CN112991479B - Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system - Google Patents
Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system Download PDFInfo
- Publication number
- CN112991479B CN112991479B CN202110244855.9A CN202110244855A CN112991479B CN 112991479 B CN112991479 B CN 112991479B CN 202110244855 A CN202110244855 A CN 202110244855A CN 112991479 B CN112991479 B CN 112991479B
- Authority
- CN
- China
- Prior art keywords
- point
- data
- frame
- dimensional
- scanning
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
- G06T11/006—Inverse problem, transformation from projection-space into object-space, e.g. transform methods, back-projection, algebraic methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2211/00—Image generation
- G06T2211/40—Computed tomography
- G06T2211/428—Real-time
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Algebra (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
- Image Processing (AREA)
Abstract
The invention provides an ultrasonic three-dimensional scanning image reconstruction method and a reconstruction system, wherein the reconstruction method comprises the following steps: step S1, inputting volume data in a polar coordinate space, wherein the volume data is composed of one or more data frames, and the data frames are composed of data of scanning planes which form a certain angle with the axle center, the axis or the axial plane of a three-dimensional scanning track in a Cartesian space; s2, constructing a three-dimensional Cartesian coordinate system, and establishing a mapping relation between the three-dimensional Cartesian coordinate and a polar coordinate; and S3, obtaining the image value of the pixel corresponding to the Cartesian coordinate space through tri-linear interpolation on the data point in the polar coordinate space. By adopting the technical scheme of the invention, the polar coordinate space matrix data index in the memory is reversely pushed according to the Cartesian space three-dimensional scanning track, so that the data in the original memory is reconstructed into the corresponding three-dimensional form during scanning, the method is simple, and the result is more accurate.
Description
Technical Field
The invention relates to the technical field of image three-dimensional reconstruction methods in three-dimensional scanning imaging, in particular to an ultrasonic three-dimensional scanning image reconstruction method and an ultrasonic three-dimensional scanning image reconstruction system.
Background
Non-invasive or minimally invasive intra-cavity images of the human body are important means for clinical pre-operative examination auxiliary diagnosis, intra-operative surgical guidance and postoperative effect monitoring. Existing intracavity imaging technologies, such as EUS (ultrasound endoscope), IVUS (intravascular ultrasound), IVOCT (intravascular optical coherence tomography), ICE (array intracardiac ultrasound), and the like, acquire a plurality of tomographic images, and although a three-dimensional image can be obtained through rotation and pullback, the three-dimensional image is an image of the axial side of the probe, and some other instruments are usually inserted in parallel with the probe during the operation, and the image of the tissue in front of the probe is a clinically better observation and operation area, so that forward-looking scanning imaging, especially forward-looking three-dimensional/four-dimensional real-time scanning imaging of a moving organ, is of great clinical significance.
In the existing products, the "forest ICE" series products of CONAVI company in Canada are a three-dimensional cone image formed by a plurality of conical surfaces with different included angles, wherein a single transducer is arranged at the top end of a rotating shaft center, and the shaft center drives the transducer to rotate during scanning, so that the surface of the transducer deviates from the shaft center by a certain angle, and an image of a conical surface can be obtained after one circle of rotation, and the angle of deviation from the shaft center is continuously changed. The patent US20180158190A1 discusses a method for reconstructing and displaying an image in such a three-dimensional scanning mode. However, as can be seen from the description, the number of scan lines for each cone is the same, so that the cone is smaller at smaller angles of departure, the scan lines are too dense and are repeated.
Another three-dimensional scanning of the forward looking probe is a non-rotational scanning mode. The transducer is also placed on top of the probe shaft, which does not rotate, but rather swings the transducer surface through an angle in the other two directions of the vertical axis. The oscillation in one direction may form a sector, which may be formed when superimposing the oscillation in the other vertical direction. The advantage of such scanning is that no rotation of the probe shaft is required, thus making the signal transmission path simpler (omitting modular components such as slip rings or resolvers).
In addition, as the data volume of the three-dimensional volume data is much larger than that of the two-dimensional image, particularly for organs such as a heart, the real-time 3D image (4D image) of the organ is obtained, the processing frame rate requirement on the three-dimensional data is higher, and the three-dimensional reconstruction is a module with larger calculation amount. In addition, for high-quality volume rendering techniques since the high-quality display of three-dimensional data, advanced illumination and rendering functions can be involved, which is also a computationally intensive link. In addition, typically allowing a user to perform a particular display process on a particular region of interest (ROI) may still increase the computational effort of the reconstruction and display process.
Disclosure of Invention
Aiming at the technical problems, the invention discloses an ultrasonic three-dimensional scanning image reconstruction method and a reconstruction system, wherein the method is simple and easy to implement, and can be used for three-dimensional scanning reconstruction under a motion state; and the GPU can achieve higher frame rate and speed up data processing, and has important significance for living body imaging of moving organs (such as hearts).
In this regard, the invention adopts the following technical scheme:
an ultrasonic three-dimensional scanning image reconstruction method comprises the following steps:
step S1, inputting volume data in a polar coordinate space, wherein the volume data is composed of one or more data frames, and the data frames are composed of data of scanning planes which form a certain angle with the axle center, the axis or the axial plane of a three-dimensional scanning track in a Cartesian space;
s2, constructing a three-dimensional Cartesian coordinate system, and establishing a mapping relation between the three-dimensional Cartesian coordinate and a polar coordinate;
and S3, obtaining the image value of the pixel corresponding to the Cartesian coordinate space through tri-linear interpolation on the data point in the polar coordinate space.
As a further improvement of the present invention, in step S1, the data point index of the volume data of the polar coordinate space volume includes a frame number, a line number, a point number; wherein, the data frame is composed of a plurality of data scanning lines, and the scanning lines are composed of a plurality of data points.
