An image processing method, apparatus, and electronic device

Abstract

The present invention discloses an image processing method, apparatus and electronic device, belonging to the technical field of medical image processing. The method includes: segmenting the bone surface of an ultrasound image through a denoising channel Unet model; converting the 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data through an ultrasound and navigation system; segmenting a bone and joint CT image to obtain CT bone surface point cloud data; performing ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data. Through the image processing method disclosed by the present invention, the registration efficiency of ultrasound and CT images can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular, to an image processing method, an apparatus, and an electronic device. Background Art

[0002] In view of the advantages of high safety, low cost, real-time performance, etc. of ultrasound, traditional orthopedic surgeries are usually guided by ultrasound. Due to tissue occlusion, difficult spatial matching, and especially the high echo phenomenon of low-water tissues in the elderly population, accurately locating the spatial position of the lesion is a challenge and there are certain surgical risks. With the development of computer technology and medical technology, computer-aided diagnosis and treatment technology has great potential in solving existing clinical problems. Although important structure three-dimensional reconstruction based on images such as CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) can be implemented through computer image segmentation technology, the difference in the patient's body posture before and during surgery and tissue deformation will affect the matching of the three-dimensional reconstruction structure to the real patient tissue.

[0003] In view of the fact that the existing registration methods for osteoarthritic ultrasound images and CT images cannot meet the indexes such as accuracy rate and registration time. Therefore, there is an urgent need for a registration method for osteoarthritic ultrasound images and CT images to effectively improve the registration efficiency of ultrasound and CT images. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide an image processing method, an apparatus, and an electronic device, which can solve the problem of low registration accuracy of osteoarthritic ultrasound images and CT images in the prior art.

[0005] To solve the above technical problems, the present invention provides the following technical solutions:

[0006] An image processing method, wherein the method includes: segmenting the bone surface of the ultrasound image through a denoising channel Unet model; converting the 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data through an ultrasound and navigation system; segmenting the osteoarthritic CT image to obtain CT bone surface point cloud data; performing ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data.

[0007] Among them, the steps of segmenting the bone surface of the ultrasound image through the denoising-channel Unet model include: preprocessing the data of the osteoarticular ultrasound image based on the histogram equalization method; adding Gaussian filtering and bilateral filtering denoising channels to the Skip-layer layer in the Unet neural network model structure to obtain the denoising-channel Unet model; inputting the preprocessed osteoarticular ultrasound image into the denoising-channel Unet model to obtain the ultrasound bone surface segmentation result.

[0008] Among them, the steps of converting the 2D bone surface data in the segmented ultrasound image into 3D ultrasound bone surface point cloud data through the ultrasound and navigation system include: obtaining the spatial position and attitude information of the ultrasound probe through the ultrasound and navigation system; combining the spatial position and attitude information with the ultrasound bone surface segmentation result to generate 3D ultrasound bone surface sparse point cloud data; using the Bezier interpolation algorithm for the 3D bone surface sparse point cloud data to generate 3D ultrasound bone surface dense point cloud data.

[0009] Among them, the steps of segmenting the osteoarticular CT image to obtain CT bone surface point cloud data include: using the Unet neural network model to segment the osteoarticular CT image to obtain the segmentation result; generating CT bone surface point cloud data based on the segmentation result.

[0010] Among them, the steps of performing ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data include: defining the 3D ultrasound bone surface point cloud data as the point cloud to be registered P = {p1, p2, …, p n}, and the CT bone surface point cloud data as the target point cloud Q = {q1, q2, …, q n}; performing initial registration on the point cloud to be registered P and the target point cloud Q; using the ICP point cloud registration algorithm to register the point clouds P and Q.

[0011] Among them, the steps of registering the point clouds P and Q using the ICP point cloud registration algorithm include: establishing the relationship between the point cloud to be registered P and the target point cloud Q as

[0012] where R is the rotation matrix and T is the translation matrix; calculating the rotation matrix R and the translation matrix T according to the initial registration relationship between the point cloud to be registered P and the target point cloud Q, and obtaining the new point cloud P' after transformation; iteratively obtaining the rotation matrix R and the translation matrix T until the average distance between the point cloud to be registered and the target point cloud is less than the set threshold or the number of iterations reaches the maximum set number; obtaining the optimal rotation matrix R and the translation matrix T; registering the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data based on the optimal rotation matrix R and the translation matrix T.

[0013] An image processing apparatus, wherein the apparatus includes: a first segmentation module for segmenting the bone surface of an ultrasound image through a denoising-channel Unet model; a conversion module for converting 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data through an ultrasound and navigation system; a second segmentation module for segmenting a bone joint CT image to obtain CT bone surface point cloud data; and a registration module for performing ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data.

[0014] Wherein, the first segmentation module includes:

[0015] A first sub-module for preprocessing data of a bone joint ultrasound image based on a histogram equalization method;

[0016] A second sub-module for adding Gaussian filtering and bilateral filtering denoising channels to the Skip-layer layer in the Unet neural network model structure to obtain a denoising-channel Unet model;

[0017] A third sub-module for inputting the preprocessed bone joint ultrasound image into the denoising-channel Unet model to obtain an ultrasound bone surface segmentation result.

[0018] Wherein, the conversion module includes:

[0019] A fourth sub-module for obtaining spatial position and attitude information of an ultrasound probe through an ultrasound and navigation system;

[0020] A fifth sub-module for combining the spatial position and attitude information with the ultrasound bone surface segmentation result to generate 3D ultrasound bone surface sparse point cloud data;

[0021] A sixth sub-module for using a Bezier interpolation algorithm on the 3D bone surface sparse point cloud data to generate 3D ultrasound bone surface dense point cloud data.

[0022] An embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program or instruction stored on the memory and executable on the processor. When the program or instruction is executed by the processor, the steps of any of the above image processing methods are implemented.

[0023] An embodiment of the present invention provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of any of the above image processing methods are implemented.

[0024] The image processing method provided by the embodiment of the present invention can be applied to the registration of bone and joint ultrasound images and CT images. When registering the bone and joint ultrasound images and CT images, on the one hand, based on the Unet model with a denoising channel of Gaussian filtering and bilateral filtering, the contrast of texture information in the ultrasound image is effectively improved, and the segmentation effect of the bone surface in the ultrasound image is enhanced; on the other hand, based on the point cloud registration technology for multi-modal images, the registration accuracy of the bone and joint ultrasound image and the bone and joint CT image can be effectively improved, and the registration time can be reduced. Description of the Drawings

[0025] Figure 1 It is a flowchart showing the steps of an image processing method according to an embodiment of the present application;

[0026] Figure 2 It is a block diagram showing the structure of an image processing device according to an embodiment of the present application;

[0027] Figure 3 It is a block diagram showing the structure of an electronic device according to an embodiment of the present application. Detailed Embodiments

[0028] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0029] The image processing method provided by the present invention can solve the positioning problem during orthopedic surgery. Specifically, through the registration of bone and joint ultrasound images and CT images by this image processing method, accurate positioning of the position to be processed during orthopedic surgery can be achieved. Specifically, it is manifested by collecting three-dimensional medical images such as CT and MRI of the patient before surgery, and reconstructing the three-dimensional surface point cloud data of the lesion through computer image segmentation technology; during surgery, through the ultrasound and navigation system, the ultrasound three-dimensional information is obtained, and the 3D bone surface point cloud data is generated from the 2D bone surface data; finally, the ultrasound bone surface point cloud data and the CT bone surface point cloud data are subjected to ICP point cloud registration, so as to provide guidance for the doctor during the surgery process.

[0030] The following will, with reference to the accompanying drawings, describe in detail the image processing solution provided by the embodiments of the present application through specific embodiments and their application scenarios.

