CN117274334A - Real-time bone model reconstruction method and system based on point cloud - Google Patents
Real-time bone model reconstruction method and system based on point cloud Download PDFInfo
- Publication number
- CN117274334A CN117274334A CN202311274511.8A CN202311274511A CN117274334A CN 117274334 A CN117274334 A CN 117274334A CN 202311274511 A CN202311274511 A CN 202311274511A CN 117274334 A CN117274334 A CN 117274334A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- bone
- real
- patient
- registration
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 210000000988 bone and bone Anatomy 0.000 title claims abstract description 286
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000000700 radioactive tracer Substances 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 13
- 230000003287 optical effect Effects 0.000 claims description 11
- 239000011159 matrix material Substances 0.000 claims description 8
- 230000009466 transformation Effects 0.000 claims description 8
- 238000005096 rolling process Methods 0.000 claims description 4
- 230000003247 decreasing effect Effects 0.000 claims 1
- 239000000523 sample Substances 0.000 description 10
- 210000004872 soft tissue Anatomy 0.000 description 4
- 230000007423 decrease Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 210000000629 knee joint Anatomy 0.000 description 3
- 210000002303 tibia Anatomy 0.000 description 3
- 210000000689 upper leg Anatomy 0.000 description 3
- 230000005855 radiation Effects 0.000 description 2
- 238000001356 surgical procedure Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000002082 fibula Anatomy 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/35—Determination of transform parameters for the alignment of images, i.e. image registration using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Multimedia (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a point cloud-based real-time bone model reconstruction method and a point cloud-based real-time bone model reconstruction system, wherein the real-time bone model reconstruction method comprises the following steps: s1, acquiring position information of bone key points of a patient, and performing rough registration on the position information and a standard bone model; s2, acquiring points on the surface of a patient bone, combining the points with the previous acquired real-time point cloud of the patient bone every time a set number of points are acquired to acquire a current real-time point cloud of the patient bone, and registering the current real-time point cloud of the patient bone with a standard bone model on the basis of S1 coarse registration; s3, after the collected points on the surface of the bone of the patient meet the set conditions, reconstructing the bone model of the patient according to the corresponding registration results. The invention can register with the standard bone model in real time in the process of point cloud acquisition, can finish three-dimensional reconstruction of the bone model of the patient while finishing bone surface information acquisition, and can directly carry out planning and navigation after the reconstruction is finished, thereby further reducing the operation flow and improving the operation efficiency.
Description
Technical Field
The invention relates to the technical field of medical robots, in particular to a real-time bone model reconstruction method and system based on point cloud.
Background
At present, in joint surgery, especially knee joint surgery, it is generally required to obtain joint image data of a patient, such as CT images, MRI images, and the like. Because CT images have higher resolution and 3D modeling capability, a doctor can conveniently conduct operation planning before operation, and therefore CT images are generally adopted to acquire image information of bones and the like of a patient. And a doctor formulates an operation scheme according to the image information.
When the preoperative operation is planned, the CT image is required to be segmented, so that a mask of a skeleton (comprising femur, tibia, fibula and the like) is obtained, 3D reconstruction is performed through a mask result, and preoperative planning is performed on the reconstructed result, wherein the steps comprise osteotomy position, prosthesis model, prosthesis placement and the like.
At present, domestic manufacturers perform preoperative planning and intra-operative navigation based on CT schemes, but traditional knee joint operation does not need CT images, and the problems of long preoperative CT segmentation time, high patient cost, more radiation, difficult reimbursement and the like exist.
At present, some manufacturers also use an optical tracking system to directly acquire bone surface information during operation, and perform 3D reconstruction on bones in real time through a point cloud registration technology. But requires the whole smearing of the bone surface and then registration. Additionally, waiting time in the doctor operation is increased, and real-time smearing and real-time registration cannot be achieved. In addition, the registration effect cannot be observed in real time in the smearing process, once the smearing deviation occurs, the error correction cost is increased only after registration.
Disclosure of Invention
The invention aims to: aiming at the defects, the invention provides a real-time bone model reconstruction method and a real-time bone model reconstruction system based on point cloud, which greatly reduce the time of an operation flow, improve the operation efficiency and reduce the cost of patients compared with a CT scheme.