As a further improvement of the present invention, in step S2, when the mapping relationship between the three-dimensional cartesian coordinates and the polar coordinates is established, the frame number, line number and point number index of the data point of the volume data in the polar coordinate space corresponding to any pixel point in the cartesian space are deduced, and all points in all the cartesian space are traversed to form a set of spatial coordinate mapping relationships.
As a further improvement of the invention, in step S3, the frame number, line number and point number index values of the polar coordinate space obtained by the mapping in step S2 are rounded up and down respectively to obtain eight groups of index values of any point, corresponding to eight data points in the polar coordinate space, and then the image values of the pixels corresponding to the Cartesian coordinate space are obtained by tri-linear interpolation.
As a further improvement of the invention, when the ultrasonic three-dimensional scanning adopts an axial rotation conical body scanning mode, each scanning conical surface forms a frame of data frame, the volume data comprises n frames of data frames, the deviation angle with the maximum rotation axis in the scanning conical surface is phi, the corresponding angle of each frame is changed at equal intervals, the scanning lines of each frame are different, the scanning lines of the ith frame are nLine (i), the data frames are arrayed at equal angles to form a 360-degree complete conical surface, the number of data points of each line is nPoint, the length, the width and the height of the three-dimensional data matrix after reconstruction are L, W and H respectively, and the transducer after reconstruction is positioned at (x 0, y0 and z 0), wherein z0=0;
the dot pitch on each scan line is: h_p=h/(nPoint-1);
the angle interval between adjacent lines of the conical surface of the ith frame data frame is: line_θi=2 pi/(nLine (i) -1);
cone angle of adjacent frames: frame_Φ= (Φ/2)/(n-1);
the original memory frame, line and point number indexes of any point P (xp, yp, zp) in the three-dimensional space after reconstruction are as follows:
the OP distance is:
the point number where the P point is located is: nPoint_P= (dist_OP)/(h_p) +1 (1);
the included angle between the straight line OP and the straight line OO' is as follows: angle_op_oo '=arccos ((zp-Z0)/dist_op), wherein point O' is the projection point of the O point at the same Z-direction height as the P point;
the frame number where the P point is located is: nframe_p= (angle_op_oo')/(frame_Φ) +1 (2);
the included angle θ_p between the scan line projection O 'P where the P point is located and the initial scan line projection (i.e., the angle rotated by P') is:
point P, line number: nLine_P=θ_P/(line_θi) +1 (3);
the indexes of frame numbers, line numbers and point numbers corresponding to any point in the Cartesian three-dimensional space can be obtained by the formulas (1), (2) and (3), and the image forms corresponding to the scanning tracks are reconstructed in the Cartesian three-dimensional space according to the indexes.
As a further improvement of the invention, when the ultrasonic three-dimensional scanning adopts a non-pivoting bidirectional sector cone scanning mode, each scanning cone forms a frame of data frame, the volume data comprises n frames of data frames, the central angle of a sector of a scanning frame is alpha, the included angle between a first frame of scanning sector and a last frame of scanning sector is beta, if the data in the memory are stored according to the alpha angle sector as a frame, the corresponding beta direction angle of each frame is changed at equal intervals, the scanning line number of each frame is nLine, and the number of data points of each line is nPoint; the length, width and height of the reconstructed three-dimensional data matrix are L, W and H respectively, and the reconstructed transducer is positioned at (x 0, y0, z 0), wherein z0=0;
the angle between the sectors of adjacent data frames is: frame_β= (β)/(n-1);
adjacent line angular spacing in each sector of a data frame: line_α=α/(nlie-1);
the point spacing on each scanning line is as follows: h_p=h/(nPoint-1);
the original memory frame, line and point number indexes of any point P (xp, yp, zp) in the three-dimensional space after reconstruction are as follows:
the OP distance is:
the point number of the P point is: nPoint_P= (dist_OP)/(h_p) +1 (5);
the point O 'is the projection point of the O point at the bottom of d, OO' is the body axis of the body data, OQ is the frame axis of the frame where the Q point P point is located,
the angle between the frame axis OQ and the body axis oo″: Δβ=atan ((yp-y 0)/(zq-z 0));
the frame number of the P point is: nframe_p= (Δβ)/(frame_β) +1 (6);
the angle between the frame axis OQ and the scan line OP: Δα=atan ((xp-x 0)/(zq-z 0)));
the line number of the P point is: nLine_P= Δα/(line_α) +1 (7);
the positions of memory data points corresponding to any point in the Cartesian three-dimensional space are obtained through formulas (5), (6) and (7), the indexes of the frame numbers, the line numbers and the point numbers are included, and the image forms corresponding to the scanning tracks are reconstructed in the Cartesian three-dimensional space according to the indexes.
As a further improvement of the invention, in step S3, the process is accelerated by the GPU while traversing all points of all cartesian space.
Adopting CUDA programming to allocate W.L.H threads, wherein one allocation mode of the threads is thread (32, 32), block (ceil (W/32), ceil (L/32), H; ensuring that each thread handles the index of one point.
As a further improvement of the invention, the original data is used as a three-dimensional texture, the coordinate index is manufactured into a lookup table which is stored in the GPU in advance as a two-dimensional texture, the indexing process is accelerated, and only texture pickup and tri-linear interpolation calculation are performed in the thread.
