Three-dimensional ultrasonic image reconstruction method and device

Abstract

The invention relates to the technical field of image analysis and processing, and provides a three-dimensional ultrasonic image reconstruction method and device. After a two-dimensional ultrasonic image sequence is obtained, the two-dimensional ultrasonic image sequence is mapped to a space rectangular coordinate system; dividing each frame of two-dimensional ultrasonic image into a plurality of polygonal frames; then, a voxel space is established between two adjacent frames of two-dimensional ultrasonic images, and coordinates of voxel points and coordinates of voxel points in the voxel space are determined; and finally, determining voxel values of the voxel points in the voxel spaces, and generating reconstruction data of the three-dimensional ultrasonic image according to the coordinates and the voxel values of the voxel points in all the voxel spaces. Therefore, the characteristics of the two-dimensional ultrasonic image sequence acquired by the side-emitting type rotary ultrasonic probe can be fully exerted, the coordinates and the voxel values of a plurality of voxel points can be synchronously determined in a plurality of voxel spaces by adopting an interpolation mode, and the reconstruction efficiency of the three-dimensional ultrasonic image is greatly improved through parallel calculation in the plurality of voxel spaces.

Description

Technical Field

[0001] The present application relates to the technical field of image analysis and processing, and in particular to a three-dimensional ultrasonic image reconstruction method and device. Background Art

[0002] like Figure 1 As shown in the figure, the side-firing ultrasound probe can realize two-dimensional ultrasound scanning in a single plane through a row of ultrasound transmitters and receivers on the axial direction of the ultrasound probe. The side-firing rotary ultrasound probe is based on the traditional side-firing ultrasound probe, and a rotary motor is added to drive the ultrasound transmitter and receiver to perform rotational scanning, replacing the manual rotation of the ultrasound probe for scanning. Figure 2 As shown in FIG. 1 , after the side-firing rotating ultrasound probe scans, a two-dimensional ultrasound image sequence around the central axis of the ultrasound probe can be obtained. Then, as shown in FIG. Figure 3 As shown, the two-dimensional ultrasound image sequence can be converted into a three-dimensional ultrasound volume data image through a three-dimensional volume data reconstruction algorithm, thereby providing three-dimensional ultrasound information.

[0003] When converting a two-dimensional ultrasound image sequence into a three-dimensional ultrasound volume data image, according to different reconstruction principles, the existing three-dimensional volume data reconstruction methods can be divided into three categories: pixel-based three-dimensional reconstruction algorithm, voxel-based three-dimensional reconstruction algorithm and function-based three-dimensional reconstruction algorithm.

[0004] Among them, the three-dimensional reconstruction algorithm based on pixels or voxels needs to traverse each voxel in the three-dimensional imaging space, or traverse each empty voxel, and then calculate the value at each voxel position through a certain interpolation algorithm. The three-dimensional reconstruction method based on functions is to establish a certain functional relationship between the pixel points of the two-dimensional ultrasound image and its three-dimensional space position, map the two-dimensional image pixels to the standard grid in the three-dimensional space, and calculate the voxel value in the three-dimensional grid by interpolating the function. The above three types of reconstruction methods are all general global three-dimensional reconstruction algorithms, which fail to combine the two-dimensional image sequence characteristics of the side-firing rotating ultrasound probe, resulting in low three-dimensional reconstruction efficiency. Summary of the invention

[0005] A three-dimensional ultrasonic image reconstruction method and device are used to fully consider the characteristics of a two-dimensional ultrasonic image sequence collected by a side-firing rotating ultrasonic probe to improve the reconstruction efficiency of three-dimensional ultrasonic images.

[0006] The present application provides a three-dimensional ultrasound image reconstruction method, comprising:

[0007] Acquire a two-dimensional ultrasound image sequence, and map the two-dimensional ultrasound image sequence to a pre-established spatial rectangular coordinate system, wherein the two-dimensional ultrasound image sequence includes multiple frames of two-dimensional ultrasound images continuously acquired along the same rotation axis;

[0008] Based on preset rules, multiple polygonal frames are divided on each frame of the two-dimensional ultrasound image, and the vertices of the polygonal frames are the centers of the pixels;

[0009] Establishing multiple voxel spaces between two adjacent frames of two-dimensional ultrasound images, wherein the voxel spaces are surrounded by polygonal frames at corresponding positions on the two adjacent frames of two-dimensional ultrasound images;

[0010] In the voxel space, the coordinates of the voxel point are determined, and the voxel value of the voxel point in the voxel space is determined according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space;

[0011] The reconstruction data of the three-dimensional ultrasound image is generated according to the coordinates and voxel values ​​of all voxel points in the voxel space.