The technical scheme is as follows:
a method for reconstructing a real-time bone model based on a point cloud, comprising:
s1, acquiring position information of bone key points of a patient, and performing rough registration on the position information and a standard bone model;
s2, acquiring points on the surface of a patient bone, combining the points with the previous acquired real-time point cloud of the patient bone to acquire a current real-time point cloud of the patient bone, and registering the current real-time point cloud of the patient bone with a standard bone model on the basis of rough registration in S1 so that the standard bone model is attached to the current real-time point cloud of the patient bone;
s3, after the collected points on the surface of the bone of the patient meet the set conditions, reconstructing the bone model of the patient according to the corresponding registration results.
Specifically, the registering includes:
s21, taking the current bone real-time point cloud of the patient as a source point cloud, taking a standard bone model as a target point cloud, and performing fine registration on the basis of S1 coarse registration;
and S22, elastically registering the current patient bone real-time point cloud with a standard bone model on the basis of the S21 fine registration, so that the standard bone model is attached to the current patient bone real-time point cloud.
More specifically, in S2, performing fine registration on the current patient bone real-time point cloud and the standard bone model specifically includes:
the current patient bone real-time point cloud is subjected to fine registration with a standard bone model based on the previous fine registration, wherein the first registration is based on the coarse registration of S1.
More specifically, before elastically registering the current patient bone real-time point cloud with the standard bone model, the method further comprises the steps of: and meshing the current patient bone real-time point cloud with the standard bone model, and elastically registering the meshed current patient bone real-time point cloud with the standard bone model.
More specifically, the elastic registration is: and taking the current bone real-time point cloud of the patient as a target point cloud, taking the standard bone model as a source point cloud, and adopting an NICP algorithm to perform elastic registration on the basis of the fine registration.
Still further, the elastic registration is specifically:
(1) Searching a source point cloud and a target point cloud through a nearest neighbor searching algorithm on the basis of the fine registration;
(2) Calculating the distance between two corresponding points in each pair of nearest neighbors, and judging whether each distance is smaller than a set threshold value;
if the position is smaller than the first threshold value, the registration of the pair of nearest neighbors is considered to be successful, and the corresponding two points are kept unchanged;
otherwise, the transformation matrix between the two corresponding points is decomposed through a Kerisky to obtain a corresponding rotation matrix in the pair of nearest neighbors, and the corresponding points in the source point cloud are subjected to rotation transformation and update the source point cloud, and the iteration times are increased by 1;
(3) Judging whether the iteration times exceeds the set times; if yes, ending iteration to obtain a final updated source point cloud, and finishing the elastic registration; otherwise, searching for the nearest neighbor point between the updated source point cloud and the target point cloud, and returning to the step (2).
Further, in the step (2), after the distance d between the two corresponding points in each pair of nearest neighbors is calculated, a distance coefficient k is added to the distance d, and whether the pair of nearest neighbors is successfully registered is judged by judging whether kd is smaller than a set threshold value.
Further, the value of the distance coefficient k gradually decreases with the number of iterations.
Furthermore, the meshing of the bone real-time point cloud and the bone model of the current patient adopts an Alpha reconstruction method, a poisson reconstruction method or a rolling ball reconstruction method.
A point cloud based real-time bone model reconstruction system, comprising:
the point cloud acquisition unit is provided with a detection tip and at least three non-collinear tracer balls, and is used for carrying out real-time point cloud acquisition by abutting the detection tip on the surface of the bone;
the optical tracking system is used for acquiring the position information of the tracer ball on the point cloud acquisition unit in real time;
the processing unit acquires the position information of the tracer ball on the point cloud acquisition unit, which is obtained by the optical tracking system, and accordingly obtains the position information of the corresponding point on the bone surface according to the position information of the detection tip;
the processing unit is used for acquiring a set number of points at each time of the point cloud acquisition unit, combining the set number of points with the previous acquired real-time point cloud of the patient bone to acquire a current real-time point cloud of the patient bone, registering the current real-time point cloud of the patient bone with the standard bone model, enabling the standard bone model to be attached to the current real-time point cloud of the patient bone, and completing reconstruction of the patient bone model according to the registration result after the points on the surface of the patient bone acquired by the point cloud acquisition unit meet the set condition.