The invention also discloses an ultrasonic three-dimensional scanning image reconstruction system, which comprises:
the system comprises a polar coordinate space volume data input module, a display module and a display module, wherein the polar coordinate space volume data input module is used for inputting volume data in a polar coordinate space, the volume data is composed of one or more data frames, and the data frames are composed of data of scanning planes which form a certain angle with the axle center, the axis or the axial plane of a three-dimensional scanning track in a Cartesian space;
the three-dimensional Cartesian coordinate and polar coordinate mapping module is used for establishing a mapping relation between the three-dimensional Cartesian coordinate and the polar coordinate;
and the interpolation module is used for obtaining the image value of the pixel corresponding to the Cartesian coordinate space through the tri-linear interpolation on the data points in the polar coordinate space.
As a further improvement of the present invention, the data point index of the volume data of the polar coordinate space volume includes a frame number, a line number, a point number;
when the mapping relation between the three-dimensional Cartesian coordinates and the polar coordinates is established, frame numbers, line numbers and point number indexes of data points of volume data in the polar coordinate space corresponding to any pixel point in the Cartesian space are deduced, and all points in all the Cartesian space are traversed to form a group of space coordinate mapping relation.
As a further improvement of the invention, the interpolation module respectively rounds up and down the frame number, the line number and the point number index values of the polar coordinate space obtained by the three-dimensional Cartesian coordinate and polar coordinate mapping module to obtain eight groups of index values of any point, corresponding to eight data points in the polar coordinate space, and then obtains the image value of the pixel corresponding to the Cartesian coordinate space through tri-linear interpolation.
Compared with the prior art, the invention has the beneficial effects that:
by adopting the technical scheme of the invention, the polar coordinate space matrix data index in the memory is reversely pushed according to the Cartesian space three-dimensional scanning track, so that the data in the original memory is reconstructed into the corresponding three-dimensional form during scanning, and the method is simple. The method can be suitable for imaging of moving organs, such as the field of intracardiac ultrasonic imaging, and because of heart beating (60-100 times per minute), the technical scheme of the invention actually shows the effect of 4D, namely at least one three-dimensional scanning and reconstruction are completed under the condition that the moving organs are relatively static, for example, the intracardiac ultrasonic imaging requires the completion of real-time scanning and three-dimensional reconstruction of partial areas in the human intracardiac cavity within 100ms, wherein the related scanning is a high-speed intracardiac ultrasonic three-dimensional scanning mode, and the organs can be approximately considered to be not moved within the scanning time (within 100 ms), and the three-dimensional scanning under the relatively static state is considered within a short time, so that the result is more accurate. Further, the index transformation relationship can be fixed as a texture cache of the GPU, and the GPU is used for accelerating the three-dimensional reconstruction process.
Drawings
Fig. 1 is a flow chart of an ultrasound three-dimensional scanning image reconstruction method of the present invention.
Fig. 2 is a schematic diagram of a scan trajectory and data storage of the pivoting cone scan method according to embodiment 1 of the present invention. Fig. 2 (a) is a schematic diagram of a trajectory of a cartesian space scan, and fig. 2 (b) is a schematic diagram of a storage form of corresponding acquired data in a polar coordinate space.
FIG. 3 is a schematic diagram of the scan trajectory and data storage of a non-pivoting bi-directional sector cone scan mode of example 2 of the present invention. Fig. 3 (a) is a schematic diagram of a trajectory of a cartesian space scan, and fig. 3 (b) is a schematic diagram of a storage form of corresponding acquired data in a polar coordinate space.
Fig. 4 is a diagram illustrating the coordinate relationship of the three-dimensional scan trajectory in cartesian space in embodiment 1 of the invention. Wherein fig. 4 (a) is a constructed cartesian three-dimensional coordinate; FIG. 4 (b) is a diagram of the projected geometry of the origin of the Z-direction volume data, at any point in the volume; fig. 4 (c) is a projection geometry diagram of Y-direction volume data axis and arbitrary points.
FIG. 5 is a graph of the coordinate relationship of eight adjacent points used for any point in example 1 of the present invention; fig. 5 (a) is a position diagram of the P point, and fig. 5 (b) is a coordinate relationship diagram of the P point and the adjacent eight points.
Fig. 6 is a diagram illustrating the coordinate relationship of the three-dimensional scan trajectory in cartesian space in embodiment 2 of the invention. Wherein fig. 6 (a) is a constructed cartesian three-dimensional coordinate; FIG. 6 (b) is a diagram of the projected geometry of the origin of the Z-direction volume data, at any point in the volume; fig. 6 (c) is a projection geometry diagram of Y-direction volume data axis and arbitrary points.
Fig. 7 is a graph showing the coordinate relationship of eight adjacent points used for an arbitrary point in embodiment 2 of the present invention. Fig. 7 (a) is a position diagram of the P point, and fig. 7 (b) is a coordinate relationship diagram of the P point and the adjacent eight points.
Detailed Description
Preferred embodiments of the present invention are described in further detail below.
As shown in fig. 1, a method for reconstructing an ultrasonic three-dimensional scan image includes the steps of:
step S1, input of volume data is performed in a polar coordinate space.
The volume data is comprised of one or more data frames, the data frames being comprised of a plurality of data scan lines, the scan lines being comprised of a plurality of data points. The indexes of specific points constituting the polar coordinate space volume data are frame numbers, line numbers and point numbers. The data frame is defined by data for each scan plane that forms an angle with the axis, axis or axial plane of the three-dimensional scan trajectory in cartesian space.
S2, constructing a three-dimensional Cartesian coordinate system, and establishing a mapping relation between the three-dimensional Cartesian coordinate and a polar coordinate; and constructing Cartesian space coordinates of the three-dimensional scanning track, deducing frame numbers, line numbers and point number indexes of data points of volume data in polar coordinate space corresponding to any pixel point of the Cartesian space, traversing all points of all the Cartesian space, and forming a group of space coordinate mapping relations.