[0031] The image processing method of the embodiment of the present application includes the following steps:

[0032] Step 101: Segment the bone surface of the ultrasound image through a Unet model with a denoising channel.

[0033] This step is the bone surface segmentation step of the ultrasound image. An optional way to segment the bone surface of the ultrasound image through a Unet model with a denoising channel can be as follows:

[0034] First, based on the histogram equalization method, preprocess the data of the osteoarticular ultrasound images;

[0035] Preprocessing the data of the osteoarticular ultrasound images using the histogram equalization-based method can enhance the contrast of the ultrasound images.

[0036] Secondly, add Gaussian filtering and bilateral filtering denoising channels to the Skip-layer in the Unet neural network model structure to obtain a Unet model with denoising channels;

[0037] By adding Gaussian filtering and bilateral filtering denoising channels to the Skip-layer in the Unet neural network model structure, the contrast of the texture information of the ultrasound images can be enhanced, and the osteoarticular segmentation effect can be improved.

[0038] Finally, input the preprocessed osteoarticular ultrasound images into the Unet model with denoising channels to obtain the segmentation result of the ultrasound bone surface.

[0039] The described Unet model structure is a U-shaped structure neural network. On the left side, there are 4 convolutional layers with denoising channels, and on the right side, there are 4 upsampling layers. The feature maps obtained by each convolutional layer will be merged into the corresponding upsampling layer to apply each layer of feature maps to subsequent calculations, and finally output the segmentation result of the ultrasound bone surface.

[0040] Step 102: Through the ultrasound and navigation system, convert the 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data.

[0041] An optional way to convert the 2D bone surface data in the segmented ultrasound image into 3D ultrasound bone surface point cloud data through the ultrasound and navigation system can be as follows:

[0042] First, through the ultrasound and navigation system, obtain the spatial position and attitude information of the ultrasound probe;

[0043] Secondly, combine the spatial position and attitude information with the ultrasound bone surface segmentation result to generate 3D ultrasound bone surface sparse point cloud data;

[0044] Finally, use the Bezier interpolation algorithm for the 3D bone surface sparse point cloud data to generate 3D ultrasound bone surface dense point cloud data.

[0045] Step 103: Segment the osteoarticular CT images to obtain CT bone surface point cloud data.

[0046] This step is the CT image bone surface point cloud reconstruction step. An optional way to segment the osteoarticular CT images to obtain CT bone surface point cloud data can be as follows:

[0047] Segment the CT images of bones and joints using the Unet neural network model to obtain the segmentation results; generate the CT bone surface point cloud data based on the segmentation results.

[0048] Step 104: Perform ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data.

[0049] When performing ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data, it can be specifically implemented through the following steps:

[0050] First, define the 3D ultrasound bone surface point cloud data as the point cloud to be registered P = {p1, p2,..., p n}, and the CT bone surface point cloud data as the target point cloud Q = {q1, q2,..., q n};

[0051] Secondly, perform initial registration on the point cloud P to be registered and the target point cloud Q;

[0052] In the specific implementation process, the mean and covariance matrix of the point clouds P and Q can be calculated, and the reference coordinate systems can be adjusted to be consistent to complete the initial registration of the point cloud P to be registered and the target point cloud Q.

[0053] Finally, use the ICP point cloud registration algorithm to register the point clouds P and Q.

[0054] Specifically, the steps of using the ICP point cloud registration algorithm to register the point clouds P and Q include:

[0055] Establish the relationship between the point cloud P to be registered and the target point cloud Q as

[0056]

[0057] where R is the rotation matrix and T is the translation matrix;

[0058] According to the initial registration relationship between the point cloud P to be registered and the target point cloud Q, calculate the rotation matrix R and the translation matrix T, and obtain the new point cloud P' after transformation;

[0059] Iteratively obtain the rotation matrix R and the translation matrix T until the average distance between the point cloud to be registered and the target point cloud is less than the set threshold or the number of iterations reaches the maximum set number of times, which is expressed by the following formula:

[0060]

[0061] where n is the upper limit of the set number of iterations, and the specific value of n can be flexibly set by those skilled in the art, and the embodiments of the present application do not make specific limitations on this.