Specifically, the point cloud acquisition unit is abutted against key points of the bones of the patient through the detection tip of the point cloud acquisition unit to acquire bone key points, and the processing unit obtains position information of the corresponding key points through the processing and performs rough registration with a standard bone model according to the position information;
the processing unit is used for registering the current patient bone real-time point cloud with the standard bone model through an ICP algorithm based on the previous registration, wherein the first registration is based on the coarse registration.
More specifically, after each registration, the processing unit elastically registers the current patient bone real-time point cloud with the standard bone model through an NICP algorithm, so that the standard bone model fits the current patient bone real-time point cloud.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention utilizes the optical tracking system to directly collect the bone surface information in the operation, can register with the standard bone model in real time in the collection process, can finish the three-dimensional reconstruction of the bone model of the patient while finishing the bone surface information collection through the real-time registration, greatly reduces the time of the operation flow compared with a CT scheme on the premise of using the operation robot, improves the operation efficiency and reduces the cost of the patient.
2. Each registration of the invention can be based on the previous registration, and the registration result is more accurate, thereby enabling the three-dimensional reconstruction of the bone model of the patient to be more accurate.
3. The invention performs point cloud completion aiming at the bone parts where the point cloud is difficult to collect, and ensures the integrity and the authenticity of the bone point cloud of the patient.
4. Aiming at the situation that some surface points are mismatched or have larger errors between the patient bone point cloud and the standard bone model, or the situation that the standard bone model is not attached to the patient bone point cloud enough, and some points have larger errors, the corresponding points in the patient bone point cloud and the standard bone model are registered, so that the attachment between the standard bone model and the acquired patient bone point cloud is ensured, the accuracy of registration is ensured, and the accuracy of reconstruction of the patient bone model is further improved.
Drawings
FIG. 1 is a flow chart of a point cloud based real-time bone model reconstruction method of the present invention;
fig. 2 is an exemplary diagram of point cloud completion in an embodiment of the present invention, where fig. 2 (a) is an exemplary diagram before point cloud completion and fig. 2 (b) is an exemplary diagram after point cloud completion.
Detailed Description
The invention is further elucidated below in connection with the drawings and the specific embodiments.
The invention discloses a real-time bone model reconstruction method based on point cloud, which is shown in figure 1 and comprises the following steps:
s1, acquiring position information of bone key points of a patient, and performing rough registration on the position information and a standard bone model;
specifically:
s11, clicking the patient bone key points through the probe, and identifying the position information of the tracer ball through the optical tracking system, so that the position information of the patient bone key points is obtained through identification, and the position information of the patient bone key points can be obtained.
In the invention, point cloud information of key points is acquired through a point cloud acquisition unit such as a probe, at least three coplanar and non-collinear tracer balls are arranged on the probe, and the key points of the skeleton of a patient are points with anatomical significance on the skeleton of the patient, such as 3 key points acquired on femur: intercondylar fossa, lateral condyle, and medial condyle; 3 key points on tibia: a proximal medial tibial tangent point, a proximal lateral tibial tangent point, and a tibial tuberosity.
In the present invention, soft tissue may exist at some critical points on the bone of a patient, and information corresponding to the critical points is acquired after the soft tissue is penetrated by a probe.
S12, performing rough registration on the bone of the patient and a standard bone model;
and (3) carrying out rough registration on the patient bone and the standard bone model based on the position information of the patient bone key points obtained in the step (S11) and combining the corresponding points on the standard bone model, so that the acquired patient bone key points are overlapped with the corresponding points on the standard bone model. In the invention, because the corresponding points involved in registration are only key points, the accuracy is not high, so the corresponding points cannot be completely overlapped in general, but the step is only to roughly align the bones of the patient with the standard bone model so as to provide an initial registration basis for the subsequent fine registration; therefore, the coarse registration of the invention only needs to make the acquired bone key points of the patient coincide with the corresponding points on the standard bone model as much as possible, if the set condition is met, for example, the sum or average distance between the distances between each bone key point and the corresponding points on the standard bone model is smaller than the set value. Wherein, the standard bone model adopts an average bone model for setting the number of people within the set age range.
Specifically, the invention can rigidly register the critical points of the patient bone with the corresponding points on the standard bone model through RANSAC (Random Sample Consensus) algorithm.