And S3, obtaining the image value of the pixel corresponding to the Cartesian coordinate space through tri-linear interpolation on the data point in the polar coordinate space. The frame number, line number and point number indexes of the polar coordinate space obtained in the step S2 are often not integer, eight groups of index values are obtained by respectively rounding up and down the frame number, line number and point number index values, eight data points in the corresponding polar coordinate space are obtained, and the image values of pixels corresponding to the Cartesian coordinate space are obtained by tri-linear interpolation.
Further, in step S3, the GPU is used to accelerate the process while traversing all points of all cartesian spaces. Even if CUDA programming is used, w×l×h threads are allocated, one allocation manner of threads is thread (32, 32), block (ceil (W/32), ceil (L/32), H); ensuring that each thread handles the index of one point.
Furthermore, the original data is used as a three-dimensional texture, the coordinate index is manufactured into a lookup table and is stored in the GPU in advance to be used as a two-dimensional texture, the indexing process is accelerated, and only texture pickup and tri-linear interpolation calculation are performed in the thread.
The invention also discloses an ultrasonic three-dimensional scanning image reconstruction system, which comprises:
and a polar coordinate space volume data input module. The volume data is comprised of one or more data frames, the data frames being comprised of a plurality of data scan lines, the scan lines being comprised of a plurality of data points. The indexes of specific points constituting the polar coordinate space volume data are frame numbers, line numbers and point numbers. The data frame is defined by data for each scan plane forming an angle with the axis (axis or axial plane) of the three-dimensional scan trajectory in cartesian space.
And the three-dimensional Cartesian coordinate and polar coordinate mapping module. And constructing Cartesian space coordinates of the three-dimensional scanning track, and deducing frame numbers, line numbers and point number indexes of data points of volume data in polar coordinate space corresponding to any pixel point of the Cartesian space. And traversing all points of all Cartesian spaces to form a set of space coordinate mapping relations.
And a neighborhood eight-point tri-linear interpolation module. The frame number, line number and point number indexes of the polar coordinate space obtained by the three-dimensional Cartesian coordinate and polar coordinate mapping module are often not integer, eight groups of index values are obtained by respectively rounding up and down the frame number, line number and point number index values, eight data points in the corresponding polar coordinate space are obtained, and the image values of pixels corresponding to the Cartesian coordinate space are obtained by tri-linear interpolation.
Ultrasound three-dimensional scanning generally requires obtaining volume data of a complete three-dimensional spatial scan, and for each transmission and reception, a line of data is formed in three-dimensional space that expresses an ultrasound echo signal within a depth interval. The three-dimensional geometrical data is constructed by recording a scanning track formed by the gestures which are transmitted and received each time by the probe in the three-dimensional space and acquiring the obtained three-dimensional matrix data so as to restore the three-dimensional geometrical form of the imaging target.
For three-dimensional scanning of a forward looking probe, there are two common ways: a method for scanning conical surface by rotating it around axle features that the transducer is arranged at the top of probe, and when it is scanned, the axle center drives the transducer to rotate, resulting in a surface of transducer being deviated by a certain angle from axle center. Another approach, which is a non-rotating bi-directional angular sweep, is to place the transducer on top of the probe shaft, which does not rotate, but rather swings the transducer surface through an angle in the other two directions of the vertical axis. The oscillation in one direction may form a sector and when superimposed with the oscillation in the other vertical direction, a sector is formed, the scanning trajectory of which is shown in fig. 3 (a).
For data acquisition organizations, data for one scan plane (cone or sector) typically constitutes one frame. The arrangement in the memory is shown in fig. 2 (b) and fig. 3 (b), respectively. On the single frame image, some preliminary processing such as noise reduction, enhancement, etc. can be performed. And the data in the memory are arranged according to the definite frame number, line number and point number.
The needle is further described below in connection with specific examples.
Example 1
As shown in fig. 2, the ultrasonic three-dimensional scanning adopts a pivoting conical body scanning mode, and as shown in fig. 2, the three-dimensional reconstruction calculation process is as follows:
first, a cone-shaped body data is formed by a plurality of scan cone frames in an axial rotation scan mode, n scan cones are assumed to be used for scanning the body data, the deviation angle of the cone surface and the rotation axis with the maximum value is phi, the corresponding angle of each frame is changed at equal intervals (space sampling is kept uniform), the number of scan lines of each frame is different, the number of scan lines of the ith frame is nLine (i), the complete cone surface of 360 degrees is formed by equiangular arrangement, the number of data points of each line is fixed, and nPoint is set.
Next, in the three-dimensional cartesian coordinate system, as shown in fig. 4 (a), three directions of X, Y and Z indicate three-dimensional coordinates of any data point in space, and the start point coordinates are located at the upper left front corners (0, 0). Assuming that the reconstructed three-dimensional data matrix is L, W and H long, wide, and high, respectively, the transducer is located at (x 0, y0, z 0) after reconstruction, where z0=0, only the far field useful portion is shown.