[0012] In a possible implementation, one coordinate axis of the pre-established spatial rectangular coordinate system is the rotation axis of the acquisition of the two-dimensional ultrasound image sequence, and the other two coordinate axes are perpendicular to the rotation axis of the acquisition of the two-dimensional ultrasound image sequence.

[0013] In a possible implementation manner, the preset rule is: the shape of the polygonal frame is a quadrilateral, and the four vertices of the quadrilateral frame are respectively the centers of four adjacent pixel points.

[0014] In a possible implementation, the voxel space is an arch structure, and the upper arch surface and the lower arch surface of the arch structure are both curved surfaces;

[0015] The distances from the four vertices of the upper arch surface to the rotation axis are equal and equal to the arc radius of the upper arch surface; the distances from the four vertices of the lower arch surface to the rotation axis are equal and equal to the arc radius of the lower arch surface.

[0016] In a possible implementation, the step of determining the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space specifically includes:

[0017] According to the coordinates of the voxel point, the coordinates of the mapping point corresponding to the voxel point are determined in the polygonal frame corresponding to the voxel space;

[0018] Determine the pixel value of the mapping point according to the vertex coordinates of the polygonal frame, the vertex pixel values ​​of the polygonal frame and the coordinates of the mapping point;

[0019] According to the coordinates of the two mapping points, the pixel values ​​of the mapping points and the coordinates of the voxel point, linear interpolation is performed based on pixels to obtain the voxel value of the voxel point.

[0020] In a possible implementation, the step of determining the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space is specifically as follows:

[0021] A reference plane is established according to the coordinates of the voxel points and the ultrasonic scanning plane corresponding to the coordinates of the voxel points, and a reference frame where the reference plane intersects with the boundary of the voxel space is obtained, and the voxel points and the rotation axis of the two-dimensional ultrasonic image acquisition are in the reference plane;

[0022] Determine the vertex pixel values ​​of the reference frame according to the vertex coordinates of the two polygonal frames, the vertex coordinates of the reference frame, and the vertex pixel values ​​of the two polygonal frames;

[0023] The voxel value of the voxel point is determined according to the vertex coordinates of the reference frame, the vertex pixel values ​​of the reference frame and the voxel point coordinates.

[0024] In a possible implementation, the step of generating the reconstruction data of the three-dimensional ultrasound image according to the coordinates and voxel values ​​of all voxel points in the voxel space further includes:

[0025] In the three-dimensional volume data after the interpolation calculation is completed for all voxel points in the voxel space, the voxel values ​​of other background areas are filled with zero values.

[0026] The present application also provides a three-dimensional ultrasonic image reconstruction device, characterized in that:

[0027] An acquisition module, for acquiring a two-dimensional ultrasound image sequence;

[0028] A mapping module, used for mapping a two-dimensional ultrasound image sequence into a pre-established spatial rectangular coordinate system, wherein the two-dimensional ultrasound image sequence includes a plurality of frames of two-dimensional ultrasound images continuously acquired with the same rotation axis;

[0029] A frame division module is used to divide each frame of two-dimensional ultrasound image into multiple polygonal frames based on preset rules, where the vertices of the polygonal frames are the centers of the pixels;

[0030] A voxel space establishment module is used to establish multiple voxel spaces between two adjacent frames of two-dimensional ultrasound images, where the voxel spaces are surrounded by polygonal frames at corresponding positions on the two adjacent frames of two-dimensional ultrasound images;

[0031] A voxel value generation module is used to determine the coordinates of the voxel point in the voxel space, and determine the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space;

[0032] The 3D reconstruction module is used to generate the reconstruction data of the 3D ultrasound image according to the coordinates and voxel values ​​of all voxel points in the voxel space.

[0033] The present application also provides a three-dimensional ultrasonic image reconstruction device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0034] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the operation of the three-dimensional ultrasound image reconstruction method provided in the present application.

[0035] The present application also provides a computer-readable storage medium, in which at least one executable instruction is stored. When the executable instruction is executed on a three-dimensional ultrasonic image reconstruction device / apparatus, the three-dimensional ultrasonic image reconstruction device / apparatus performs the operation of the three-dimensional ultrasonic image reconstruction method provided in the present application.