S2, acquiring points on the surface of the bone of the patient, combining the acquired set number of points with the real-time point cloud of the bone of the patient obtained in the previous time to obtain the real-time point cloud of the bone of the current patient, and registering the real-time point cloud of the bone of the current patient with a standard bone model;
specifically:
s21, acquiring points except key points on the surface of the bone of the patient in real time through a point cloud acquisition unit such as a probe, wherein each acquisition set number of points are combined with the real-time point cloud of the bone of the patient obtained in the previous time to obtain the real-time point cloud of the bone of the current patient;
in the acquisition process, the probe is required to be close to the bone surface, and the probe is always kept to move on the bone surface, so that the acquired point cloud shape is consistent with the actual skeleton.
In the invention, the acquired point cloud needs to be displayed at a corresponding position on bones so that doctors can know the position of the acquired point cloud, and in joint replacement operation or related operations, such as knee joint replacement operation, the density of the acquired points needs to be increased in the area which is similar to the area of the relevant positions of the lateral condyle on femur, the medial condyle on tibia, the upper platform focus and the like and is used as the key area in the planning operation of the osteotomy face in order to plan the osteotomy face more accurately;
therefore, in the invention, the conditions for collecting the point clouds of the key areas are required to be satisfied, and the conditions are as follows:
the method comprises the steps of presetting indexes of points corresponding to different key areas on a standard bone model, carrying out nearest neighbor point search by adopting a nearest neighbor point search algorithm such as a KDTE algorithm on the basis of rough registration of S1 in the process of collecting point clouds except key points on the surface of a patient bone by a point cloud collecting unit such as a probe, searching points closest to the points in the point cloud on the surface of the patient bone in the standard bone model, recording the number of the found points positioned in the key areas, and if the number of the found points positioned in the key areas reaches a set condition, considering that the collection of the points corresponding to the areas on the surface of the patient bone is completed; in the invention, the set condition is that the ratio between the number of found points located in the key area and the number of points corresponding to the key area on the standard skeleton model is larger than the set value, such as 0.9.
Some key areas are located at the boundary between soft tissue and bone or in soft tissue, and cannot be acquired, as shown in fig. 2 (a), so that point cloud complementation is required, and the specific point cloud complementation is as follows:
calculating the point cloud boundary of the point cloud on the surface of the bone of the patient, calculating the distance between each key point acquired in S1 and each boundary point according to the point cloud boundary, setting the interval parameter between each key point and each boundary point, carrying out point cloud complementation in a point cloud model according to the interval parameter, and then splicing with the actual point cloud to form final point cloud data, namely obtaining the bone real-time point cloud of the patient. The invention can calculate the point cloud boundary of the point cloud on the surface of the bone of the patient by adopting algorithms such as a warp-weft scanning method, a grid dividing method, a normal estimation method, alpha shapes and the like, and preferably, the invention adopts a boundary extraction method based on normal estimation of point cloud data.
Taking the medial and lateral condyle positions as an example, after collecting the medial and lateral condyle points, carrying out point cloud completion on the region from the medial and lateral condyle key points to the middle of the point cloud on the surface of the bone, calculating the distance from the medial and lateral condyle points to the point cloud boundary point on the basis of calculating the point cloud boundary of the surface of the bone of the patient, setting interval parameters according to the distance distances, generating new points in the region from the medial and lateral condyle points to the point cloud boundary point according to the interval parameters, and splicing with the actual point cloud to form the final point cloud of the bone of the patient, as shown in fig. 2 (b).
Also, in the point cloud completion process, if the ratio between the number of the completed points in the corresponding region to be completed and the number of the points in the corresponding region on the corresponding standard bone model is greater than a set value, the points in the region are considered to be completed.
In the present invention, the amount is 100 as the set amount.
S22, based on the previous registration, performing fine registration on the real-time point cloud of the current patient bone obtained in the S21 and a standard bone model;
specifically, the real-time point cloud of the current patient bone obtained in S21 is used as a source point cloud, the standard bone model is used as a target point cloud, and the ICP algorithm is adopted for fine registration based on the previous registration. The first registration is based on the rough registration of the step S1, and the previous registration is the fine registration between the real-time point cloud of the patient bone obtained in the previous time and the standard bone model.