The point spacing on each scanning line is as follows: h_p=h/(nPoint-1);
adjacent line angular spacing of cone i (i-th frame): line_θi=2 pi/(nLine (i) -1);
adjacent cone (adjacent frame) angle: frame_Φ= (Φ/2)/(n-1);
then, for any point P (xp, yp, zp) in the three-dimensional space after reconstruction, as shown in fig. 4 (a), the original memory frame, line, point index can be deduced as follows:
the OP distance is:
the P point is the point number: nPoint_P= (dist_OP)/(h_p) +1 (1);
in fig. 4 (c), the point O 'is a projection point of the point O at a uniform height (Z direction) with the point P, and the angle between the straight line OP and the straight line OO' is: angle_op_oo' =arccos ((zp-z 0)/dist_op);
the P point is located at the frame number: nframe_p= (angle_op_oo')/(frame_Φ) +1 (2);
in fig. 4 b, the point O' is a projection point of the point O on the same height (Z direction) as the point P, and the projection of the frame where the point P is located on the height plane is a circle (the cone is projected as a circle on a certain cross section). Assume that the start scan line scan order in the frame where the P point is located is as shown in the middle head of the figure. The included angle θ_p between the scan line projection O 'P of the P point and the initial scan line projection (i.e. the angle rotated by P') is:
point P, line number: nLine_P=θ_P/(line_θi) +1 (3)
From the formulas (1), (2) and (3), the position index (frame number, line number and point number) of the memory data point corresponding to any point in the cartesian three-dimensional space can be obtained. The image form corresponding to the scanning track can be reconstructed in the Cartesian three-dimensional space.
In addition, the frame number, line number, point number index are not always integers, but instead, the index of the integer near the calculated value is generally obtained, and the coordinate relation diagram of the adjacent eight points used by any point is shown in fig. 5, wherein the index of the integer position of any point P forms a hexahedral vertex by 8 points 0 、P 1 、P 2 、P 3 、P 4 、P 5 、P 6 、P 7 The frame number and the line number corresponding to the point are respectively as follows:
P 0 ：Frame(i+1)，Line(Φi,i)，Point(i)；
P 1 ：Frame(i+1)，Line(Φi,i+1)，Point(i)；
P 2 ：Frame(i+1)，Line(Φi,i+1)，Point(i+1)；
P 3 ：Frame(i+1)，Line(Φi,i)，Point(i+1)；
P 4 ：Frame(i)，Line(Φj,j)，Point(i)；
P 5 ：Frame(i)，Line(Φj,j+1)，Point(i)；
P 6 ：Frame(i)，Line(Φj,j+1)，Point(i+1)；
P 7 ：Frame(i)，Line(Φj,j)，Point(i+1)。
to obtain the image value V (P) at point P, it can be obtained by using tri-linear interpolation through its neighboring 8 points, specifically as follows:
V(P)＝{(weigth_line(i)*V(P0)+weigth_line(i+1)*V(P1))*weigth_point(i)+(weigth_line(i)*V(P3)+weigth_line(i+1)*V(P2))*weigth_point(i+1)}*weigth_frame(i)+{(weigth_line(j)*V(P4)+weigth_line(j+1)*V(P5))*weigth_point(i)+(weigth_line()*V(P7)+weigth_line(j+1)*V(P6))*weigth_point(i+1)}*weigth_frame(i+1)；(4)
wherein:
the weight_line (i) is the line weight (contribution) occupied by the line number where P0 is located in the Frame (i); weigth_line (i) +weigth_line (i+1) =1.0;
the weight_line (j) is the line weight (contribution) occupied by the line number where P4 is located in Frame (i+1); weigth_line (j) +weigth_line (j+1) =1.0;
similarly, the weight_point (i) and the weight_point (i+1) are point weights corresponding to the vertexes, and the weight_point (i) +the weight_point (i+1) =1.0;
the weight_frame (i) and the weight_frame (i+1) are frame weights corresponding to the vertices, and weight_frame (i) +weight_frame (i+1) =1.0.
Example 2
As shown in fig. 3, the ultrasonic three-dimensional scanning adopts a non-pivoting bidirectional sector cone scanning mode, and as shown in fig. 3, the three-dimensional reconstruction calculation process is as follows:
firstly, a bidirectional sector scanning mode forms one piece of data by a plurality of scanned sector frames, the scanning forming one piece of data is assumed to have n frames, the central angle of each sector of the scanning is alpha, the included angle between the first frame of scanning sector and the last frame of scanning sector is beta, the data in the memory is assumed to be stored according to the alpha angle sector as one frame (according to the equivalent calculation process of storing the beta angle sector as one frame), the beta direction angle corresponding to each frame is changed at equal intervals (the spatial sampling is kept uniform), the scanning line number of each frame is set to be nLine, the data point number of each line is fixed, and nPoint is set.
Next, in the three-dimensional cartesian coordinate system space, three-dimensional spatial arrangement of the reconstructed volume data is shown in fig. 6 (a), three directions of XYZ indicate three-dimensional coordinates of any data point in the space, and the start point coordinates are located at the upper left front corners (0, 0). Assuming that the reconstructed three-dimensional data matrix is L, W and H long, wide, and high, respectively, the transducer is located at (x 0, y0, z 0) after reconstruction, where z0=0, only the far field useful portion is shown.
Adjacent sector (adjacent frame) angle: frame_β= (β)/(n-1);
adjacent line angular spacing in each sector: line_α=α/(nlie-1);
the point spacing on each scanning line is as follows: h_p=h/(nPoint-1);
then, for any point P (xp, yp, zp) in the three-dimensional space after reconstruction, the original memory frame, line, point number index can be deduced as follows:
the OP distance is:
the P point is the point number: nPoint_P= (dist_OP)/(h_p) +1; (5)
In fig. 6 (c), point O "is a projection point of O at the bottom of d, OO" constitutes the axis of the volume data (the volume axis), and Q point P is the axis of the frame (the frame axis), and the projection relationship in the pitch direction in fig. 6 (b) is:
the angle between the frame axis OQ and the body axis oo″: Δβ=atan ((yp-y 0)/(zq-z 0));
the P point is located at the frame number: nframe_p= (Δβ)/(frame_β) +1; (6)
The angle between the frame axis OQ and the scan line OP: Δα=atan ((xp-x 0)/(zq-z 0)));
point P, line number: nLine_P= Δα/(line_α) +1; (7)
The position index (frame number, line number, point number) of the memory data point corresponding to any point in the Cartesian three-dimensional space can be obtained according to the formulas (5), (6) and (7). The image form corresponding to the scanning track can be reconstructed in the Cartesian three-dimensional space.