[0036] The three-dimensional ultrasonic image reconstruction method and device provided by the embodiment of the present application, after acquiring the two-dimensional ultrasonic image sequence, first, maps the two-dimensional ultrasonic image sequence to a pre-established spatial rectangular coordinate system; then, based on the preset rules, divides a plurality of polygonal frames on each frame of the two-dimensional ultrasonic image; then, establishes a plurality of voxel spaces between two adjacent frames of the two-dimensional ultrasonic image, and determines the coordinates of the voxel points in the voxel space according to the vertex coordinates of the two polygonal frames corresponding to the voxel space; determines the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of the two polygonal frames corresponding to the voxel space; finally, generates the reconstruction data of the three-dimensional ultrasonic image according to the coordinates and voxel values ​​of the voxel points in all the voxel spaces. In this way, the characteristics of the two-dimensional ultrasonic image sequence acquired by the side-firing rotating ultrasonic probe can be fully utilized, and the coordinates and voxel values ​​of the plurality of voxel points can be determined by using the interpolation method in the plurality of voxel spaces simultaneously, and the reconstruction efficiency of the three-dimensional ultrasonic image can be greatly improved through parallel calculation in the plurality of voxel spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0038] Figure 1 A schematic diagram of the structure of a side-firing ultrasound probe provided in an embodiment of the present application;

[0039] Figure 2 A schematic diagram of the structure of a side-firing rotating ultrasonic probe provided in an embodiment of the present application;

[0040] Figure 3 A schematic diagram of converting a two-dimensional ultrasound image sequence provided in an embodiment of the present application into a three-dimensional ultrasound volume data image;

[0041] Figure 4 Schematic diagram of the three-dimensional ultrasound image reconstruction process provided in the embodiment of the present application Figure 1 ;

[0042] Figure 5 A schematic diagram of the position of a spatial rectangular coordinate system provided in an embodiment of the present application;

[0043] Figure 6 A schematic diagram of a rectangular frame formed by four adjacent pixels provided in an embodiment of the present application;

[0044] Figure 7 A schematic diagram of a rectangular frame formed by four non-adjacent pixels provided in an embodiment of the present application;

[0045] Figure 8 A schematic diagram of mapping a rectangular frame to a spatial rectangular coordinate system provided in an embodiment of the present application;

[0046] Fig. 9 A schematic diagram of an arch structure between two adjacent frames of two-dimensional ultrasound images provided in an embodiment of the present application;

[0047] Fig.10 Schematic diagram of the three-dimensional ultrasound image reconstruction process provided in the embodiment of the present application Figure 2 ;

[0048] Fig.11 Schematic diagram of arch interpolation provided in the embodiment of the present application Figure 1 ;

[0049] Fig.12 A schematic diagram of determining mapping point pixels by bilinear interpolation in a plane provided in an embodiment of the present application;

[0050] Fig.13 A schematic diagram of determining the voxel value of a voxel point by distance weighting provided in an embodiment of the present application;

[0051] Fig.14 Schematic diagram of the three-dimensional ultrasound image reconstruction process provided in the embodiment of the present application Figure 3 ;

[0052] Fig.15 Schematic diagram of arch interpolation provided in the embodiment of the present application Figure 2 ;

[0053] Fig.16 Schematic diagram of determining vertex pixel values ​​of a reference frame by distance weighting provided in an embodiment of the present application

[0054] Fig.17 A schematic diagram of determining voxel point pixels by bilinear interpolation in a plane provided in an embodiment of the present application. DETAILED DESCRIPTION

[0055] When converting a two-dimensional ultrasound image sequence into a three-dimensional ultrasound volume data image, the global three-dimensional reconstruction algorithm used in the prior art fails to combine the two-dimensional image sequence characteristics of the side-firing rotating ultrasound probe, resulting in low three-dimensional reconstruction efficiency.

[0056] In view of the above technical problems, the present application provides a three-dimensional ultrasound image reconstruction method. Figure 4 As shown, the three-dimensional ultrasound image reconstruction method includes S410 to S450.

[0057] S410, acquiring a two-dimensional ultrasonic image sequence, and mapping the two-dimensional ultrasonic image sequence to a pre-established spatial rectangular coordinate system, wherein the two-dimensional ultrasonic image sequence includes multiple frames of two-dimensional ultrasonic images continuously acquired with the same rotation axis.

[0058] Specifically, one coordinate axis of the spatial rectangular coordinate system can be set to coincide with the line where the rotation axis is located, and the other two coordinate axes are perpendicular to the line where the rotation axis is located, wherein the rotation axis refers to the motor rotation axis of the side-firing rotating ultrasound probe that collects the two-dimensional ultrasound image sequence. Figure 5 As shown, the z-axis of the spatial rectangular coordinate system coincides with the motor rotation axis, the y-axis is vertically arranged perpendicular to the rotation axis, wherein the y-axis coincides with the plane where one of the two-dimensional ultrasound images is located, and the x-axis is horizontally arranged perpendicular to the rotation axis.

[0059] In this way, the edge of the scanning plane of the two-dimensional ultrasonic image coincides with the z-axis, and the motor of the side-firing rotary ultrasonic probe rotates and collects the two-dimensional ultrasonic image frame by frame. According to the image sequence, each frame of the two-dimensional ultrasonic image has a unique position in the spatial coordinate system, and the position is determined by the rotation angle of the motor's rotating shaft, thereby mapping the scanned two-dimensional ultrasonic image sequence to a pre-established spatial rectangular coordinate system.

[0060] S420, based on a preset rule, divide each frame of the two-dimensional ultrasound image into a plurality of polygonal frames, where the vertices of the polygonal frames are the centers of the pixels.

[0061] Among them, Figure 6 As shown, the preset rule is: the shape of the polygonal frame is a quadrilateral, and the four vertices of the quadrilateral are the centers of four adjacent pixels, that is, Figure 6 One side of the quadrilateral shown in FIG. 1 passes through the area where two pixels are located. It should be noted that the pixel points are actually square pixel units. In practical applications, the pixel coordinate index in the two-dimensional image can be defined as (i, j). In this way, after mapping the pixel point to the spatial rectangular coordinate system, it can be represented by (z i ,y j ) represents the coordinate values ​​of the pixel point on the z-axis and y-axis in the spatial rectangular coordinate system.

[0062] Of course, in actual application, the preset rules can be adjusted according to actual needs. For example, the shape of the polygonal frame is a polygonal structure other than a quadrilateral. For another example, Figure 7 As shown, the four vertices of the quadrilateral are the centers of four pixels, and there are pixels between two pixels on the same side of the quadrilateral, that is, Figure 7 One side of the quadrilateral shown passes through the area where three pixels are located.

[0063] Taking the polygonal frame as a quadrilateral as an example, for each frame of two-dimensional ultrasound image. Figure 6 As shown, four adjacent pixel center points form a rectangular frame. The four vertices of the rectangular frame are the centers of the four pixel points, and the pixel values ​​of the pixel points are used as the pixel values ​​of the vertices. Figure 8 As shown, it is a schematic diagram showing the rectangular frame in each frame of two-dimensional ultrasound image in the spatial rectangular coordinate system.

[0064] S430, establishing a plurality of voxel spaces between two adjacent frames of two-dimensional ultrasonic images, wherein the voxel spaces are surrounded by polygonal frames at corresponding positions on the two adjacent frames of two-dimensional ultrasonic images.

[0065] The polygonal frames at corresponding positions on two adjacent two-dimensional ultrasound images refer to the two polygonal frames having the same position on the two-dimensional ultrasound images. Taking the shape of the polygonal frame as a rectangular frame as an example, for each pair of two adjacent two-dimensional ultrasound images, two rectangular frames at the same position on the two-dimensional ultrasound images are used to form a voxel space. The two rectangular frames having the same position on the two-dimensional ultrasound image refer to the polygons having the same coordinate index on the two-dimensional ultrasound image, that is, the four vertex pixels of the rectangular frame have the same coordinate index on the two-dimensional ultrasound image.

[0066] like Fig. 9 As shown, the two polygonal frames are the surfaces of the same voxel space on the two-dimensional ultrasound image, and the coordinate values ​​of the two polygonal frames on the rotation axis (z axis) are the same. In this way, the coordinate indexes of the two polygonal frames on their respective two-dimensional ultrasound images are the same, that is, after rotating one polygonal frame along the z axis (rotation axis), the two polygonal frames can be overlapped.

[0067] In this example, two adjacent frames of two-dimensional ultrasound images are taken by rotating the ultrasound probe, and the voxel space forms an arch, and the upper and lower arch surfaces of the arch structure are both curved surfaces. Fig. 9As shown, the characteristics of this arch body are: the distances from the four vertices of the lower arch surface of the arch body to the rotation axis (z axis) of the ultrasound probe are equal and equal to the arc radius of the lower arch surface. In other words, the arc radius formed by the arc surface of the lower arch surface is the distance from the lower arch surface to the rotation axis. The distances from the four vertices of the upper arch surface of the arch body to the rotation axis (z axis) of the ultrasound probe are equal and equal to the arc radius of the upper arch surface. In other words, the arc radius formed by the arc surface of the upper arch surface is the distance from the upper arch surface to the rotation axis. In this way, all two-dimensional ultrasound image sequences can be divided into several independent voxel spaces.

[0068] S440, determining the coordinates of the voxel point in the voxel space, and determining the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space.