In the invention, because the acquired real-time point cloud of the current patient bone is smaller at the beginning and has an excessively large gap compared with the standard bone model, if the standard bone model is used as the source point cloud and the real-time point cloud of the current patient bone is used as the target point cloud for registration according to the conventional ICP algorithm, registration failure can be caused, so that the accuracy of registration is ensured by using the real-time point cloud of the current patient bone as the source point cloud and the standard bone model as the target point cloud.
In the step, the real-time point cloud of the patient bone is registered with the standard bone model, in the process, the real-time point cloud of the patient bone is registered with the whole standard bone model, and the real-time point cloud of the patient bone is aligned and attached as much as possible on the premise of not changing the shape of the point cloud of the patient bone.
And S23, elastically registering the two on the basis of the fine registration of the S22.
In the invention, after the fine registration of S22, the conditions of certain point mismatch or larger error or insufficient fitting of the real-time point cloud of the standard bone model and the patient bone exist, and certain points have larger deviation, so the invention elastically registers the real-time point cloud of the patient bone and the standard bone model on the basis of the fine registration of S2, further enables the standard bone model to fit the real-time point cloud of the patient bone, and further converts the standard bone model into the shape of the actual patient bone through adjustment.
Specifically, the invention adopts NICP algorithm to elastically register the real-time point cloud of the patient bone with the standard bone model, in the invention, the real-time point cloud of the patient bone is used as the target point cloud, the standard bone model is used as the source point cloud, and the invention specifically comprises the following steps:
(1) Searching nearest neighbor points of the real-time point cloud of the bone of the patient and the standard bone model through a nearest neighbor searching algorithm on the basis of S22 fine registration;
(2) Calculating the distance between two corresponding points in each pair of nearest neighbors, and judging whether each distance is smaller than a set threshold value;
if the position is smaller than the first threshold value, the registration of the pair of nearest neighbors is considered to be successful, and the corresponding two points are kept unchanged;
otherwise, the transformation matrix between the corresponding points is decomposed through a Keliski to obtain a corresponding rotation matrix in the pair of nearest neighbors, rotation transformation is carried out on the corresponding points in the source point cloud, the source point cloud is updated, and the iteration times are increased by 1; wherein, in the first iteration, the transformation matrix between the corresponding points is the transformation matrix obtained by S22 fine registration;
(3) Judging whether the iteration times exceeds the set times P; if yes, ending iteration to obtain a final updated source point cloud, and finishing the elastic registration; otherwise, searching for the nearest neighbor point between the updated source point cloud and the target point cloud, and returning to the step (2).
In the invention, in the step (2), in order to prevent the transient registration condition between the real-time point cloud of the patient bone and the standard bone model, after the distance d between the corresponding two points in each pair of nearest neighbors is calculated, a distance coefficient k is added, and whether the pair of nearest neighbors is successfully registered is judged by judging whether kd is smaller than a set threshold value. Further, the value of the distance coefficient k gradually decreases with the number of iterations, e.g., may decrease proportionally or in fixed steps.
S3, after the collected points on the surface of the bone of the patient meet the set conditions, reconstructing the bone model of the patient according to the corresponding elastic registration results;
after the collected points on the surface of the bone of the patient meet the set conditions, namely, the final point cloud of the bone of the patient is obtained, the reconstruction of the bone model of the patient is completed according to the obtained corresponding elastic registration; wherein the points of the surface of the bone of the patient collected meet the set condition may be greater than a set number or a ratio between the number of surface points of the standard bone model is greater than a set value.
Further, prior to elastic registration, the real-time point cloud of the patient's bone may be gridded with a standard bone model, although the two may also be directly elastically registered.
In the invention, the real-time point cloud of the bone of the patient and the standard bone model are gridded by adopting the methods of Alpha reconstruction, poisson reconstruction, rolling ball reconstruction and the like to iterate the points in the real-time point cloud of the bone of the patient and the standard bone model until the whole point cloud is converted into the grid model, and preferably, the invention adopts the rolling ball reconstruction to gridde the real-time point cloud of the bone of the patient and the standard bone model.
The invention also provides a real-time standard skeleton model reconstruction system based on the point cloud, which comprises the following steps:
the point cloud acquisition unit can be a probe, and is provided with a detection tip and at least three non-collinear tracer balls; the point cloud acquisition unit is abutted against the surface of the bone through the detection tip of the point cloud acquisition unit to acquire point cloud;
the optical tracking system acquires the position information of the tracer ball on the point cloud acquisition unit in real time;
the processing unit is used for acquiring the position information of the tracer ball on the point cloud acquisition unit, which is obtained by the optical tracking system, so that the position information of the corresponding detection tip can be obtained through the position information of the tracer ball when the point cloud acquisition unit acquires the point cloud, and the position information of the corresponding point on the surface of the bone can be obtained;
and the processing unit acquires a set number of points at each point cloud acquisition unit, combines the points with the acquired point cloud to acquire real-time point cloud of the patient bone, performs fine registration on the real-time point cloud and the standard bone model, and completes reconstruction of the standard bone model of the patient according to the registration result after the points on the surface of the bone of the patient acquired by the point cloud acquisition unit meet the set condition.
Specifically, the processing unit acquires the position information of the tracer ball on the point cloud acquisition unit, which is obtained by the optical tracking system, so that the position information of the corresponding detection tip can be obtained through the position information of the tracer ball when the point cloud acquisition unit acquires the point cloud, and further the position information of the corresponding point on the surface of the skeleton can be obtained, and further the real-time point cloud of the skeleton of the patient can be obtained.
The method comprises the steps that a point cloud acquisition unit firstly abuts against key points of a patient bone through a detection tip of the point cloud acquisition unit to acquire the key points of the bone, a processing unit obtains position information of the corresponding key points through the processing and performs rough registration with a standard bone model according to the position information, then the point cloud acquisition unit acquires points on the surface of the bone, and each set number of points are acquired and combined with the point cloud acquired before to obtain real-time point cloud of the bone of the patient; wherein, the standard bone model adopts an average standard bone model for setting the number of people within the set age range.
The processing unit performs fine registration on the real-time point cloud of the bone of the patient and the standard bone model through an ICP algorithm based on the previous registration, wherein the first registration is based on coarse registration; and on the basis of the fine registration, elastically registering the real-time point cloud of the patient bone with the standard bone model through an NICP algorithm, so that the real-time point cloud of the patient bone is attached to the standard bone model, and finally, after the acquired points on the surface of the patient bone meet the set conditions, reconstructing the patient bone model based on the elastic registration result.
In the present invention, both the ICP algorithm and the NICP algorithm may be consistent with the standard bone model reconstruction methods described above.
According to the invention, CT images are not required to be acquired for reconstructing the standard bone model of the patient, bone surface information can be directly acquired in an operation through an optical tracking system, and registration with the standard bone model can be carried out in real time in the acquisition process, three-dimensional reconstruction of the bone model of the patient can be completed while the bone surface information acquisition is completed through real-time registration, and planning and navigation are directly carried out after the reconstruction is completed, so that the operation flow is further reduced, the operation efficiency is improved, the time of the patient subjected to radiation is reduced, and the hospitalizing cost of the patient is greatly reduced. And each registration is based on the previous registration, so that the registration result is more accurate, and the three-dimensional reconstruction of the bone model of the patient is more accurate. Meanwhile, in the process of acquiring the real-time point cloud of the patient bone, the invention carries out point cloud complement on the bone part which is difficult to acquire the point cloud, ensures the integrity and the authenticity of the point cloud of the patient bone, and after registration, carries out elastic registration on certain points with larger deviation, namely registration on corresponding points in the real-time point cloud of the standard bone model and the patient bone, ensures the fit between the standard bone model and the acquired real-time point cloud of the patient bone, ensures the accuracy of registration, and further improves the accuracy of reconstruction of the bone model of the patient.
The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to the specific details of the above embodiments, and various equivalent changes (such as number, shape, position, etc.) may be made to the technical solution of the present invention within the scope of the technical concept of the present invention, and these equivalent changes all fall within the scope of the present invention.
Claims (12)
1. The real-time bone model reconstruction method based on the point cloud is characterized by comprising the following steps of:
s1, acquiring position information of bone key points of a patient, and performing rough registration on the position information and a standard bone model;
s2, acquiring points on the surface of a patient bone, combining the points with the previous acquired real-time point cloud of the patient bone to acquire a current real-time point cloud of the patient bone, and registering the current real-time point cloud of the patient bone with a standard bone model on the basis of rough registration in S1 so that the standard bone model is attached to the current real-time point cloud of the patient bone;
s3, after the collected points on the surface of the bone of the patient meet the set conditions, reconstructing the bone model of the patient according to the corresponding registration results.
2. The method of real-time bone model reconstruction according to claim 1, wherein the registration comprises:
s21, taking the current bone real-time point cloud of the patient as a source point cloud, taking a standard bone model as a target point cloud, and performing fine registration on the basis of S1 coarse registration;
and S22, elastically registering the current patient bone real-time point cloud with a standard bone model on the basis of the S21 fine registration, so that the standard bone model is attached to the current patient bone real-time point cloud.
3. The method for reconstructing a real-time bone model according to claim 2, wherein in S2, performing fine registration on the current patient bone real-time point cloud and a standard bone model specifically comprises:
the current patient bone real-time point cloud is subjected to fine registration with a standard bone model based on the previous fine registration, wherein the first registration is based on the coarse registration of S1.
4. The method of real-time bone model reconstruction according to claim 2, further comprising the step of, prior to elastically registering the current patient bone real-time point cloud with a standard bone model: and meshing the current patient bone real-time point cloud with the standard bone model, and elastically registering the meshed current patient bone real-time point cloud with the standard bone model.
5. The method of real-time bone model reconstruction according to any one of claims 2 to 4, wherein the elastic registration is: and taking the current bone real-time point cloud of the patient as a target point cloud, taking the standard bone model as a source point cloud, and adopting an NICP algorithm to perform elastic registration on the basis of the fine registration.
6. The method of real-time bone model reconstruction according to claim 5, wherein the elastic registration is in particular:
(1) Searching a source point cloud and a target point cloud through a nearest neighbor searching algorithm on the basis of the fine registration;
(2) Calculating the distance between two corresponding points in each pair of nearest neighbors, and judging whether each distance is smaller than a set threshold value;
if the position is smaller than the first threshold value, the registration of the pair of nearest neighbors is considered to be successful, and the corresponding two points are kept unchanged;
otherwise, the transformation matrix between the two corresponding points is decomposed through a Kerisky to obtain a corresponding rotation matrix in the pair of nearest neighbors, and the corresponding points in the source point cloud are subjected to rotation transformation and update the source point cloud, and the iteration times are increased by 1;
(3) Judging whether the iteration times exceeds the set times; if yes, ending iteration to obtain a final updated source point cloud, and finishing the elastic registration; otherwise, searching for the nearest neighbor point between the updated source point cloud and the target point cloud, and returning to the step (2).
7. The method according to claim 6, wherein in the step (2), after calculating the distance d between the two corresponding points in each pair of nearest neighbors, a distance coefficient k is added to the distance d, and whether the pair of nearest neighbors is successfully registered is determined by determining whether kd is smaller than a set threshold.
8. The method of claim 7, wherein the value of the distance coefficient k is gradually decreased with the number of iterations.
9. The method according to claim 4, wherein the meshing of the current patient bone real-time point cloud and the bone model adopts an Alpha reconstruction method, a poisson reconstruction method or a rolling ball reconstruction method.
10. A point cloud based real-time bone model reconstruction system, comprising:
the point cloud acquisition unit is provided with a detection tip and at least three non-collinear tracer balls, and is used for carrying out real-time point cloud acquisition by abutting the detection tip on the surface of the bone;
the optical tracking system is used for acquiring the position information of the tracer ball on the point cloud acquisition unit in real time;
the processing unit acquires the position information of the tracer ball on the point cloud acquisition unit, which is obtained by the optical tracking system, and accordingly obtains the position information of the corresponding point on the bone surface according to the position information of the detection tip;
the processing unit is used for acquiring a set number of points at each time of the point cloud acquisition unit, combining the set number of points with the previous acquired real-time point cloud of the patient bone to acquire a current real-time point cloud of the patient bone, registering the current real-time point cloud of the patient bone with the standard bone model, enabling the standard bone model to be attached to the current real-time point cloud of the patient bone, and completing reconstruction of the patient bone model according to the registration result after the points on the surface of the patient bone acquired by the point cloud acquisition unit meet the set condition.
11. The real-time bone model reconstruction system according to claim 10, wherein the point cloud acquisition unit performs bone key point acquisition by abutting the detection tip thereof against key points of the bone of the patient, and the processing unit obtains position information of the corresponding key points by the processing and performs coarse registration with the standard bone model according to the position information;
the processing unit is used for registering the current patient bone real-time point cloud with the standard bone model through an ICP algorithm based on the previous registration, wherein the first registration is based on the coarse registration.
12. The real-time bone model reconstruction system according to claim 11, wherein the processing unit elastically registers the current patient bone real-time point cloud with the standard bone model by means of an NICP algorithm after each registration such that the standard bone model fits the current patient bone real-time point cloud.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311274511.8A CN117274334A (en) | 2023-09-28 | 2023-09-28 | Real-time bone model reconstruction method and system based on point cloud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311274511.8A CN117274334A (en) | 2023-09-28 | 2023-09-28 | Real-time bone model reconstruction method and system based on point cloud |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117274334A true CN117274334A (en) | 2023-12-22 |
Family
ID=89217478
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311274511.8A Pending CN117274334A (en) | 2023-09-28 | 2023-09-28 | Real-time bone model reconstruction method and system based on point cloud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117274334A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117670959A (en) * | 2024-02-01 | 2024-03-08 | 鑫君特(苏州)医疗科技有限公司 | Bone registration device and electronic equipment |
-
2023
- 2023-09-28 CN CN202311274511.8A patent/CN117274334A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117670959A (en) * | 2024-02-01 | 2024-03-08 | 鑫君特(苏州)医疗科技有限公司 | Bone registration device and electronic equipment |
| CN117670959B (en) * | 2024-02-01 | 2024-04-26 | 鑫君特(苏州)医疗科技有限公司 | Bone registration device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112641511B (en) | Joint replacement surgery navigation system and method | |
| US9842394B2 (en) | Detection of anatomical landmarks | |
| CN108765417B (en) | Femur X-ray film generating system and method based on deep learning and digital reconstruction radiographic image | |
| CN107016666B (en) | Image post-processing method for pelvis CT three-dimensional reconstruction based on coordinate system | |
| US20060204067A1 (en) | Determining shaft and femur neck axes and three-dimensional reconstruction | |
| CN113116523B (en) | Orthopedic surgery registration device, terminal equipment and storage medium | |
| CN117274334A (en) | Real-time bone model reconstruction method and system based on point cloud | |
| CN113066109B (en) | Pelvis registration method, pelvis registration device, and pelvis registration system | |
| Zheng et al. | (i) Registration techniques for computer navigation | |
| CN111105385A (en) | Method for processing human body joint data provided based on tomography technology | |
| CN107714078B (en) | Method and system for positioning three-dimensional positions of plants in bones and joints by using structural features | |
| Jaramaz et al. | 2D/3D registration for measurement of implant alignment after total hip replacement | |
| Chintalapani et al. | Statistical atlas based extrapolation of CT data | |
| CN113077499A (en) | Pelvis registration method, pelvis registration device and pelvis registration system | |
| CN113274130A (en) | Markless surgery registration method for optical surgery navigation system | |
| CN115358995A (en) | Full-automatic space registration system based on multi-mode information fusion | |
| Bousigues et al. | 3D reconstruction of the scapula from biplanar X-rays for pose estimation and morphological analysis | |
| CN115137482A (en) | Navigation and positioning data treatment method for joint replacement computer-assisted surgery | |
| CN117274335A (en) | Bone model reconstruction method and system based on point cloud | |
| CN111292306A (en) | Knee joint CT and MR image fusion method and device | |
| Guezou-Philippe et al. | Statistical shape modeling to determine the anterior pelvic plane for total hip arthroplasty | |
| CN114403911B (en) | Ulna proximal end parameter measuring method and device based on CT three-dimensional reconstruction image | |
| US12008760B2 (en) | Systems and methods for estimating the movement of a target using a universal deformation model for anatomic tissue | |
| US20240024035A1 (en) | Preoperative imaging combined with intraoperative navigation before and after alteration of a surgical site to create a composite surgical three dimensional structural dataset | |
| Song et al. | Computer-aided modeling and morphological analysis of hip joint |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |