Search

CN-122023726-A - Three-dimensional reconstruction processing method for vascular surgery operation navigation image

CN122023726ACN 122023726 ACN122023726 ACN 122023726ACN-122023726-A

Abstract

The invention discloses a three-dimensional reconstruction processing method of a vascular surgical operation navigation image, in particular relates to the technical field of medical image navigation, and is used for solving the problem of topology dislocation of internal and external images under the stress deformation of a blood vessel in operation; extracting an initial vessel center line through a preoperative static three-dimensional image, establishing a space skeleton coordinate system, combining a static vessel wall grid to form an initial mapping reference, extracting a deformed vessel center line according to monitoring data in an isomerism operation, generating a space displacement vector field, recalculating a dynamic vessel wall three-dimensional grid, mapping pixel characteristics of a two-dimensional intra-cavity surface to a dynamic inner surface, and realizing continuous three-dimensional reconstruction and topological alignment of a navigation image.

Inventors

  • ZHANG YEQING

Assignees

  • 核工业总医院

Dates

Publication Date
20260512
Application Date
20260410

Claims (10)

  1. 1. The three-dimensional reconstruction processing method of the vascular surgical operation navigation image is characterized by comprising the following steps of: s1, acquiring a preoperative static three-dimensional image, extracting an initial vessel center line and dispersing to form a space skeleton coordinate system, extracting a static vessel wall grid and establishing an initial mapping reference; S2, acquiring monitoring data in isomerism, extracting a deformed blood vessel center line, respectively extracting an optical analysis quantity and a space morphological analysis quantity from real-time instrument development characteristics and the deformed blood vessel center line, inputting the optical analysis quantity and the space morphological analysis quantity into a pre-training judgment module to obtain a deformed topology credibility coefficient; s3, applying the space displacement vector field to a space skeleton coordinate system, and recalculating the vertex positions of the grid of the static tube wall according to an initial mapping reference to generate a three-dimensional grid of the dynamic tube wall; and S4, reading a two-dimensional intra-cavity surface image corresponding to the deformed blood vessel central line, establishing an orthogonal projection tangent plane in the current node coordinate and current position tangent direction, emitting space rays to the inner surface of the three-dimensional grid of the dynamic tube wall, and mapping the two-dimensional intra-cavity surface pixel characteristics to a hit surface patch.
  2. 2. The three-dimensional reconstruction processing method of a navigation image of vascular surgical operation according to claim 1, wherein the three-dimensional reconstruction processing method is characterized by comprising the following steps: Establishing an initial mapping reference comprises the steps of performing vascular enhancement processing on a preoperative static three-dimensional image, determining an initial vascular center line penetrating through a target vascular segment, forming a space skeleton coordinate system along the initial vascular center line according to an equal arc length sequence, extracting static pipe wall grids, respectively corresponding the vertexes of the static pipe wall grids to space skeleton coordinate nodes with the minimum distance, and recording target binding nodes and relative space vectors.
  3. 3. The three-dimensional reconstruction processing method of a navigation image of vascular surgical operation according to claim 2, wherein the three-dimensional reconstruction processing method is characterized by comprising the following steps: The initial mapping reference further comprises the steps of establishing a local skeleton coordinate base formed by a tangential direction, a local cross section first direction and a local cross section second direction at each space skeleton coordinate node, decomposing each relative space vector into local components along the tangential direction, the local cross section first direction and the local cross section second direction on the local skeleton coordinate base, and storing the local components corresponding to the target binding nodes.
  4. 4. The three-dimensional reconstruction processing method of a navigation image of vascular surgical operation according to claim 1, wherein the three-dimensional reconstruction processing method is characterized by comprising the following steps: Obtaining deformation topological credibility coefficients comprises the steps of extracting real-time instrument development characteristics and deformation vessel center lines from monitoring data in isomerism, forming optical analysis quantities representing development forms in a node local analysis window, forming space form analysis quantities representing path bending states on the deformation vessel center lines, correspondingly writing the optical analysis quantities and the space form analysis quantities into the same node serial numbers, and inputting the optical analysis quantities and the space form analysis quantities into a pre-training judgment module.
  5. 5. The three-dimensional reconstruction processing method for vascular surgical operation navigation images according to claim 4, wherein the three-dimensional reconstruction processing method comprises the following steps: And after the pre-training judging module outputs the deformation topology credibility coefficient, screening according to the deformation topology safety threshold, taking the normalized arc length positions of the initial vessel center line and the deformed vessel center line as the homologous index basis, establishing a one-to-one correspondence between the spatial skeleton coordinate nodes and the deformed vessel center line nodes, and generating a spatial displacement vector field by the coordinate difference of the corresponding nodes.
  6. 6. The three-dimensional reconstruction processing method of a navigation image of vascular surgical operation according to claim 1, wherein the three-dimensional reconstruction processing method is characterized by comprising the following steps: The recalculating the positions of the grid vertices of the static pipe wall comprises the steps of applying a space displacement vector field to a space skeleton coordinate system item by item to obtain an updated space skeleton coordinate system, reconstructing an updated local skeleton coordinate base formed by an updated tangential direction, an updated local cross section first direction and an updated local cross section second direction at each updated space skeleton coordinate node, and recovering the grid vertices of the dynamic pipe wall according to the initial mapping reference.
  7. 7. The three-dimensional reconstruction processing method for vascular surgical operation navigation images according to claim 6, wherein the three-dimensional reconstruction processing method comprises the following steps: Generating the dynamic pipe wall three-dimensional grid comprises the steps of inheriting the triangular patch connection relation of the static pipe wall grid to form dynamic triangular patches, judging the patch orientation and patch area of each dynamic triangular patch, determining a patch degradation threshold value according to the static triangular patch area and the patch shape-preserving coefficient, and executing local vertex outward movement correction on the dynamic triangular patches with areas lower than the patch degradation threshold value to obtain the dynamic pipe wall three-dimensional grid.
  8. 8. The three-dimensional reconstruction processing method of a navigation image of vascular surgical operation according to claim 1, wherein the three-dimensional reconstruction processing method is characterized by comprising the following steps: establishing an orthogonal projection tangent plane comprises the steps of taking the current node coordinate of the deformed blood vessel central line as a plane origin, taking the tangential direction of the current position as a normal direction, establishing a local plane coordinate system in the orthogonal projection tangent plane in the first direction of the updated local section and the second direction of the updated local section, and forming an orthogonal projection tangent plane sampling point according to the transverse single-pixel physical distance and the longitudinal single-pixel physical distance of the two-dimensional intra-cavity tangent plane image.
  9. 9. The three-dimensional reconstruction processing method for vascular surgical operation navigation images according to claim 8, wherein the three-dimensional reconstruction processing method comprises the following steps: The method comprises the steps of transmitting space rays to the inner surface of a three-dimensional grid of a dynamic tube wall, forming space rays by taking current node coordinates as ray starting points aiming at all orthogonal projection tangent plane sampling points except a central sampling point, determining a static local radius upper bound according to the maximum value of the length of a relative space vector bound to a current space skeleton coordinate node, determining a ray searching upper limit threshold by combining radial release coefficients, and selecting a first hit space intersection point nearest to the current node coordinate within the ray searching upper limit threshold.
  10. 10. The three-dimensional reconstruction processing method of a navigation image of vascular surgical operation according to claim 9, wherein: Mapping the two-dimensional intra-cavity surface pixel characteristics to hit patches comprises the steps of firstly executing continuous gray scale values on orthogonal projection surface sampling points in a two-dimensional intra-cavity surface image, writing gray scale values corresponding to the sampling points into the hit patch where the first hit space intersection point is located, forming patch attachment records, recording position parameters of the first hit space intersection point relative to three dynamic pipe wall grid vertexes of the hit patch, and completing dynamic inner wall attachment display according to the patch attachment records.

Description

Three-dimensional reconstruction processing method for vascular surgery operation navigation image Technical Field The invention relates to the technical field of medical image navigation, in particular to an image three-dimensional reconstruction processing method for a navigation scene of endovascular surgery intracavity repair operation. Background In the endovascular repair surgery, in order to realize accurate surgical path navigation, a three-dimensional navigation picture capable of reflecting the overall topology and the local inner cavity state of a blood vessel is generally constructed by simultaneously utilizing high-resolution static three-dimensional structural images acquired before surgery and heterogeneous monitoring data such as two-dimensional angiography, an intracavity ultrasonic cross section and the like acquired in real time during surgery. Most of the existing image fusion methods use a preoperative static three-dimensional model as a fixed shape basis, and then real-time monitoring data in operation is superimposed into the static model through rigid coordinate transformation. Because the blood vessel belongs to flexible tissues, when instruments such as a guide wire, a catheter or a stent conveyor and the like are pushed in a lumen, the curvature, the trend and the local radius of the blood vessel can change in real time along with the stress state, and a preoperative static model can not reflect the deformation in the operation. If the intra-operative two-dimensional inner cavity information is directly mapped onto the preoperative static three-dimensional outer wall in a rigid superposition mode, the problems that the inner and outer images are topologically misplaced, the local textures penetrate through the outer wall boundary, the wall surface display position is wrong and the like easily occur, so that the reliability of intra-operative navigation is affected, and even the judgment of an operator on the position of a blood vessel perforation, an interlayer or an instrument can be misled. Therefore, it is necessary to provide a processing method capable of unifying preoperative static geometric precursors and intraoperative deformation information into the same three-dimensional topological carrier, so as to improve continuity, correspondence and reliability of vascular surgical navigation images. Disclosure of Invention The invention provides a three-dimensional reconstruction processing method for a navigation image of vascular surgery operation, which aims to solve the problem that the preoperative outer wall morphology and the intra-operative inner cavity image are difficult to keep the topological consistency under the condition of flexible deformation of a blood vessel in the prior art. The method comprises the steps of firstly extracting initial vascular center lines from preoperative static three-dimensional images, establishing a space skeleton coordinate system, simultaneously extracting static tube wall grids, establishing an initial mapping reference, recovering deformed vascular center lines corresponding to the current instrument propulsion state from monitoring data in heterogeneous operation, completing deformation reliability judgment at a node level, generating a space displacement vector field when a deformation topological safety threshold is met, then recalculating the space skeleton coordinate system and the static tube wall grids according to the space displacement vector field to obtain dynamic tube wall three-dimensional grids consistent with the current stress state, and finally writing pixel characteristics of two-dimensional intra-cavity surface images into the inner surface of the dynamic tube wall three-dimensional grids to form a continuous intraoperative three-dimensional navigation result. In order to achieve the above purpose, the invention adopts the following technical scheme: s1, acquiring a preoperative static three-dimensional image, extracting an initial vessel center line and dispersing to form a space skeleton coordinate system, extracting a static vessel wall grid and establishing an initial mapping reference; S2, acquiring monitoring data in isomerism, extracting a deformed blood vessel center line, respectively extracting an optical analysis quantity and a space morphological analysis quantity from real-time instrument development characteristics and the deformed blood vessel center line, inputting the optical analysis quantity and the space morphological analysis quantity into a pre-training judgment module to obtain a deformed topology credibility coefficient; s3, applying the space displacement vector field to a space skeleton coordinate system, and recalculating the vertex positions of the grid of the static tube wall according to an initial mapping reference to generate a three-dimensional grid of the dynamic tube wall; and S4, reading a two-dimensional intra-cavity surface image corresponding to the deformed blood vessel centr