[0062] Obtain the optimal rotation matrix R and translation matrix T;

[0063] Based on the optimal rotation matrix R and translation matrix T, the 3D ultrasound bone surface point cloud data is registered with the CT bone surface point cloud data, so as to realize the registration of the ultrasound point cloud data of the bone joint and the CT point cloud data of the bone joint.

[0064] The image processing method provided by the embodiments of the present application performs bone surface segmentation of ultrasound images, ultrasound three-dimensional point cloud reconstruction, and CT image bone surface point cloud reconstruction, and finally performs ICP point cloud registration based on the ultrasound bone surface point cloud data and the CT bone surface point cloud data. On the one hand, the Unet model based on the Gaussian filter and the bilateral filter denoising channel effectively improves the contrast of the texture information of the ultrasound image and the bone surface segmentation effect of the ultrasound image; on the other hand, the point cloud registration technology based on multi-modal images can effectively improve the registration accuracy of the ultrasound images of the bone joint and the CT images of the bone joint, and reduce the registration time, which is beneficial for guiding doctors to perform surgery during the operation.

[0065] Figure 2 It is a structural block diagram of an image processing device according to an embodiment of the present application.

[0066] The image processing device of the embodiments of the present application includes the following functional modules:

[0067] The first segmentation module 201 is used to segment the bone surface of the ultrasound image through the Unet model with a denoising channel.

[0068] The conversion module 202 is used to convert the 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data through the ultrasound and the navigation system.

[0069] The second segmentation module 203 is used to segment the bone joint CT image to obtain CT bone surface point cloud data.

[0070] The registration module 204 is used to perform ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data.

[0071] Optionally, the first segmentation module includes:

[0072] The first sub-module is used to perform data preprocessing on the bone joint ultrasound image based on the histogram equalization method.

[0073] The second sub-module is used to add a Gaussian filter and a bilateral filter denoising channel to the Skip-layer layer in the Unet neural network model structure to obtain the Unet model with a denoising channel.

[0074] The third sub-module inputs the preprocessed bone joint ultrasound image into the Unet model with a denoising channel to obtain the ultrasound bone surface segmentation result.

[0075] Optionally, the transformation module includes:

[0076] A fourth sub-module for obtaining the spatial position and attitude information of the ultrasound probe through an ultrasound and navigation system;

[0077] A fifth sub-module for combining the spatial position and attitude information with the ultrasound bone surface segmentation result to generate 3D ultrasound bone surface sparse point cloud data;

[0078] A sixth sub-module for using the Bezier interpolation algorithm on the 3D bone surface sparse point cloud data to generate 3D ultrasound bone surface dense point cloud data.

[0079] Optionally, the second segmentation module is specifically configured to: segment the bone and joint CT image using a Unet neural network model to obtain a segmentation result; and generate CT bone surface point cloud data based on the segmentation result.

[0080] Optionally, the registration module includes:

[0081] A seventh sub-module for defining the 3D ultrasound bone surface point cloud data as the point cloud P to be registered = {p1, p2,..., p n}, and the CT bone surface point cloud data as the target point cloud Q = {q1, q2,..., q n};

[0082] An eighth sub-module for performing initial registration on the point cloud P to be registered and the target point cloud Q;

[0083] A ninth sub-module for registering the point clouds P and Q using the ICP point cloud registration algorithm.

[0084] Optionally, the ninth sub-module is specifically configured to:

[0085] Establish the relationship between the point cloud P to be registered and the target point cloud Q as

[0086]

[0087] where R is the rotation matrix and T is the translation matrix;

[0088] Calculate the rotation matrix R and the translation matrix T according to the initial registration relationship between the point cloud P to be registered and the target point cloud Q, and obtain the new point cloud P' after transformation;

[0089] Iteratively obtain the rotation matrix R and the translation matrix T until the average distance between the point cloud to be registered and the target point cloud is less than the set threshold or the number of iterations reaches the maximum set number;

[0090] Obtain the optimal rotation matrix R and translation matrix T;

[0091] Based on the optimal rotation matrix R and translation matrix T, register the 3D ultrasonic bone surface point cloud data and the CT bone surface point cloud data.

[0092] When the image processing device provided by the embodiment of the present application registers the osteoarticular ultrasonic image and the CT image, on the one hand, based on the Unet model of the Gaussian filtering and bilateral filtering denoising channels, it effectively improves the contrast of the texture information of the ultrasonic image and improves the segmentation effect of the bone surface of the ultrasonic image; on the other hand, based on the point cloud registration technology of multi-modal images, it can effectively improve the registration accuracy of the osteoarticular ultrasonic image and the osteoarticular CT image, and reduce the registration time.

[0093] In the embodiment of the present application Figure 2 The image processing device shown can be a device, or a component, an integrated circuit, or a chip in a server. In the embodiment of the present application Figure 2 The image processing device shown can be a device with an operating system. The operating system can be the Android operating system, can be the iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0094] The image processing device provided by the embodiment of the present application Figure 2 shown can implement Figure 1 each process implemented by the method embodiment, and for the sake of brevity, it will not be repeated here.

[0095] Optionally, as Figure 3 shown, the embodiment of the present application further provides an electronic device 300, including a processor 301, a memory 302, a program or instruction stored on the memory 302 and executable on the processor 301. When the program or instruction is executed by the processor 301, it implements each process of the above-mentioned image processing method embodiment, and can achieve the same technical effect, and for the sake of brevity, it will not be repeated here.

[0096] It should be noted that the electronic device in the embodiment of the present application includes the above-mentioned server.

[0097] The embodiment of the present application further provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, it implements each process of the above-mentioned image processing method embodiment, and can achieve the same technical effect, and for the sake of brevity, it will not be repeated here.

[0098] Among them, the processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes computer-readable storage media such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disc, etc.

[0099] Another embodiment of the present application provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement each process of the above embodiment of the image processing method and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0100] It should be understood that the chip mentioned in the embodiments of the present application may also be referred to as a system-on-chip, system chip, chip system, or system-on-chip, etc.

[0101] It should be noted that in this article, 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 limitations, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including that element.

[0102] The above is the preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. An image processing method, characterized in that, The method includes: Segmenting the bone surface of the ultrasound image through a denoising-channel Unet model, including: Performing data preprocessing on the osteoarthritic ultrasound image based on the histogram equalization method; adding Gaussian filtering and bilateral filtering denoising channels to the Skip-layer layer in the Unet neural network model structure to obtain a denoising-channel Unet model; inputting the preprocessed osteoarthritic ultrasound image into the denoising-channel Unet model to obtain the ultrasound bone surface segmentation result; Converting the 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data through an ultrasound and navigation system, including: Obtaining the spatial position and attitude information of the ultrasound probe through the ultrasound and navigation system; combining the spatial position and attitude information with the ultrasound bone surface segmentation result to generate 3D ultrasound bone surface sparse point cloud data; using the Bezier interpolation algorithm for the 3D bone surface sparse point cloud data to generate 3D ultrasound bone surface dense point cloud data; Segmenting the osteoarthritic CT image to obtain CT bone surface point cloud data, including: Segmenting the osteoarthritic CT image using the Unet neural network model to obtain a segmentation result; generating CT bone surface point cloud data based on the segmentation result; Performing ICP point cloud registration on the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data, including: Define the 3D ultrasound bone surface point cloud data as the point cloud to be registered \(P = \{p_1, p_2, \ldots, p\) n \}, and the CT bone surface point cloud data as the target point cloud \(Q = \{q_1, q_2, \ldots, q\) n \). Performing initial registration on the point cloud P to be registered and the target point cloud Q; Registering the point clouds P and Q using the ICP point cloud registration algorithm; The step of registering the point clouds P and Q using the ICP point cloud registration algorithm includes: Establishing the relationship between the point cloud P to be registered and the target point cloud Q as where R is the rotation matrix and T is the translation matrix; Calculating the rotation matrix R and the translation matrix T according to the initial registration relationship between the point cloud P to be registered and the target point cloud Q, and obtaining the transformed new point cloud P'; Iteratively obtaining the rotation matrix R and the translation matrix T until the average distance between the point cloud to be registered and the target point cloud is less than the set threshold or the number of iterations reaches the maximum set number; Obtaining the optimal rotation matrix R and the translation matrix T; Registering the 3D ultrasound bone surface point cloud data and the CT bone surface point cloud data based on the optimal rotation matrix R and the translation matrix T.

