloud data registration method includes: step S100 to step S500.

[0078] Step S100: Obtain first point cloud data and second point cloud data.

[0079] Among them, the first point cloud data is obtained by scanning a target object using a first scanning method, and the second point cloud data is obtained by scanning the target object using a second scanning method. The first scanning method is different from the second scanning method.

[0080] Optionally, the accuracy of the first point cloud data is higher than that of the second point cloud data.

[0081] Optionally, the first scanning method is to scan a target part of the target object using a three-dimensional MRI (Magnetic Resonance Imaging) scanner.

[0082] Optionally, the second scanning method is to scan a target part of the target object using a structured light camera.

[0083] Optionally, the target part is the face.

[0084] In some implementation manners, after obtaining the initial first point cloud data, data preprocessing is further performed on the initial first point cloud data to obtain preprocessed first point cloud data.

[0085] In some implementation manners, after obtaining the initial second point cloud data, data preprocessing is further performed on the initial second point cloud data to obtain preprocessed second point cloud data.

[0086] Optionally, the data preprocessing includes at least one operation of noise removal, gray normalization, image enhancement, facial tissue segmentation, and surface reconstruction.

[0087] In some implementation manners, the first point cloud data and the second point cloud data processed in subsequent steps are both preprocessed.

[0088] In some implementation manners, the first point cloud data and the second point cloud data processed in subsequent steps may not be preprocessed either.

[0089] Step S200: Receive a first selection instruction, and determine multiple first feature points in the first point cloud data according to the first selection instruction.

[0090] In some embodiments, surface reconstruction is performed based on the first point cloud data to obtain a first reconstruction model, and a plurality of first feature points in the first reconstruction model are determined according to a first selection instruction. The plurality of first feature points in the first reconstruction model are also the plurality of first feature points in the first point cloud data.

[0091] Optionally, the first feature points are points with prominent features on the face of the target object, such as the corner points of the eyes and the wing points of the nose.

[0092] Optionally, the number of the first feature points is more than 2, for example, the number of the first feature points is 2, 3, 4, 5, 6, 10, etc.

[0093] Please refer to Figure 2A , Figure 2A which is a schematic diagram of a plurality of first feature points provided by an embodiment of the present application. As Figure 2A shown, in some embodiments, surface reconstruction is performed based on the first point cloud data to obtain a first reconstruction model 10, and the first feature point 11, the first feature point 12, the first feature point 13, the first feature point 14, the first feature point 15, and the first feature point 16 in the first reconstruction model 10 are determined according to the first selection instruction.

[0094] Optionally, the plurality of first feature points include a plurality of first feature point pairs.

[0095] Optionally, the two first feature points in the first feature point pair are located on the same organ of the target object.

[0096] As Figure 2A shown, optionally, the first feature point 11 and the first feature point 12 are the two corner points of the right eye of the target object, and the first feature point 11 and the first feature point 12 are a first feature point pair. The first feature point 13 and the first feature point 14 are the two corner points of the left eye of the target object, and the first feature point 13 and the first feature point 14 are a first feature point pair. The first feature point 15 and the first feature point 16 are the two wing points of the target object, and the first feature point 15 and the first feature point 16 are a first feature point pair.

[0097] In some other embodiments, a plurality of first feature points are selected from the first point cloud data according to the first selection instruction. The selection method of the plurality of first feature points is as described above.

[0098] Step S300: Receive a second selection instruction, and determine a plurality of second feature points in the second point cloud data according to the second selection instruction.

[0099] Wherein, the second feature points correspond to the first feature points one by one. The positions of the second feature points and the first feature points in the point cloud data are the same or approximately the same.

[0100] The method for determining a plurality of second feature points in the second point cloud data according to the second selection instruction refers to the method for determining a plurality of first feature points in the first point cloud data in step S200.

[0101] Optionally, the plurality of second feature points include a plurality of second feature point pairs. The method for determining the plurality of second feature point pairs refers to the method for determining the plurality of first feature point pairs in step S200.

[0102] Step S400: Determine a plurality of first registration subdomains based on the plurality of first feature points.

[0103] In some embodiments, step S400 includes steps S410 to S420.

[0104] Step S410: Receive a selection instruction, and select at least one target subdomain determination method from a plurality of subdomain determination methods according to the selection instruction.

[0105] In some embodiments, the plurality of subdomain determination methods include a sphere subdomain determination method, a cube subdomain determination method, and an oblique prism subdomain determination method.

[0106] Step S420: For each target subdomain determination method, use the target subdomain determination method to determine a plurality of first registration subdomains based on the plurality of first feature points, so as to obtain all the first registration subdomains.

[0107] In some embodiments, when there are multiple target subdomain determination methods, sequentially use one target subdomain determination method among the multiple target subdomain determination methods to determine a plurality of first registration subdomains based on the plurality of first feature points. When all the target subdomain determination methods have been executed, all the first registration subdomains are obtained.