[0069] Taking the shape of the polygonal frame as a quadrilateral as an example, in the embodiment of the present application, the scheme is described by taking any voxel point in the arch body. In practical applications, the number or position of voxel points in each arch body can be set differently, wherein the position of the voxel point is relative to the arch body. Specifically, the number and position of the voxel points are determined according to the total voxel space size reconstructed and the reconstructed voxel density, and after determination, the coordinates of all voxel points in the arch body corresponding to the spatial coordinate system can be obtained.

[0070] In the process of determining the voxel value of a voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space, the embodiment of the present application provides two implementation methods.

[0071] like Fig.10 As shown, in some embodiments, a first implementation manner of determining a voxel value includes steps S1010 to S1030.

[0072] S1010, according to the coordinates of the voxel point, determine the coordinates of a mapping point corresponding to the voxel point in a polygonal frame corresponding to the voxel space.

[0073] Take the polygonal frame as a rectangular frame as an example, Fig.11 As shown, first, according to the coordinates of the voxel point, the coordinates of the two mapping points corresponding to the voxel point are determined in the two polygonal frames corresponding to the voxel space. Among them, the two mapping points corresponding to the voxel point refer to the coordinate points of the voxel point in the polygonal frame when the voxel point is rotated along the collected rotation axis so that the voxel point is rotated into the two-dimensional ultrasound image.

[0074] In some embodiments, the coordinates of the mapping point can be established in the following manner: through the voxel point, with the distance from the voxel point to the rotation axis as the radius, a reference circle is established, and the reference circle is parallel to the front and back faces of the arch. Among them, the front and back faces are parallel to the XOY plane of the spatial rectangular coordinate system, and the intersection of the reference circle and the plane area in the two rectangular frames is the mapping point. The arch is a hexahedron, including an upper arch surface, a lower arch surface, and two parallel fan-shaped surfaces. The two parallel fan-shaped surfaces are the front and back faces of the arch described above, and the remaining two faces are the faces formed by the polygonal frames on two adjacent frames of two-dimensional ultrasound images, that is, the faces where the mapping points are located. In this example, the front and back faces are parallel to the XOY plane of the spatial rectangular coordinate system, and O is the origin of the spatial rectangular coordinate system.

[0075] In some embodiments, the coordinates of the mapping point can also be established in the following manner: determine the position of the mapping point by coordinate calculation. For example, if the z-axis coincides with the rotation axis, and the x-axis coincides with the plane where one of the two-dimensional ultrasound images is located, there are three constraints. The first constraint is that the coordinate value of the mapping point on the z-axis is equal to the coordinate value of the voxel point on the z-axis. The second constraint is that the distance from the mapping point to the z-axis is equal to the distance from the voxel point to the z-axis. The third constraint is that the mapping point is in the plane area within the rectangular frame, and the plane area within the rectangular frame can be expressed by a function. Then, through the above three constraints, the position of the mapping point can be uniquely determined in the plane area within the rectangular frame.

[0076] S1020, determining the pixel value of the mapping point according to the vertex coordinates of the polygonal frame, the vertex pixel values ​​of the polygonal frame, and the coordinates of the mapping point.

[0077] Specifically, taking the polygonal frame as a rectangular frame as an example, the pixel value at the mapping point is calculated using the four vertex pixel values ​​of the rectangular frame through the bilinear interpolation method in the plane. Fig.12As shown, the four vertices of the rectangular box are m1, m2, m3 and m4, and the pixel values ​​of the four vertices are f(m1), f(m2), f(m3) and f(m4) respectively. pA is the mapping point of the voxel point to be determined on the plane where the rectangular box is located. Then, e1 and e2 can be determined based on the vertex coordinates of the polygonal box and the coordinates of the mapping point. Among them, e1 is the ratio of the distance from pA to edge m1m2 to the length of edge m1m3, and e2 is the ratio of the distance from pA to edge m1m3 to the length of edge m1m2. The pixel value f(pA) at pA can be obtained by the following model: f(pA) = (1-e2) × ((1-e1) × f(m1) + e1 × f(m3)) + e2 × ((1-e1) × f(m2) + e1 × f(m4)); or, the pixel value f(pA) at pA can be obtained by the following model: f(pA) = (1-e1) × ((1-e2) × f(m1) + e2 × f(m2)) + e1 × ((1-e2) × f(m3) + e2 × f(m4)).

[0078] S1030, obtaining a voxel value of the voxel point according to the coordinates of the two mapping points, the pixel values ​​of the mapping points, and the coordinates of the voxel point.