2. An image processing apparatus for implementing the image processing method according to claim 1, characterized in that, The device includes: A first segmentation module for segmenting the bone surface of the ultrasound image through a denoising-channel Unet model; The first segmentation module includes: A first sub-module for performing data preprocessing on the osteoarthritic ultrasound image based on the histogram equalization method; A second sub-module for adding Gaussian filtering and bilateral filtering denoising channels to the Skip-layer layer in the Unet neural network model structure to obtain a denoising-channel Unet model; A third sub-module for inputting the preprocessed osteoarthritic ultrasound image into the denoising-channel Unet model to obtain the ultrasound bone surface segmentation result; A conversion module for converting the 2D bone surface data in the ultrasound bone surface segmentation result into 3D ultrasound bone surface point cloud data through an ultrasound and navigation system; The conversion module includes: The fourth sub-module is used to obtain the spatial position and attitude information of the ultrasonic probe through the ultrasonic and navigation systems; The fifth sub-module is used to combine the spatial position and attitude information with the ultrasonic bone surface segmentation result to generate 3D ultrasonic bone surface sparse point cloud data; The sixth sub-module is used to use the Bezier interpolation algorithm for the 3D bone surface sparse point cloud data to generate 3D ultrasonic bone surface dense point cloud data; The second segmentation module is used to segment the osteoarticular CT image to obtain CT bone surface point cloud data; Specifically, the second segmentation module is used to: segment the osteoarticular CT image using the Unet neural network model to obtain a segmentation result; based on the segmentation result, generate CT bone surface point cloud data; The registration module is used to perform ICP point cloud registration on the 3D ultrasonic bone surface point cloud data and the CT bone surface point cloud data; The registration module includes: The seventh sub-module is used to define the 3D ultrasonic bone surface point cloud data as the point cloud to be registered \(P = \{p_1, p_2, \ldots, p\) n \}, and the CT bone surface point cloud data as the target point cloud \(Q = \{q_1, q_2, \ldots, q\) n \}; The eighth sub-module is used to perform initial registration on the point cloud P to be registered and the target point cloud Q; The ninth sub-module is used to register the point clouds P and Q using the ICP point cloud registration algorithm; Specifically, the ninth sub-module is used to: Establish the relationship between the point cloud P to be registered and the target point cloud Q as where R is the rotation matrix and T is the translation matrix; According to the initial registration relationship between the point cloud P to be registered and the target point cloud Q, calculate the rotation matrix R and the translation matrix T, and obtain the new point cloud P' after transformation; Iteratively obtain the rotation matrix R and the translation matrix T until the average distance between the point cloud to be registered and the target point cloud is less than the set threshold or the number of iterations reaches the maximum set number; Obtain the optimal rotation matrix R and the translation matrix T; Based on the optimal rotation matrix R and the translation matrix T, register the 3D ultrasonic bone surface point cloud data and the CT bone surface point cloud data.

3. An electronic device, including a processor, a memory, and a program or instruction stored on the memory and executable on the processor. When the program or instruction is executed by the processor, the steps of the image processing method according to claim 1 are implemented.

Images