[0108] In some embodiments, when the target subdomain determination method is the sphere subdomain determination method, determine a first registration subdomain with the first feature point as the center of the sphere and the preset radius as the radius of the sphere based on the coordinates of each first feature point and the preset radius, so as to obtain a plurality of first registration subdomains determined by the sphere subdomain determination method.

[0109] The target subdomain determination method corresponding to the first registration subdomain determined by the sphere subdomain determination method is the sphere subdomain determination method.

[0110] In some embodiments, when the target subdomain determination method is the sphere subdomain determination method, the calculation formula for the first registration subdomain is:

[0111]

[0112] where Ri represents the i-th first registration subdomain, m represents a point in Ri, and m is a point in the first point cloud data. It represents a three-dimensional space, ti represents the i-th first feature point, r represents a preset radius, and 2 represents the Euclidean norm.

[0113] Please refer to Figure 2B , Figure 2B FIG. is a schematic diagram of a plurality of first registration sub-domains determined by the sphere sub-domain determination method provided in an embodiment of the present application. As Figure 2B shown, exemplarily, based on the coordinates of each first feature point and the preset radius, a first registration sub-domain with the first feature point as the center point of the sphere and the preset radius as the radius of the sphere is determined, so as to obtain a plurality of first registration sub-domains A1 determined by the sphere sub-domain determination method. At this time, the first registration sub-domain A1 is a sphere sub-domain.

[0114] In some embodiments, when the target sub-domain determination method is the cube sub-domain determination method, based on the coordinates of each first feature point and the preset cube side length, a first registration sub-domain with the first feature point as the center point of the cube and the preset cube side length as the side length of the cube is determined, so as to obtain a plurality of first registration sub-domains determined by the cube sub-domain determination method.

[0115] The target sub-domain determination method corresponding to the first registration sub-domain determined by the cube sub-domain determination method is the cube sub-domain determination method.

[0116] In some embodiments, when the target sub-domain determination method is the cube sub-domain determination method, the calculation formula for the first registration sub-domain is:

[0117]

[0118] where l represents the preset cube side length, and ∞ represents the maximum norm.

[0119] Please refer to Figure 2C , Figure 2C FIG. is a schematic diagram of a plurality of first registration sub-domains determined by the cube sub-domain determination method provided in an embodiment of the present application. As Figure 2C shown, exemplarily, based on the coordinates of each first feature point and the preset cube side length, a first registration sub-domain with the first feature point as the center point of the cube and the preset cube side length as the side length of the cube is determined, so as to obtain a plurality of first registration sub-domains A1 determined by the cube sub-domain determination method. At this time, the first registration sub-domain A1 is a cube sub-domain.

[0120] As described above, optionally, the plurality of second feature points include a plurality of second feature point pairs.

[0121] In some embodiments, when the target subdomain determination method is the inclined prism subdomain determination method, a first registration subdomain including a pair of first feature points is determined based on the coordinates of the two first feature points in each pair of first feature points and a preset distance value, thereby obtaining a plurality of first registration subdomains determined by the inclined prism subdomain determination method.

[0122] The target subdomain determination method corresponding to the first registration subdomain determined by the inclined prism subdomain determination method is the inclined prism subdomain determination method.

[0123] In some embodiments, when the target subdomain determination method is the inclined prism subdomain determination method, the calculation formula for the first registration subdomain is:

[0124]

[0125] Where (m)x represents the coordinate value of the point in Ri on the X-axis, (m)y represents the coordinate value of the point in Ri on the Y-axis, (m)z represents the coordinate value of the point in Ri on the Z-axis, xmin represents the minimum value of the coordinate values on the X-axis, xmax represents the maximum value of the coordinate values on the X-axis, ymin represents the minimum value of the coordinate values on the Y-axis, ymax represents the maximum value of the coordinate values on the Y-axis, zmin represents the minimum value of the coordinate values on the Z-axis, zmax represents the maximum value of the coordinate values on the Z-axis. ti represents the i-th first feature point, ti+1 represents the (i + 1)-th first feature point, and ti and ti+1 are the two first feature points in a pair of first feature points. (t i ) x represents the coordinate value of ti on the X-axis, (t i+1 ) x represents the coordinate value of ti+1 on the X-axis, (t i ) y represents the coordinate value of ti on the Y-axis, (t i+1 ) y represents the coordinate value of ti+1 on the Y-axis, (t i ) z represents the coordinate value of ti on the Z-axis, (t i+1 ) z represents the coordinate value of ti+1 on the Z-axis. d represents the preset distance value.

[0126] Please refer to Figure 2D , Figure 2D which is a schematic diagram of a plurality of first registration subdomains determined by the inclined prism subdomain determination method provided in the embodiments of the present application. As Figure 2DAs shown, exemplarily, a first registration sub-domain including the first feature point pair is determined based on the coordinates of two first feature points in each first feature point pair and a preset distance value, so as to obtain a plurality of first registration sub-domains A1 determined by using the inclined prism sub-domain determination method. At this time, the first registration sub-domain A1 is an inclined prism sub-domain.

[0127] Step S500: Determine a plurality of second registration sub-domains based on a plurality of second feature points.

[0128] Among them, the second registration sub-domains and the first registration sub-domains are in one-to-one correspondence.

[0129] In some embodiments, step S500 includes: for each target sub-domain determination method, use the target sub-domain determination method to determine a plurality of second registration sub-domains based on a plurality of second feature points, so as to obtain all second registration sub-domains.

[0130] In some embodiments, when there are multiple target sub-domain determination methods, sequentially use one of the multiple target sub-domain determination methods to determine a plurality of second registration sub-domains based on a plurality of second feature points. When all target sub-domain determination methods have been executed, all second registration sub-domains are obtained.

[0131] In some embodiments, when the target sub-domain determination method is the sphere sub-domain determination method, a second registration sub-domain with the second feature point as the center of the sphere and the preset radius as the radius of the sphere is determined based on the coordinates of each second feature point and the preset radius, so as to obtain a plurality of second registration sub-domains determined by using the sphere sub-domain determination method.

[0132] The method for determining a plurality of second registration sub-domains by using the sphere sub-domain determination method refers to the method for determining a plurality of first registration sub-domains by using the sphere sub-domain determination method.

[0133] The target sub-domain determination method corresponding to the second registration sub-domain determined by using the sphere sub-domain determination method is the sphere sub-domain determination method.

[0134] In some embodiments, when the target sub-domain determination method is the cube sub-domain determination method, a second registration sub-domain with the second feature point as the center of the cube and the preset side length of the cube as the side length of the cube is determined based on the coordinates of each second feature point and the preset side length of the cube, so as to obtain a plurality of second registration sub-domains determined by using the cube sub-domain determination method.

[0135] The method for determining a plurality of second registration sub-domains by using the cube sub-domain determination method refers to the method for determining a plurality of first registration sub-domains by using the cube sub-domain determination method.

[0136] The target sub-domain determination method corresponding to the second registration sub-domain determined by using the cube sub-domain determination method is the cube sub-domain determination method.

[0137] As described above, optionally, a plurality of second feature point pairs are included among the plurality of second feature points.

[0138] In some embodiments, when the target sub-domain determination method is the inclined prism column sub-domain determination method, based on the coordinates of two second feature points in each second feature point pair and a preset distance value, a second registration sub-domain including the second feature point pair is determined, so as to obtain a plurality of second registration sub-domains determined by the inclined prism column sub-domain determination method.

[0139] The method for determining a plurality of second registration sub-domains by the inclined prism column sub-domain determination method refers to the method for determining a plurality of first registration sub-domains by the inclined prism column sub-domain determination method.

[0140] The target sub-domain determination method corresponding to the second registration sub-domain determined by the inclined prism column sub-domain determination method is the inclined prism column sub-domain determination method.

[0141] By adopting different sub-domain determination methods, registration sub-domains of different shapes can be obtained, and local features of the face can be captured from different angles, so as to improve the registration accuracy for the facial features of the target object. The first registration sub-domains obtained by the spherical sub-domain determination method are spherically distributed with the first feature points as the center, and can uniformly obtain information around the first feature points from all directions, and can improve the registration accuracy for facial features with an approximately spherical distribution. The cube sub-domain determination method can regularly divide the registration area in the direction of a specific rectangular coordinate system. For facial features with obvious right-angle or rectangular features (such as the relatively regular shape features of the orbital region), the feature details of this area can be better captured, so as to improve the registration accuracy for facial features with right-angle or rectangular features. The inclined prism column sub-domain determination method can capture features within a specific direction and depth range according to the preset distance value and the preset generation direction, so as to improve the registration accuracy for facial features with inclined or specific-direction structures (such as the inclined structure facial features on the side of the nose bridge).

[0142] Step S600: Register each second registration sub-domain with the corresponding first registration sub-domain to obtain the first registration matrix of each second registration sub-domain.

[0143] In some embodiments, an improved ICP (Iterative Closest Point) algorithm is used to register each second registration sub-domain with the corresponding first registration sub-domain to obtain the first registration matrix of each second registration sub-domain.

[0144] Wherein, the second feature points or second feature point pairs used to determine the second registration sub-domain and the first feature points or first feature point pairs used to determine the first registration sub-domain corresponding to the second registration sub-domain are the feature points or feature point pairs at the same position in the point cloud data.

[0145] In some embodiments, the calculation formula of the first registration matrix is as follows:

[0146]

[0147] where Ti represents the first registration matrix between the i-th second registration sub-domain and the corresponding first registration sub-domain, q represents a point in Hi, Hi is the i-th second registration sub-domain corresponding to Ri, p represents a point in Ri, and Ti(q) represents the position transformation of q by Ti.

[0148] Optionally, that is, Ti is a 4×4 matrix.

[0149] Step S700: Register the second point cloud data with the first point cloud data based on the first registration matrices of all the second registration sub-domains.

[0150] By performing registration using different sub-domain determination methods, the local feature information in the point cloud data can be fully utilized to avoid global noise and outliers, thereby improving the registration accuracy.

[0151] Please refer to Figure 3 , Figure 3 which is a schematic flowchart of the second embodiment of the point cloud data registration method provided by the embodiments of the present application. As Figure 3 shown, in some embodiments, before step S400, the point cloud data registration method further includes: step S800 to step S900.

[0152] Step S800: Calculate the second registration matrix of the second point cloud data based on each first feature point and the corresponding second feature point.

[0153] In some embodiments, an improved ICP (Iterative Closest Point) algorithm is used to calculate the second registration matrix of the second point cloud data based on each first feature point and the corresponding second feature point, and the method refers to the above description.

[0154] Step S900: Multiply the second registration matrix on the left by the coordinates of each point in the second point cloud data, so as to perform the first registration of the second point cloud data with the first point cloud data.

[0155] After the first registration, the coordinates of each point in the second point cloud data have been updated.

[0156] In some embodiments, after performing step S800 and step S900, in step S500, multiple second registration sub-domains are determined based on multiple feature points in the second point cloud data after the first registration.

[0157] In some embodiments, step S700 includes steps S710 to S720.

[0158] Step S710: Weightedly fuse all the first registration matrices to obtain a third registration matrix.

[0159] In some embodiments, step S710 includes steps S711 to S713.

[0160] Step S711: Calculate the registration error between each second registration sub-domain and the corresponding first registration sub-domain.

[0161] In some embodiments, the calculation formula for the registration error between the second registration sub-domain and the corresponding first registration sub-domain is:

[0162]

[0163] where ei represents the registration error between the i-th second registration sub-domain and the corresponding first registration sub-domain, and n represents the number of points in Hi.

[0164] Step S712: Calculate the weight of each second registration sub-domain based on the registration error of each second registration sub-domain and the number of points in the second point cloud data included in the second registration sub-domain.

[0165] In some embodiments, step S712 includes steps S7121 to S7122.

[0166] Step S7121: Calculate the initial weight of each second registration sub-domain based on the registration error of each second registration sub-domain and the number of points in the second point cloud data included in the second registration sub-domain.

[0167] In some embodiments, step S7121 includes steps (7121.1) to (7121.3).

[0168] (7121.1) For each second registration sub-domain, calculate a first weight calculation parameter by taking the exponential of the opposite of the value obtained by dividing the registration error of the second registration sub-domain by the value of a preset first error weight as the preset base.

[0169] Optionally, the preset base is the base of the natural logarithm.

[0170] In some embodiments, the first error weight is used to adjust the influence of the registration error of the second registration sub-domain on the initial weight of the second registration sub-domain.

[0171] Optionally, the value of the first error weight is set according to the registration accuracy requirement. The larger the value of the first error weight, the higher the registration accuracy requirement.

[0172] Optionally, the numerical range of the first error weight is from 0 mm (millimeter) to 1 mm.

[0173] Exemplarily, the first error weight is 1 mm.

[0174] (7121.2) Calculate the second weight calculation parameter by taking the exponent with the opposite value of the number of points in the second point cloud data included in the second registration sub-domain divided by the preset second error weight as the preset base number.

[0175] In some embodiments, the second error weight is used to adjust the influence of the number of points in the second point cloud data included in the second registration sub-domain on the initial weight of the second registration sub-domain.

[0176] Optionally, set the numerical value of the second error weight according to the point cloud sampling density of the structured light camera. The larger the numerical value of the second error weight, the higher the point cloud sampling density of the structured light camera.

[0177] Optionally, the numerical range of the second error weight is from 1000 to 10000.

[0178] Exemplarily, the second error weight is 5000.

[0179] (7121.3) Multiply the first weight calculation parameter by the second weight calculation parameter to obtain the initial weight of the second registration sub-domain.

[0180] In some embodiments, the formula for calculating the initial weight of the second registration sub-domain is:

[0181]

[0182] Wherein, Φ(Hi) represents the initial weight of Hi, exp represents the base of the natural logarithm, σ1 represents the first error weight, σ2 represents the second error weight, represents the value of the first weight calculation parameter, represents the value of the second weight calculation parameter, and vi represents the number of points in the second point cloud data included in Hi.

[0183] In some embodiments, the closer the initial weight of the second registration sub-domain is to 1, the more reliable the first registration matrix of the second registration sub-domain is.

[0184] Step S7122: Calculate the normalized weight of each second registration sub-domain based on the initial weight of each second registration sub-domain and the sum of the initial weights of all second registration sub-domains.

[0185] In some embodiments, when the multiple first registration sub-domains and the multiple second registration sub-domains are determined by the same target sub-domain determination method, divide the initial weight of each second registration sub-domain by the sum of the initial weights of all second registration sub-domains to obtain the normalized weight of each second registration sub-domain.

[0186] In some embodiments, when the multiple first registration sub-domains and the multiple second registration sub-domains are determined by multiple target sub-domain determination methods, each target sub-domain determination method corresponds to a preset sub-domain weight, and step S7122 includes steps (7122.1) to step (7122.3).

[0187] (7122.1) For each second registration sub-domain, divide the initial weight of the second registration sub-domain by the sum of the initial weights of all second registration sub-domains to obtain a first calculated value.

[0188] (7122.2) Divide the sub-domain weight of the target sub-domain determination method corresponding to the second registration sub-domain by the sum of the sub-domain weights of all target sub-domain determination methods to obtain a second calculated value.

[0189] (7122.3) Multiply the first calculated value by the second calculated value to obtain the normalized weight of the second registration sub-domain.

[0190] Step S713: Based on the weights of each second registration sub-domain and the first registration matrix of the second registration sub-domain, perform weighted fusion on all first registration matrices to obtain a third registration matrix.

[0191] In some embodiments, step S713 includes steps S7131 to step S7133.

[0192] Step S7131: Convert each first registration matrix into the Lie algebra space to obtain the corresponding first registration matrix element.

[0193] In some embodiments, the calculation formula of the first registration matrix element is:

[0194] ξi = log(Ti) ∈ se(3),

[0195] where ξi represents the first registration matrix element corresponding to the i-th first registration matrix, and se(3) represents the Lie algebra space.

[0196] Step S7132: Based on the weights of each second registration sub-domain and the first registration matrix elements corresponding to the second registration sub-domain, perform weighted fusion on all first registration matrix elements to obtain the weighted-fused first registration matrix elements.

[0197] In some embodiments, the calculation formula of the weighted-fused first registration matrix elements:

[0198]

[0199] Among them, ξfine represents the elements of the first registration matrix after weighted fusion, wi represents the normalized weight of the i-th second registration sub-domain, and a represents the total number of second registration sub-domains.

[0200] Step S7133: Convert the elements of the first registration matrix after weighted fusion into a third registration matrix in Euclidean space.

[0201] In some embodiments, the calculation formula of the third registration matrix is:

[0202] Tfine = exp(ξfine) ∈ SE(3),

[0203] where Tfine represents the third registration matrix, and SE(3) represents the situation of the special Euclidean group in three-dimensional space, which contains the set of all possible three-dimensional rigid body transformations.

[0204] Step S720: Left-multiply the coordinates of each point in the second point cloud data by the third registration matrix, so as to perform the second registration of the second point cloud data and the first point cloud data.

[0205] By registering the first point cloud data and the second point cloud data twice, it is possible to use both global feature information and local feature information for registration, and by registering the first point cloud data and the second point cloud data a second time, it is possible to make full use of the local feature information in the point cloud data to avoid global noise and outliers, thereby improving the registration accuracy.

[0206] Please refer to Figures 4A to 4D , Figure 4A which is a schematic diagram of the first reconstruction model provided by the embodiment of the present application. Figure 4B which is a schematic diagram of the second point cloud data provided by the embodiment of the present application. Figure 4C which is a schematic diagram of the result after the first registration of the second point cloud data and the first point cloud data. Figure 4D which is a schematic diagram of the result after the second registration of the second point cloud data and the first point cloud data. As Figures 4A to 4D shown, after the first registration of the second point cloud data and the first point cloud data, the first reconstruction model 10 and the second point cloud data 20 are roughly overlapped, but there are still obvious non-overlapping parts of the second point cloud data 20. After the second registration of the second point cloud data and the first point cloud data, the first reconstruction model 10 and the second point cloud data 20 almost completely overlap, and there are no obvious non-overlapping parts of the second point cloud data 20. This shows that by using the point cloud data registration method of the present application, a high registration accuracy can still be achieved without pasting or implanting marker points on the target object.

[0207] In some embodiments, when the point cloud data registration method does not include step S800 and step S900, the third registration matrix is left-multiplied by the coordinates of each point in the second point cloud data, so as to register the second point cloud data with the first point cloud data.

[0208] In summary, the point cloud data registration method provided by the embodiments of the present application has the following advantages:

[0209] 1. By registering the second point cloud data with the first point cloud data through the first registration matrix based on all the second registration sub-domains, a very high registration accuracy can still be achieved without pasting or implanting marker points on the target object, thereby reducing the dependence on marker points during the operation, reducing the risk of infection of the target object, and reducing the risk of displacement and detachment of the marker points during the operation, improving the safety of the operation. In addition, without pasting or implanting markers before the operation, the operation preparation time is significantly reduced.

[0210] 2. By adopting different sub-domain determination methods for registration, the local feature information in the point cloud data can be fully utilized to avoid global noise and outliers, thereby improving the registration accuracy.

[0211] 3. By adopting different sub-domain determination methods, registration sub-domains of different shapes can be obtained, and the local features of the face can be captured from different angles, thereby improving the registration accuracy for the facial features of the target object. The first registration sub-domain obtained by the spherical sub-domain determination method is spherically distributed with the first feature point as the center, and can uniformly obtain the information around the first feature point from all directions, and can improve the registration accuracy for facial features with an approximately spherical distribution. The cube sub-domain determination method can regularly divide the registration area in the direction of a specific rectangular coordinate system. For facial features with obvious right-angle or rectangular features (such as the relatively regular shape features of the orbital region), the feature details of this area can be better captured, thereby improving the registration accuracy for facial features with right-angle or rectangular features. The oblique prism sub-domain determination method can capture the features within a specific direction and depth range according to the preset distance value and the preset generation direction, thereby improving the registration accuracy for facial features with inclined or specific-direction structures (such as the inclined structure facial features on the side of the nose bridge).

[0212] 4. By registering the first point cloud data and the second point cloud data twice, the global feature information and the local feature information can be utilized simultaneously for registration, and by registering the first point cloud data and the second point cloud data for the second time, the local feature information in the point cloud data can be fully utilized to avoid global noise and outliers, thereby improving the registration accuracy.

[0213] Please refer to Figure 5 , Figure 5It is a schematic structural diagram of a point cloud data registration device provided by an embodiment of the present application. As Figure 5 shown, the point cloud data registration device 300 includes an acquisition module 310 and a processing module 320.

[0214] In some embodiments, the acquisition module 310 is configured to acquire first point cloud data and second point cloud data. The first point cloud data is obtained by scanning a target object using a first scanning method, and the second point cloud data is obtained by scanning the target object using a second scanning method. The first scanning method is different from the second scanning method.

[0215] In some embodiments, the processing module 320 is configured to receive a first selection instruction, determine a plurality of first feature points in the first point cloud data according to the first selection instruction; receive a second selection instruction, determine a plurality of second feature points in the second point cloud data according to the second selection instruction, and the second feature points correspond to the first feature points one by one; determine a plurality of first registration sub-domains based on the plurality of first feature points; determine a plurality of second registration sub-domains based on the plurality of second feature points, and the second registration sub-domains correspond to the first registration sub-domains one by one; register each second registration sub-domain with the corresponding first registration sub-domain to obtain a first registration matrix of each second registration sub-domain; register the second point cloud data with the first point cloud data based on the first registration matrices of all second registration sub-domains.

[0216] Please refer to Figure 6 , Figure 6 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 6 shown, the electronic device 400 includes: one or more processors 410 and a memory 420. Figure 6 Here, one processor 410 is taken as an example.

[0217] In some embodiments, the processor 410 and the memory 420 can be connected through a bus or other means. Figure 6 Here, connection through a bus is taken as an example.

[0218] In some embodiments, the processor 410 is configured to obtain first point cloud data and second point cloud data. The first point cloud data is obtained by scanning a target object using a first scanning method, and the second point cloud data is obtained by scanning the target object using a second scanning method, where the first scanning method is different from the second scanning method. The processor 410 receives a first selection instruction and determines a plurality of first feature points in the first point cloud data according to the first selection instruction. The processor 410 receives a second selection instruction and determines a plurality of second feature points in the second point cloud data according to the second selection instruction, where the second feature points correspond to the first feature points one by one. The processor 410 determines a plurality of first registration sub-domains based on the plurality of first feature points, and determines a plurality of second registration sub-domains based on the plurality of second feature points, where the second registration sub-domains correspond to the first registration sub-domains one by one. The processor 410 registers each second registration sub-domain with the corresponding first registration sub-domain to obtain a first registration matrix for each second registration sub-domain, and registers the second point cloud data with the first point cloud data based on the first registration matrices of all the second registration sub-domains.

[0219] In some embodiments, the memory 420, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as program instructions / modules for the point cloud data registration method in the embodiments of the present application. By running the non-volatile software programs, instructions, and modules stored in the memory 420, the processor 410 executes various functional applications and data processing of the electronic device 400, that is, implements the point cloud data registration method in the above method embodiments.

[0220] In some embodiments, the memory 420 may include a program storage area and a data storage area. The program storage area can store an operating system and application programs required for at least one function. The data storage area can store data created according to the use of the electronic device 400. In addition, the memory 420 may include a high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 420 may optionally include a memory remotely provided with respect to the processor 410, and these remote memories can be connected to the controller through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0221] In some embodiments, one or more modules are stored in the memory 420 and, when executed by one or more processors 410, execute the point cloud data registration method in any of the above method embodiments. For example, execute the method steps S100 to S700 described above. Figure 1 in

[0222] Please refer to Figure 7 , Figure 7It is a structural block diagram of a computer-readable storage medium provided by an embodiment of the present application. Program code 510 is stored in the computer-readable storage medium 500, and the program code 510 can be called by a processor to execute the point cloud data registration method described in the above method embodiment.

[0223] The computer-readable storage medium 500 can be an electronic memory such as a flash memory, EEPROM (electrically erasable programmable read-only memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium includes a non-transitory computer-readable storage medium. The computer-readable storage medium 500 has a storage space for program code that executes any method step in the above point cloud data registration method. These program codes can be read from or written to one or more computer program products. The program codes can be compressed in an appropriate form, for example.

[0224] In some embodiments, the embodiments of the present application further provide a computer program product, including a computer program, which implements the above point cloud data registration method when executed by a processor.

[0225] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in the computer-readable storage medium 500. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the various embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the various embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the various embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0226] In summary, the present application provides a point cloud data registration method and an electronic device. The point cloud data registration method includes: obtaining first point cloud data and second point cloud data, where the first point cloud data is obtained by scanning a target object using a first scanning method, and the second point cloud data is obtained by scanning the target object using a second scanning method, and the first scanning method is different from the second scanning method; receiving a first selection instruction, and determining a plurality of first feature points in the first point cloud data according to the first selection instruction; receiving a second selection instruction, and determining a plurality of second feature points in the second point cloud data according to the second selection instruction, and the second feature points correspond to the first feature points one by one; determining a plurality of first registration sub-domains based on the plurality of first feature points; determining a plurality of second registration sub-domains based on the plurality of second feature points, and the second registration sub-domains correspond to the first registration sub-domains one by one; registering each second registration sub-domain with the corresponding first registration sub-domain to obtain a first registration matrix of each second registration sub-domain; registering the second point cloud data with the first point cloud data based on the first registration matrices of all the second registration sub-domains. By registering the second point cloud data with the first point cloud data based on the first registration matrices of all the second registration sub-domains, the present application can still achieve a very high registration accuracy without pasting or implanting marker points on the target object, thereby reducing the dependence on marker points during the operation, reducing the risk of infection of the target object, and reducing the risks of displacement and falling off of the marker points during the operation, and improving the safety of the operation.

[0227] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A point cloud data registration method, characterized in that: include: Acquire first point cloud data and second point cloud data, wherein the first point cloud data is obtained by scanning the target object using a first scanning method, and the second point cloud data is obtained by scanning the target object using a second scanning method, and the first scanning method is different from the second scanning method; receiving a first selection instruction, and determining a plurality of first feature points in the first point cloud data according to the first selection instruction; receiving a second selection instruction, and determining a plurality of second feature points in the second point cloud data according to the second selection instruction, wherein the second feature points correspond one-to-one to the first feature points; determining a plurality of first registration sub-domains based on the plurality of first feature points; Determine a plurality of second registration sub-domains based on the plurality of second feature points, wherein the second registration sub-domains correspond one-to-one to the first registration sub-domains; Registering each of the second registration subdomains with the corresponding first registration subdomain to obtain a first registration matrix for each second registration subdomain; The second point cloud data is registered with the first point cloud data based on the first registration matrix of all the second registration subdomains.

2. The point cloud data registration method according to claim 1, characterized in that: The determining of a plurality of first registration sub-domains based on the plurality of first feature points comprises: receiving a selection instruction, and selecting at least one target subdomain determination method from among a plurality of subdomain determination methods according to the selection instruction; For each of the target subdomain determination methods, the target subdomain determination method is used to determine a plurality of the first registration subdomains based on the plurality of first feature points, thereby obtaining all the first registration subdomains; The determining a plurality of second registration sub-domains based on the plurality of second feature points comprises: For each of the target subdomain determination methods, the target subdomain determination method is used to determine multiple second registration subdomains based on the multiple second feature points, thereby obtaining all the second registration subdomains.

3. The point cloud data registration method according to claim 2, characterized in that: The multiple subdomain determination methods include a sphere subdomain determination method, a cube subdomain determination method, and an oblique prism subdomain determination method, the multiple first feature points include multiple first feature point pairs, and the multiple second feature points include multiple second feature point pairs. For each of the target subdomain determination methods, the target subdomain determination method is used to determine a plurality of the first registration subdomains based on the plurality of first feature points, thereby obtaining all the first registration subdomains, including: When the target subdomain determination method is the spherical subdomain determination method, a first registration subdomain with the first feature point as the center point of the sphere and the preset radius as the radius of the sphere is determined based on the coordinates of each of the first feature points and a preset radius, thereby obtaining a plurality of first registration subdomains determined by the spherical subdomain determination method; When the target subdomain determination method is the cube subdomain determination method, a first registration subdomain is determined based on the coordinates of each of the first feature points and the preset cube side length, with the first feature point as the center point of the cube and the preset cube side length as the cube side length, thereby obtaining a plurality of first registration subdomains determined by the cube subdomain determination method; When the target subdomain determination method is the oblique prism subdomain determination method, determining a first registration subdomain including the first feature point pair based on the coordinates of two first feature points in each first feature point pair and a preset distance value, thereby obtaining a plurality of first registration subdomains determined by the oblique prism subdomain determination method; For each of the target subdomain determination methods, using the target subdomain determination method to determine a plurality of second registration subdomains based on the plurality of second feature points, thereby obtaining all the second registration subdomains, includes: When the target subdomain determination method is the spherical subdomain determination method, a second registration subdomain with the second feature point as the sphere center point and the preset radius as the sphere radius is determined based on the coordinates of each of the second feature points and the preset radius, thereby obtaining a plurality of second registration subdomains determined by the spherical subdomain determination method; When the target subdomain determination method is the cube subdomain determination method, a second registration subdomain is determined based on the coordinates of each of the second feature points and the preset cube side length, with the second feature point as the cube center point and the preset cube side length as the cube side length, thereby obtaining a plurality of second registration subdomains determined by the cube subdomain determination method; When the target subdomain determination method is the oblique prism subdomain determination method, the second registration subdomain including the second feature point pair is determined based on the coordinates of two second feature points in each second feature point pair and the preset distance value, thereby obtaining a plurality of second registration subdomains determined by the oblique prism subdomain determination method.

4. The point cloud data registration method according to claim 1, characterized in that: Before determining a plurality of first registration sub-domains based on the plurality of first feature points, the method further includes: Calculate a second registration matrix of the second point cloud data based on each of the first feature points and the corresponding second feature points; Multiplying the second registration matrix by the coordinates of each point in the second point cloud data, thereby performing a first registration on the second point cloud data and the first point cloud data; The registering the second point cloud data with the first point cloud data based on the first registration matrix of all the second registration subdomains includes: Performing weighted fusion on all the first registration matrices to obtain a third registration matrix; The third registration matrix is ​​left-multiplied by the coordinates of each point in the second point cloud data, so as to perform a second registration on the second point cloud data and the first point cloud data.

5. The point cloud data registration method according to claim 4, characterized in that: The step of weighted fusion of all the first registration matrices to obtain a third registration matrix includes: Calculating a registration error between each of the second registration sub-domains and the corresponding first registration sub-domain; calculating a weight of each of the second registration subdomains based on a registration error of each of the second registration subdomains and the number of points in the second point cloud data included in the second registration subdomain; All the first registration matrices are weightedly fused based on the weight of each of the second registration subdomains and the first registration matrix of the second registration subdomain to obtain the third registration matrix.

6. The point cloud data registration method according to claim 5, characterized in that: The step of calculating the weight of each second registration subdomain based on the registration error of each second registration subdomain and the number of points in the second point cloud data included in the second registration subdomain comprises: Calculating an initial weight of each second registration subdomain based on the registration error of each second registration subdomain and the number of points in the second point cloud data included in the second registration subdomain; A normalized weight of each second registration sub-domain is calculated based on an initial weight of each second registration sub-domain and a sum of the initial weights of all second registration sub-domains.

7. The point cloud data registration method according to claim 6, characterized in that: The calculating the initial weight of each second registration subdomain based on the registration error of each second registration subdomain and the number of points in the second point cloud data included in the second registration subdomain comprises: For each of the second registration subdomains, a first weight calculation parameter is calculated by dividing the registration error of the second registration subdomain by the inverse of the value of the preset first error weight as an exponent of a preset base; A second weight calculation parameter is calculated by dividing the number of points in the second point cloud data included in the second registration subdomain by the inverse of a value of a preset second error weight as an exponent of a preset base; The first weight calculation parameter is multiplied by the second weight calculation parameter to obtain an initial weight of the second registration subdomain.

8. The point cloud data registration method according to claim 6, characterized in that: When the plurality of first registration subdomains and the plurality of second registration subdomains are determined by using a plurality of target subdomain determination methods, each of the target subdomain determination methods corresponds to a preset subdomain weight, The calculating the normalized weight of each second registration subdomain based on the initial weight of each second registration subdomain and the sum of the initial weights of all the second registration subdomains comprises: For each of the second registration subdomains, dividing the initial weight of the second registration subdomain by the sum of the initial weights of all the second registration subdomains to obtain a first calculated value; Dividing the subdomain weight of the target subdomain determination method corresponding to the second registration subdomain by the sum of the subdomain weights of all the target subdomain determination methods to obtain a second calculated value; The first calculated value is multiplied by the second calculated value to obtain a normalized weight of the second registration subdomain.

9. The point cloud data registration method according to claim 5, characterized in that: The step of weighting and fusing all the first registration matrices based on the weight of each of the second registration subdomains and the first registration matrix of the second registration subdomain to obtain the third registration matrix comprises: Convert each of the first registration matrices into the Lie algebraic space to obtain corresponding first registration matrix elements; weightedly fusion all the first registration matrix elements based on the weight of each second registration subdomain and the first registration matrix element corresponding to the second registration subdomain to obtain a weighted fusion first registration matrix element; The weighted fused first registration matrix elements are converted into the third registration matrix in the Euclidean space.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the point cloud data registration method according to any one of claims 1 to 9.

Images