[0079] For example, Fig.13 As shown, pX is the voxel point to be determined, pA and pB are the mapping points of pX in two rectangular boxes, respectively. Then, d1 and d2 can be determined by the coordinates of the two mapping points and the coordinates of the voxel point. Among them, d1 is the distance from pX to pA, d2 is the distance from pX to pB, f(pA) and f(pB) are the pixel values ​​at pA and pB, respectively. Then, the distance weighting can be used to determine the pixel value f(pX) at pX. Specifically, it can be obtained by the following model: f(pX) = (f(pA) / d1+f(pB) / d2) / (1 / d1+1 / d2).

[0080] like Fig.14 As shown, in some embodiments, the second implementation manner of determining the voxel value includes: S1410 to S1430.

[0081] S1410, establishing a reference plane according to the coordinates of the voxel points and the ultrasonic scanning plane corresponding to the coordinates of the voxel points, and obtaining a reference frame where the reference plane intersects with the boundary of the voxel space.

[0082] like Fig.15As shown, firstly, a plane passing through the ultrasound rotation axis where the desired voxel point is located is established as a reference plane. In addition, the mathematical expression of the reference plane can also be directly determined by mathematical calculation methods. For example, if the coordinates of the voxel point in the reference plane are known, and the reference plane is perpendicular to the XOY plane of the spatial rectangular coordinate system, the mathematical expression of the reference plane can be determined according to the mathematical expression of the XOY plane. Then, when the reference plane is determined, the reference frame formed by the intersection of the reference plane and the arch is calculated, and the coordinates of the four vertices of the reference frame are obtained. In this example, the shape of this rectangular frame is a rectangle.

[0083] S1420 , determining vertex pixel values ​​of the reference frame according to vertex coordinates of the two polygonal frames, vertex coordinates of the reference frame, and vertex pixel values ​​of the two polygonal frames.

[0084] The four vertex pixel values ​​of the reference frame are obtained by using the distance weighting method through the four vertex pixel values ​​of the two rectangular frames, wherein the distance weighting method is the same as the principle of calculating the pixel value at pX mentioned above. Fig.16 As shown, taking two vertices pX′ and pX″′ of the reference frame as an example, m2 and m4 are two vertices of a rectangular frame, m2′ and m4′ are two vertices of another rectangular frame, f(pX′) is the pixel value of vertex pX′, f(pX″′) is the pixel value of vertex pX″′, f(m2) is the pixel value of vertex m2, f(m4) is the pixel value of vertex m4, f(m2′) is the pixel value of vertex m2′, and f(m4′) is the pixel value of vertex m4′. Then we can adopt Use distance weighting to determine the pixel value f(pX′) and pixel value f(pX”') at pX′ and pX”'. Specifically, f(pX') can be obtained through the following model: f(pX′) = (f(m2′) / d1'+f(m2) / d2') / (1 / d1'+1 / d2'). And, f(pX”') can be obtained through the following model: f(pX”') = (f(m4') / d1"+f(m4) / d2") / (1 / d1"+1 / d2").

[0085] S1430 , determining a voxel value of the voxel point according to the vertex coordinates of the reference frame, the vertex pixel values ​​of the reference frame, and the voxel point coordinates.

[0086] By referring to the pixel values ​​of the four vertices of the frame, the bilinear interpolation method in the plane can be used to obtain the voxel value at the desired voxel point, where the bilinear interpolation method in the plane is the same as the principle of calculating the pixel value of the mapping point mentioned above. Fig.17As shown in the figure, the four vertices of the rectangular frame are pX, pX', pX" and pX"', the pixel values ​​of the four vertices are f(pX), f(pX'), f(pX") and f(pX"'), pA is the mapping point of the voxel point to be determined on the plane where the reference frame is located, e1 is the ratio of the distance from pA to the edge pXpX' to the length of the edge pXpX", and e2 is the ratio of the distance from pA to pXpX" to the length of the edge pXpX'. The pixel value f(pA) at pA can be obtained by the following model: Alternatively, the pixel value f(pA) at pA can be obtained by the following model: f(pA) = (1-e2)×((1-e1)×f(pX)+e1×f(pX"))+e2×((1-e1)×f(pX′)+e1×f(pX”')); or, the pixel value f(pA) at pA can be obtained by the following model: f(pA) = (1-e1)×((1-e2)×f(pX)+e2×f(pX′))+e1×((1-e2)×f(pX")+e2×f(pX”')).

[0087] It should be noted that when the voxel point coordinates and the voxel value of the voxel point are interpolated and calculated in each arch, the interpolation calculation of each arch does not affect each other and can be calculated in parallel. The GPU (Graphics Processing Unit) can be used in parallel to accelerate the interpolation calculation process, thereby improving the speed of determining the voxel value of the voxel point position.

[0088] S450: Generate reconstruction data of a three-dimensional ultrasonic image according to the coordinates and voxel values ​​of all voxel points in the voxel space.

[0089] In the three-dimensional volume data after the interpolation calculation is completed for all voxel points in the voxel space, the voxel values ​​of other background areas are filled with zero values, so as to obtain complete three-dimensional ultrasonic image reconstruction data.

[0090] The three-dimensional ultrasonic image reconstruction method provided by the embodiment of the present application is applied. After acquiring the two-dimensional ultrasonic image sequence, first, the two-dimensional ultrasonic image sequence is mapped to a pre-established spatial rectangular coordinate system; then, based on the preset rules, multiple polygonal frames are divided on each frame of the two-dimensional ultrasonic image; then, multiple voxel spaces are established between two adjacent frames of two-dimensional ultrasonic images, and the coordinates of the voxel points in the voxel space are determined according to the coordinates of the voxel points and the vertex coordinates of the two polygonal frames corresponding to the voxel space; the voxel values ​​of the voxel points in the voxel space are determined according to the vertex pixel values ​​of the two polygonal frames corresponding to the voxel space; finally, the reconstruction data of the three-dimensional ultrasonic image is generated according to the coordinates and voxel values ​​of the voxel points in all voxel spaces. In this way, the characteristics of the two-dimensional ultrasonic image sequence collected by the side-firing rotating ultrasonic probe can be fully utilized, and the interpolation method can be used synchronously in multiple voxel spaces to determine the coordinates and voxel values ​​of multiple voxel points. Through parallel calculation in multiple voxel spaces, the reconstruction efficiency of the three-dimensional ultrasonic image is greatly improved.

[0091] The embodiment of the present application also provides a three-dimensional ultrasound image reconstruction device, including an acquisition module, a mapping module, a frame division module, a voxel space establishment module, a voxel value generation module and a three-dimensional reconstruction module.

[0092] The acquisition module acquires a two-dimensional ultrasound image sequence.

[0093] The mapping module is used to map a two-dimensional ultrasonic image sequence into a pre-established spatial rectangular coordinate system, wherein the two-dimensional ultrasonic image sequence includes multiple frames of two-dimensional ultrasonic images continuously acquired with the same rotation axis.

[0094] The frame division module is used to divide each frame of two-dimensional ultrasound image into multiple polygonal frames based on preset rules, and the vertices of the polygonal frames are the centers of the pixels.

[0095] The voxel space establishment module is used to establish multiple voxel spaces between two adjacent frames of two-dimensional ultrasonic images. The voxel space is surrounded by polygonal frames at corresponding positions on the two adjacent frames of two-dimensional ultrasonic images.

[0096] The voxel value generation module is used to determine the coordinates of the voxel point in the voxel space, and determine the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space.

[0097] The three-dimensional reconstruction module is used to generate reconstruction data of the three-dimensional ultrasound image according to the coordinates and voxel values ​​of all voxel points in the voxel space.

[0098] An embodiment of the present application also provides a three-dimensional ultrasonic image reconstruction device, including: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus; the memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the operation of the above-mentioned three-dimensional ultrasonic image reconstruction method.

[0099] An embodiment of the present application further provides a computer-readable storage medium, in which at least one executable instruction is stored. When the executable instruction is executed on a three-dimensional ultrasonic image reconstruction device / apparatus, the three-dimensional ultrasonic image reconstruction device / apparatus performs the operation of the above-mentioned three-dimensional ultrasonic image reconstruction method.

[0100] The above specific implementation methods further illustrate the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above are only specific implementation methods of the present application and are not used to limit the protection scope of the present application. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solutions of the present application should be included in the protection scope of the present application.

Claims

1. A three-dimensional ultrasonic image reconstruction method, characterized in that: include: Acquire a two-dimensional ultrasound image sequence, and map the two-dimensional ultrasound image sequence to a pre-established spatial rectangular coordinate system, wherein the two-dimensional ultrasound image sequence includes multiple frames of two-dimensional ultrasound images continuously acquired along the same rotation axis; Based on a preset rule, a plurality of polygonal frames are divided on each frame of the two-dimensional ultrasound image, and the vertices of the polygonal frames are the centers of the pixels; Establishing a plurality of voxel spaces between two adjacent frames of two-dimensional ultrasonic images, wherein the voxel spaces are surrounded by polygonal frames at corresponding positions on the two adjacent frames of two-dimensional ultrasonic images; In the voxel space, the coordinates of the voxel point are determined, and the voxel value of the voxel point in the voxel space is determined according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space; The reconstruction data of the three-dimensional ultrasound image is generated according to the coordinates and voxel values ​​of all voxel points in the voxel space.

2. The three-dimensional ultrasonic image reconstruction method according to claim 1, characterized in that: One coordinate axis of the pre-established spatial rectangular coordinate system is the rotation axis of the acquisition of the two-dimensional ultrasonic image sequence, and the other two coordinate axes are perpendicular to the rotation axis of the acquisition of the two-dimensional ultrasonic image sequence.

3. The three-dimensional ultrasonic image reconstruction method according to claim 1, characterized in that: The preset rule is: the shape of the polygonal frame is a quadrilateral, and the four vertices of the quadrilateral frame are respectively the centers of four adjacent pixel points.

4. The three-dimensional ultrasonic image reconstruction method according to claim 3, characterized in that: The voxel space is an arch structure, and the upper arch surface and the lower arch surface of the arch structure are both curved surfaces; The distances from the four vertices of the upper arch surface to the rotation axis are equal and equal to the arc radius of the upper arch surface; the distances from the four vertices of the lower arch surface to the rotation axis are equal and equal to the arc radius of the lower arch surface.

5. The three-dimensional ultrasonic image reconstruction method according to claim 1, characterized in that: The step of determining the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space specifically includes: According to the coordinates of the voxel point, within the polygonal frame corresponding to the voxel space, the coordinates of a mapping point corresponding to the voxel point are determined; Determine the pixel value of the mapping point according to the vertex coordinates of the polygonal frame, the vertex pixel values ​​of the polygonal frame and the coordinates of the mapping point; According to the coordinates of the two mapping points, the pixel values ​​of the mapping points and the coordinates of the voxel point, pixel linear interpolation is performed to obtain the voxel value of the voxel point.

6. The three-dimensional ultrasonic image reconstruction method according to claim 1, characterized in that: The step of determining the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space is specifically: A reference plane is established according to the coordinates of the voxel point and the ultrasonic scanning plane corresponding to the coordinates of the voxel point, and a reference frame where the reference plane intersects with the boundary of the voxel space is obtained, wherein the voxel point and the rotation axis of the two-dimensional ultrasonic image acquisition are within the reference plane; Determine the vertex pixel values ​​of the reference frame according to the vertex coordinates of the two polygonal frames, the vertex coordinates of the reference frame, and the vertex pixel values ​​of the two polygonal frames; The pixel of the voxel point is determined according to the vertex coordinates of the reference frame, the vertex pixel values ​​of the reference frame and the voxel point coordinates.

7. The three-dimensional ultrasonic image reconstruction method according to claim 1, characterized in that: The step of generating the reconstruction data of the three-dimensional ultrasonic image according to the coordinates and voxel values ​​of all voxel points in the voxel space also includes: In the three-dimensional volume data after the interpolation calculation is completed for all voxel points in the voxel space, the voxel values ​​of other background areas are filled with zero values.

8. A three-dimensional ultrasonic image reconstruction device, characterized in that: include: An acquisition module, for acquiring a two-dimensional ultrasound image sequence; A mapping module, used for mapping a two-dimensional ultrasound image sequence into a pre-established spatial rectangular coordinate system, wherein the two-dimensional ultrasound image sequence includes multiple frames of two-dimensional ultrasound images continuously acquired with the same rotation axis; A frame division module is used to divide each frame of two-dimensional ultrasound image into multiple polygonal frames based on preset rules, where the vertices of the polygonal frames are the centers of the pixels; A voxel space establishment module is used to establish a plurality of voxel spaces between two adjacent frames of two-dimensional ultrasonic images, wherein the voxel spaces are surrounded by polygonal frames at corresponding positions on the two adjacent frames of two-dimensional ultrasonic images; A voxel value generating module, used to determine the coordinates of a voxel point in the voxel space, and determine the voxel value of the voxel point in the voxel space according to the coordinates of the voxel point and the vertex pixel values ​​of two polygonal frames corresponding to the voxel space; The three-dimensional reconstruction module is used to generate reconstruction data of the three-dimensional ultrasound image according to the coordinates and voxel values ​​of all voxel points in the voxel space.

9. A three-dimensional ultrasonic image reconstruction device, characterized in that: include: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform the operation of the three-dimensional ultrasound image reconstruction method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one executable instruction, and when the executable instruction is executed on the three-dimensional ultrasonic image reconstruction device / apparatus, the three-dimensional ultrasonic image reconstruction device / apparatus performs the operation of the three-dimensional ultrasonic image reconstruction method according to any one of claims 1 to 7.

Images