The frame number, line number and point number index are not integers, and the coordinate relation diagram of eight adjacent points used by any point is shown in FIG. 7, wherein 8 points are combined into a hexahedral vertex for the integer position index of any point P, wherein P is calculated by substituting the integer index near the calculated value 0 、P 1 、P 2 、P 3 、P 4 、P 5 、P 6 、P 7 The frame number and the line number corresponding to the point are respectively as follows:
P 0 ：Frame(i)，Line(i)，Point(i)；
P 1 ：Frame(i+1)，Line(i)，Point(i)；
P 2 ：Frame(i+1)，Line(i+1)，Point(i)；
P 3 ：Frame(i)，Line(i+1)，Point(i)；
P 4 ：Frame(i)，Line(i)，Point(i+1)；
P 5 ：Frame(i+1)，Line(i)，Point(i+1)；
P 6 ：Frame(i+1)，Line(i+1)，Point(i+1)；
P 7 ：Frame(i)，Line(i+1)，Point(i+1)。
to obtain the image value V (P) at the P point, the image value of the P point can be obtained by using the three-linear interpolation method (same as the formula (4)) through the 8 points adjacent thereto.
In addition to the embodiments of the present invention, for all points w×l×h in cartesian space to traverse computations, the computation with the CPU is time consuming, and the process can be accelerated with the GPU. Adopting CUDA programming to allocate W.L.H threads, wherein one allocation mode of the threads is thread (32, 32), block (ceil (W/32), ceil (L/32), H; ensuring that each thread handles the index of one point.
Furthermore, in order to accelerate the indexing speed, the original data is used as a three-dimensional texture, the coordinate indexes from the polar coordinate space (scanning storage data) to the Cartesian space (imaging) are made into a lookup table, the lookup table is stored in the GPU in advance to be used as a two-dimensional texture, the indexing process is accelerated, and only texture pickup and tri-linear interpolation calculation are performed in the thread.
The foregoing is a further detailed description of the invention in connection with the preferred embodiments, and it is not intended that the invention be limited to the specific embodiments described. It will be apparent to those skilled in the art that several simple deductions or substitutions may be made without departing from the spirit of the invention, and these should be considered to be within the scope of the invention.
Claims (9)
1. The ultrasonic three-dimensional scanning image reconstruction method is characterized by comprising the following steps of:
step S1, inputting volume data in a polar coordinate space, wherein the volume data is composed of one or more data frames, and the data frames are composed of data of scanning planes which form a certain angle with the axle center, the axis or the axial plane of a three-dimensional scanning track in a Cartesian space;
s2, constructing a three-dimensional Cartesian coordinate system, and establishing a mapping relation between the three-dimensional Cartesian coordinate and a polar coordinate;
step S3, obtaining image values of pixels corresponding to the Cartesian coordinate space through tri-linear interpolation on data points in the polar coordinate space;
wherein the data frame is composed of a plurality of data scanning lines, and the scanning lines are composed of a plurality of data points;
the ultrasonic three-dimensional scanning adopts a conical scanning mode with pivoting or a bidirectional sector conical scanning mode with non-pivoting, and each scanning conical surface forms a frame of data frame;
in step S1, the data point index of the volume data of the polar coordinate space volume includes a frame number, a line number, and a point number;
in step S2, when the mapping relationship between the three-dimensional cartesian coordinates and the polar coordinates is established, the frame number, line number and point number indexes where the data points of the volume data in the polar coordinate space corresponding to any pixel point in the cartesian space are located are deduced, and all points in all the cartesian space are traversed to form a set of spatial coordinate mapping relationship.
2. The ultrasound three-dimensional scan image reconstruction method according to claim 1, wherein: in step S3, the frame number, line number, and point number index values of the polar coordinate space obtained by the mapping in step S2 are rounded up and down respectively to obtain eight groups of index values of any point, corresponding to eight data points in the polar coordinate space, and then the image values of the pixels corresponding to the cartesian coordinate space are obtained by tri-linear interpolation.
3. The ultrasound three-dimensional scan image reconstruction method according to claim 2, wherein: when an ultrasonic three-dimensional scanning mode adopts an axial rotation conical body scanning mode, the body data comprise n frames of data frames, the maximum off angle between the scanning conical surface and the rotation axis is phi, the angle corresponding to each frame is changed at equal intervals, the number of scanning lines of the ith frame is nLine (i), the data frames are arrayed at equal angles to form a complete conical surface of 360 degrees, the number of data points of each line is nPoint, the length, the width and the height of a three-dimensional data matrix after reconstruction are L, W and H respectively, and a transducer after reconstruction is positioned at (x 0, y0 and z 0), wherein z0=0;
the dot pitch on each scan line is: h_p=h/(nPoint-1);
the angle interval between adjacent lines of the conical surface of the ith frame data frame is: line_θi=2 pi/(nLine (i) -1);
cone angle of adjacent frames: frame_Φ= (Φ/2)/(n-1);
the original memory frame, line and point number indexes of any point P (xp, yp, zp) in the three-dimensional space after reconstruction are as follows:
the OP distance is:
the point number where the P point is located is: nPoint_P= (dist_OP)/(h_p) +1 (1);
the included angle between the straight line OP and the straight line OO' is as follows: angle_op_oo '=arccos ((zp-Z0)/dist_op), wherein point O' is the projection point of the O point at the same Z-direction height as the P point;
the frame number where the P point is located is: nframe_p= (angle_op_oo')/(frame_Φ) +1 (2);
the included angle θ_p between the scan line projection O' P where the P point is located and the initial scan line projection is:
point P, line number: nLine_P=θ_P/(line_θi) +1 (3);
the indexes of frame numbers, line numbers and point numbers corresponding to any point in the Cartesian three-dimensional space can be obtained by the formulas (1), (2) and (3), and the image forms corresponding to the scanning tracks are reconstructed in the Cartesian three-dimensional space according to the indexes.
4. The ultrasound three-dimensional scan image reconstruction method according to claim 2, wherein: when the ultrasonic three-dimensional scanning adopts a non-pivoting bidirectional sector cone scanning mode, the volume data comprises n frames of data frames, the central angle of a sector for scanning one frame is alpha, the included angle between a first frame of scanning sector and a last frame of scanning sector is beta, if the data in the memory are stored according to the alpha angle sector as one frame, the corresponding beta direction angle of each frame is changed at equal intervals, the number of scanning lines of each frame is nLine, and the number of data points of each line is nPoint; the length, width and height of the reconstructed three-dimensional data matrix are L, W and H respectively, and the reconstructed transducer is positioned at (x 0, y0, z 0), wherein z0=0;
the angle between the sectors of adjacent data frames is: frame_β= (β)/(n-1);
adjacent line angular spacing in each sector of a data frame: line_α=α/(nlie-1);
the point spacing on each scanning line is as follows: h_p=h/(nPoint-1);
the original memory frame, line and point number indexes of any point P (xp, yp, zp) in the three-dimensional space after reconstruction are as follows:
OP distance：
The point number of the P point is: nPoint_P= (dist_OP)/(h_p) +1 (5);
the point O 'is the projection point of the O point at the bottom of d, OO' is the body axis of the body data, OQ is the frame axis of the frame where the Q point and the P point are located,
the angle between the frame axis OQ and the body axis oo″: Δβ=atan ((yp-y 0)/(zp-z 0));
the frame number of the P point is: nframe_p= (Δβ)/(frame_β) +1 (6);
the angle between the frame axis OQ and the scan line OP: Δα=atan ((xp-x 0)/(zp-z 0)));
the line number of the P point is: nLine_P= Δα/(line_α) +1 (7);
the positions of memory data points corresponding to any point in the Cartesian three-dimensional space are obtained through formulas (5), (6) and (7), the indexes of the frame numbers, the line numbers and the point numbers are included, and the image forms corresponding to the scanning tracks are reconstructed in the Cartesian three-dimensional space according to the indexes.
5. The ultrasonic three-dimensional scanning image reconstruction method according to any one of claims 1 to 4, characterized in that: in step S3, accelerating the process by using the GPU when traversing all points of all Cartesian spaces;
and adopting CUDA programming to allocate W.L.H threads, and ensuring that each thread processes the index of one point.
6. The ultrasound three-dimensional scan image reconstruction method according to claim 5, wherein: and taking the original data as a three-dimensional texture, preparing a lookup table by using a coordinate index, storing the lookup table in the GPU in advance as a two-dimensional texture, accelerating the indexing process, and only performing texture pickup and tri-linear interpolation calculation in a thread.
7. An ultrasound three-dimensional scan image reconstruction system, comprising:
the system comprises a polar coordinate space volume data input module, a display module and a display module, wherein the polar coordinate space volume data input module is used for inputting volume data in a polar coordinate space, the volume data is composed of one or more data frames, and the data frames are composed of data of scanning planes which form a certain angle with the axle center, the axis or the axial plane of a three-dimensional scanning track in a Cartesian space;
the three-dimensional Cartesian coordinate and polar coordinate mapping module is used for establishing a mapping relation between the three-dimensional Cartesian coordinate and the polar coordinate; the interpolation module is used for obtaining the image value of the pixel corresponding to the Cartesian coordinate space through the tri-linear interpolation on the data points in the polar coordinate space;
wherein the data frame is composed of a plurality of data scanning lines, and the scanning lines are composed of a plurality of data points;
the ultrasonic three-dimensional scanning adopts a conical scanning mode with pivoting or a bidirectional sector conical scanning mode with non-pivoting, and each scanning conical surface forms a frame of data frame;
the data point index of the volume data of the polar coordinate space body comprises a frame number, a line number and a point number;
when the mapping relation between the three-dimensional Cartesian coordinates and the polar coordinates is established, frame numbers, line numbers and point number indexes of data points of volume data in the polar coordinate space corresponding to any pixel point of the Cartesian space are deduced, and all points of all the Cartesian space are traversed to form a group of space coordinate mapping relation.
8. The ultrasound three-dimensional scanning image reconstruction system according to claim 7, wherein: the data point index of the volume data of the polar coordinate space volume comprises a frame number, a line number and a point number;
when the mapping relation between the three-dimensional Cartesian coordinates and the polar coordinates is established, frame numbers, line numbers and point number indexes of data points of volume data in the polar coordinate space corresponding to any pixel point in the Cartesian space are deduced, and all points in all the Cartesian space are traversed to form a group of space coordinate mapping relation.
9. The ultrasound three-dimensional scanning image reconstruction system according to claim 7, wherein: the interpolation module respectively rounds up and down the frame number, the line number and the point number index value of the polar coordinate space obtained by the three-dimensional Cartesian coordinate and polar coordinate mapping module to obtain eight groups of index values of any point, corresponds to eight data points in the polar coordinate space, and then obtains the image value of the pixel corresponding to the Cartesian coordinate space through tri-linear interpolation.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110244855.9A CN112991479B (en) | 2021-03-05 | 2021-03-05 | Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system |
| PCT/CN2022/079064 WO2022184138A1 (en) | 2021-03-05 | 2022-03-03 | Ultrasonic three-dimensional scanning image reconstruction method and ultrasonic three-dimensional scanning image reconstruction system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110244855.9A CN112991479B (en) | 2021-03-05 | 2021-03-05 | Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112991479A CN112991479A (en) | 2021-06-18 |
| CN112991479B true CN112991479B (en) | 2023-08-08 |
Family
ID=76353017
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110244855.9A Active CN112991479B (en) | 2021-03-05 | 2021-03-05 | Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112991479B (en) |
| WO (1) | WO2022184138A1 (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112991479B (en) * | 2021-03-05 | 2023-08-08 | 深圳英美达医疗技术有限公司 | Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system |
| WO2023240584A1 (en) | 2022-06-17 | 2023-12-21 | 之江实验室 | Cross-media knowledge semantic expression method and apparatus |
| CN114820954B (en) * | 2022-06-29 | 2022-11-15 | 武汉中仪物联技术股份有限公司 | Three-dimensional reconstruction method and device, electronic equipment and storage medium |
| CN117406194A (en) * | 2023-12-15 | 2024-01-16 | 中科星图维天信科技股份有限公司 | Methods, devices, equipment and storage media for generating contour surface data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6002738A (en) * | 1995-07-07 | 1999-12-14 | Silicon Graphics, Inc. | System and method of performing tomographic reconstruction and volume rendering using texture mapping |
| CN1891159A (en) * | 2005-07-01 | 2007-01-10 | 深圳迈瑞生物医疗电子股份有限公司 | Quick scanning conversion method for 3-D supersonic imaging |
| CN101292883A (en) * | 2007-04-23 | 2008-10-29 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic three-dimensional quick imaging method and apparatus |
| CN104183013A (en) * | 2014-06-26 | 2014-12-03 | 宋和平 | Three-dimensional reconstruction method and system of ultrasonic probe and eye two-dimensional ultrasound image |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5787889A (en) * | 1996-12-18 | 1998-08-04 | University Of Washington | Ultrasound imaging with real time 3D image reconstruction and visualization |
| CN112991479B (en) * | 2021-03-05 | 2023-08-08 | 深圳英美达医疗技术有限公司 | Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system |
-
2021
- 2021-03-05 CN CN202110244855.9A patent/CN112991479B/en active Active
-
2022
- 2022-03-03 WO PCT/CN2022/079064 patent/WO2022184138A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6002738A (en) * | 1995-07-07 | 1999-12-14 | Silicon Graphics, Inc. | System and method of performing tomographic reconstruction and volume rendering using texture mapping |
| CN1891159A (en) * | 2005-07-01 | 2007-01-10 | 深圳迈瑞生物医疗电子股份有限公司 | Quick scanning conversion method for 3-D supersonic imaging |
| CN101292883A (en) * | 2007-04-23 | 2008-10-29 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic three-dimensional quick imaging method and apparatus |
| CN104183013A (en) * | 2014-06-26 | 2014-12-03 | 宋和平 | Three-dimensional reconstruction method and system of ultrasonic probe and eye two-dimensional ultrasound image |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022184138A1 (en) | 2022-09-09 |
| CN112991479A (en) | 2021-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112991479B (en) | Ultrasonic three-dimensional scanning image reconstruction method and reconstruction system | |
| JP5175275B2 (en) | Apparatus for rendering for display forward view image data | |
| JP3537594B2 (en) | Ultrasonic diagnostic equipment | |
| Belohlavek et al. | Three-and four-dimensional cardiovascular ultrasound imaging: a new era for echocardiography | |
| US5396890A (en) | Three-dimensional scan converter for ultrasound imaging | |
| JP4424799B2 (en) | Three-dimensional imaging method and system | |
| JP4204095B2 (en) | 3D imaging system and method for subject volume | |
| JP6611880B2 (en) | Data display and processing algorithms for 3D imaging systems | |
| CN101919707B (en) | Ultrasonic diagnosis apparatus, medical image processing method and image processing method | |
| US20050267366A1 (en) | Ultrasonic diagnostic apparatus and image processing method | |
| US20090018448A1 (en) | Three-dimensional ultrasonic diagnostic apparatus | |
| WO2016140116A1 (en) | Diagnostic imaging apparatus and image construction method | |
| JP2003061956A (en) | Ultrasonic diagnostic apparatus, medical diagnosing apparatus and image processing method | |
| US20250248687A1 (en) | Mechanical four-dimensional intracardiac ultrasound imaging system | |
| US20210093303A1 (en) | Medical image diagnostic apparatus, ultrasonic diagnostic apparatus, medical imaging system, and imaging control method | |
| JP4119497B2 (en) | Ultrasound diagnostic imaging equipment | |
| US7529431B2 (en) | Method and device for reconstructing and representing multidimensional objects from one-dimensional or two-dimensional image data | |
| JPH11128224A (en) | Ultrasonic 3-dimensional image reorganizing device | |
| Chandrasekaran et al. | Three-dimensional volumetric ultrasound imaging of arterial pathology from two-dimensional intravascular ultrasound: an in vitro study | |
| Kupferwasser et al. | Three-dimensional imaging of cardiac mass lesions by transesophageal echocardiographic computed tomography | |
| WO2024114021A1 (en) | 4d intracardiac echocardiography imaging system, echocardiography imaging method and echocardiography imaging apparatus | |
| JP3579320B2 (en) | Ultrasound image processing device | |
| JPH10216127A (en) | Ultrasonic diagnostic apparatus and adapter device for image processing | |
| Sugata et al. | Three dimensional display of ocular region using an array transducer | |
| US20230252749A1 (en) | Image processing device, image processing system, image display method, and image processing program |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |