CN104739442A - Pressure elastic imaging displacement detection method, pressure elastic imaging displacement detection device and ultrasonic imaging device - Google Patents

Abstract

The application discloses an elastic imaging displacement detection method. The elastic imaging displacement detection method includes: when performing displacement detection on nodes in target image data which is arranged in multiple rows, using the node in the last row of each row and highly matched with the node in each row as a guide node of the node in each row, confirming initial offset of each node according to offset of the corresponding guild node, searching a matching point based on a certain range of each guild node, and thereby reducing cumulative effects of miscalculations. The application furthermore provides a pressure elastic imaging displacement detection device applicable to the pressure elastic imaging displacement detection method, and discloses an ultrasonic imaging device.

Description

Compressive resilience imaging displacement detection method, device and supersonic imaging apparatus

Technical field

The application relates to a kind of armarium, is specifically related to a kind of compressive resilience imaging displacement detection method and device thereof and a kind of supersonic imaging apparatus.

Background technology

Medical supersonic elastogram mainly refers to a series of imaging for the purpose of display organization elastic difference and signal processing technology.Current existing a few macrotaxonomy comprises compressive resilience imaging, acoustic radiation force elastogram (AcousticRadiation Force Imaging, ARFI), shearing wave elastogram (Shear Wave Elastography, SWE) etc.Wherein the time of compressive resilience imaging progress is the longest, and technology is also ripe.Compressive resilience imaging, as cancer detection, especially in the optimum pernicious differentiation of breast carcinoma, to the important supplementary means that B-mode ultrasonogram detects, is applied to clinical fast.

Compressive resilience imaging mainly applies pressure (pressure) by hand-held ultrasound probe to destination organization, obtain destination organization and compressed front and back two frame ultrasonic echo information, the displacement (displacement) that before and after compression, correspondence position occurs is calculated again by specific algorithm, be destination organization two not spatial position change information in the same time, by asking axially (axial) gradient to displacement, and then obtain the strain value (strain) of target tissue region each point, under identical outer force compresses, strain larger, expression is organized softer, strain less, then expression is organized harder.Strain value according to target tissue region each point shows with pictorial form, intuitively can reflect the soft or hard difference between different tissues or elasticity difference.

In above-mentioned processing procedure, whether displacement detecting is accurate, whether calculating is quick, and joint effect the Contrast-to-noise ratio (contrast-noise-ratio, CNR), the real-time of image, clinical frame per second etc. of final strain pattern.Application number be 201110159110.9 Chinese patent " displacement detecting method in a kind of elastogram, Apparatus and system " propose guide zero phase estimate (Guided Phase Zero Estimation, GPZE) algorithm, this GPZE displacement detecting algorithm, use optimum default displacement result to guide next line displacement to calculate on the one hand, decrease volumes of searches; The method of phase estimation is used to carry out displacement calculating on the other hand, less demanding to initial data sample rate, greatly reduce amount of calculation.But GPZE algorithm still has following weak point:

Existing GPZE algorithm is when calculating one frame elastic image, adopt and guide account form line by line, because guide position is too fixing, when lastrow point occurs that mistake in computation causes bad point, due to the existence guided, this position of the every a line after extending to that can lead to errors, image shows as longitudinal wire error.

Summary of the invention

The application provides a kind of compressive resilience imaging displacement detection method and device thereof and a kind of supersonic imaging apparatus, reduces the probability of the downward conduction eror of pilot point.

According to the first aspect of the application, the application provides a kind of compressive resilience imaging displacement detection method, comprising:

Obtain two frame image datas, respectively as compression before destination image data and compression after by matching image data;

Using the node of the first row in destination image data as impact point, in by matching image data, find corresponding match point respectively;

According to the impact point of the first row and the displacement result of each node of match point calculating the first row thereof;

Calculate the displacement result of each node of N-th row in destination image data, wherein N be successively from 2 to n integer, n is the line number that a two field picture divides, and comprising:

The highest node of matching degree is found out as row initial guide point in the matching result of the capable each node of N-1;

The each node searching N-th row based on row initial guide point is by the match point in matching image data;

According to each node of N-th row and the displacement result of each node of match point calculating N-th row thereof.

According to the second aspect of the application, the application provides displacement detector in a kind of compressive resilience imaging, comprising:

Image collection module, for obtaining two frame image datas, respectively as compression before destination image data and compression after by matching image data;

First matching module, for using the node of the first row in destination image data as impact point, find corresponding match point in by matching image data respectively;

Search module, for finding out the highest node of matching degree as row initial guide point in the matching result of the capable each node of N-1, wherein N is the current line needing to search match point, N be successively from 2 to n integer, n is the line number that a two field picture divides;

Second matching module, for each node of searching N-th row based on row initial guide point by the match point in matching image data;

Displacement computing module, for calculating the displacement result of each node of the first row according to the impact point of the first row and match point thereof, and calculates the displacement result of each node of N-th row according to each node of N-th row and match point thereof.

According to the third aspect of the application, the application provides a kind of supersonic imaging apparatus, comprising:

Probe, for receiving ultrasonic echo to scanning objective emission ultrasound wave;

Signal processor, for processing ultrasonic echo, generates ultrasound image data;

Image processor, for processing ultrasound image data, and generate elastic image, image processor comprises: the displacement result of each node that above-mentioned displacement detector and deformation based checkout gear detect generates the elastic image generating apparatus of elastic image.

The beneficial effect of the application is: improve guidance mode, by existing fixing guidance mode, change into by the highest node of matching degree as pilot point, thus reduce the probability of the downward conduction eror of pilot point, improve the reliability of guiding, avoid the accumulative effect of mistake in computation.

Accompanying drawing explanation

Fig. 1 is the embodiment of the present application supersonic imaging apparatus structure chart;

Fig. 2 is the embodiment of the present application displacement detector structure chart;

Fig. 3 is the embodiment of the present application second matching module structure chart;

Fig. 4 is each modal displacement overhaul flow chart of the embodiment of the present application;

Fig. 5 is that the embodiment of the present application video data block divides schematic diagram;

Fig. 6 is the flow chart that the embodiment of the present application searches the match point of each node;

Fig. 7 is the one strategy schematic diagram that the embodiment of the present application determines region of search.

Detailed description of the invention

By reference to the accompanying drawings the present invention is described in further detail below by detailed description of the invention.

Medical elastic imaging mainly refers to a series of imaging for the purpose of display organization elastic difference and signal processing technology.For medical ultrasound imaging technology, please refer to Fig. 1, Figure 1 shows that the structure of supersonic imaging apparatus, comprise probe 1, signal processor 2, image processor 3 and display 4.Wherein:

Probe 1 is for receiving ultrasonic echo to scanning objective emission ultrasound wave.Ultrasonic output circuit 11 produces Wave data, connected the array element of probe 1 by transmission channel 12, to being detected tissue emissions ultrasound wave, ultrasound wave forms ultrasonic echo through Tissue reflectance with after absorbing, probe 1 receives ultrasonic echo, exports signal processor 2 to by receive path 13.

Signal processor 2, for processing ultrasonic echo, generates ultrasound image data.First the ultrasonic echo that receive path 13 receives is obtained radio frequency (radiofrequency, RF) signal by Beam synthesis link by signal processor 2; The baseband signal of quadrature demodulation is obtained again after quadrature demodulation.In processing procedure, can also after Beam synthesis, radio frequency signal carries out a liter sampling, increases the sample rate of RF signal, then after quadrature demodulation again through down-sampled.Displacement detecting precision can be increased by rising sampling, rising sample rate and being preset by system.Ultrasound image data after process outputs to image processor.

Image processor 3 for processing ultrasound image data, and generates elastic image.Image processor 3 comprises displacement detector and elastic image generating apparatus, displacement detector processes to the ultrasound image data that signal processor 2 exports the displacement result obtaining each node, and the displacement result of each node of then elastic image generating apparatus deformation based checkout gear detection generates elastic image.

The elastic image that display 4 generates for display image processor 3.

The improvement of the application is the displacement detector in image processor 3, is illustrated in figure 2 the structure of displacement detector, comprises image collection module 301, first matching module 302, searches module 303, second matching module 304 and displacement computing module 305.Image collection module 301 for obtaining two frame image datas, respectively as compression before destination image data and compression after by matching image data; First matching module 302 for using the node of the first row in destination image data as impact point, find corresponding match point in by matching image data respectively; Search module 303 for finding out the highest node of matching degree as row initial guide point in the matching result of the capable each node of N-1, wherein N is the current line needing to search match point, N be successively from 2 to n integer, n is the line number that a two field picture divides; Second matching module 304 for each node of searching N-th row based on row initial guide point by the match point in matching image data; Displacement computing module 305 for calculating the displacement result of each node of the first row according to the impact point of the first row and match point thereof, and calculates the displacement result of each node of N-th row according to each node of N-th row and match point thereof.In a kind of instantiation, displacement computing module 305 comprises cross-correlation phase calculation unit and displacement computing unit, cross-correlation phase calculation unit, for calculating the cross-correlation phase place of this node based on the ultrasonic radio frequency complex signal of this node and match point thereof; Displacement computing unit is used for the final mean annual increment movement result based on this node of cross-correlation phase calculation.

In one embodiment, second matching module 304 structure as shown in Figure 3, comprises first object point determining unit 341, first side-play amount acquiring unit 342, first region of search determining unit 343, first search unit 344, second impact point determining unit 345, pilot point determining unit 346, second side-play amount acquiring unit 347, second region of search determining unit 348 and the second search unit 349.First object point determining unit 341 is for the first object point of N-th row in the highest node determination destination image data of matching degree in capable based on N-1; First side-play amount acquiring unit 342 is for obtaining the shift offset of initial guide point; First region of search determining unit 343 is for using the initial offset of the shift offset of row initial guide point as N-th row first object point, by in matching image data, with the position after the position of N-th row first object point skew initial offset for core searching position, determine region of search based on core searching position; First search unit 344 for searching for the match point of the N-th row first object point in destination image data in region of search; Second impact point determining unit 345 for after the match point finding N-th row first object point, successively with N-th row first object point colleague both sides node for impact point; Pilot point determining unit 346 selects node that matching degree is the highest as pilot point for calculating in the node of shift offset around impact point; Second side-play amount acquiring unit 347 is for obtaining the shift offset of pilot point; Second region of search determining unit 348 is for using the initial offset of the shift offset of pilot point as impact point, by in matching image data, with the position after the position of impact point skew initial offset for core searching position, determine region of search based on core searching position; Second search unit 349 for mating impact point in destination image data in region of search, and obtains the match point of the N-th row impact point in destination image data.

The present embodiment also discloses displacement detecting method in a kind of compressive resilience imaging, and the method is applicable to above-mentioned supersonic imaging apparatus, the displacement detector especially in image processor 3.The thinking of the method is: at displacement detection, based on two frame image datas before and after compression, when utilizing guiding zero phase to estimate (guided phasezero estimation, GPZE) algorithm displacement calculating result, using node the highest for matching degree as initial guide point.GPZE algorithm can be the Chinese patent " displacement detecting method in a kind of elastogram, Apparatus and system " of 201110159110.9 see application number.

GPZE algorithm carrys out the phase place of the cross-correlation function detected more quickly and accurately before and after compression between two frame signals mainly through the thinking of guiding search, and the corresponding relation of deriving between this phase place and length travel amount, calculate the length travel amount between two frame signals, while greatly reducing amount of calculation, ensure that the quality of Displacement Estimation.In addition, the scope of displacement detecting is also expanded, and GPZE algorithm is not only applicable to thin tail sheep situation, is also applicable to the situation of Large travel range.

Suppose being expressed as by matching image signal after the target image before compressing and compression:

$f_u(t,x)=A_u(t,x)e^{j({\mathrm{\ω}}_{c}t+\mathrm{\θ})}$

$f_c(t,x)=A_c(t-u_y,{x-u}_x)e^{j[{\mathrm{\ω}}_c(t-u_y)+\mathrm{\θ}]}=A_c({t-u}_y,x-u_x)e^{j[{\mathrm{\ω}}_c(t-n{T}_c/2-\mathrm{\τ})+\mathrm{\θ}]}$

Wherein, ω _cbe signal center frequency, θ is signal initial phase, T _cfor the signal period, with mid frequency ω _ccorresponding.U _xfor compressing the space transversal displacement caused, u _yfor compressing the vertical misalignment amount caused.U _yalways u can be expressed as _y=τ+nT _cthe form of/2, wherein n is integer, and τ is always distributed in-T _c/ 2 ~ T _cin/2 scopes.

After quadrature demodulation, become baseband signal, the plural form of baseband signal can be expressed as:

f _ub(t,x)=A _u(t,x)e ^jθ

$f_{\mathrm{cb}}(t,x)=A_c(t-u_y,{x-u}_x)e^{j(-{\mathrm{\ω}}_c\mathrm{\τ}-{\mathrm{\ω}}_c{\mathrm{nT}}_c/2+\mathrm{\θ})}$

Or,

S _u=I _u+iQ _u

S _c=I _c+iQ _c

So the cross-correlation function between baseband signal is expressed as:

$\begin{array}{c}{R}_b(u_0,x_0)=\underset{t_0-\mathrm{\Δt}}{\overset{t_0-\mathrm{\Δt}}{\∫}}{f}_{\mathrm{ub}}(t,x)\·{f_{\mathrm{cb}}}^{*}(t+u_0,x+x_0)\mathrm{dt}\\ =\underset{t_0-\mathrm{\Δt}}{\overset{t_0-\mathrm{\Δt}}{\∫}}{A}_u(t,x)e^{\mathrm{j\θ}}\·A_c({t-u}_y+u_0,{x-u}_x+x_0)e^{-j[{\mathrm{\ω}}_c(-{\mathrm{nT}}_c/2-\mathrm{\τ})+\mathrm{\θ}]}\mathrm{dt}\\ =e^{\mathrm{jn\π}}(\underset{t_0-\mathrm{\Δt}}{\overset{t_0-\mathrm{\Δt}}{\∫}}{A}_u(t,x)\·A_c({t-u}_y+u_0,x-u_x+x_0)\mathrm{dt})e^{{\mathrm{j\ω}}_c\mathrm{\τ}}\\ =e^{\mathrm{jn\π}}{R}_{\mathrm{eb}}(u_0,x_0)e^{j{\mathrm{\ω}}_c\mathrm{\τ}}\end{array}$

Wherein, u ₀and x ₀longitudinal direction between two frame target datas during representative calculating cross-correlation function and transversal displacement, R _ebbe two frame data envelope signal between cross-correlation function.

Envelope cross-correlation function R is made as long as find _b(u ₀, x ₀) maximum time u ₀and x ₀namely longitudinal, the transversal displacement between two frame envelope data is found, the match point of impact point in target image before utilizing vertical and horizontal side-play amount can find compression upon compression in image, thus calculate cross-correlation phase place according to impact point and match point, and calculate final mean annual increment movement result further.

Each modal displacement testing process as shown in Figure 4, comprises the following steps:

Step 410. obtains view data

The displacement detecting of the present embodiment calculates based on the data obtained.Obtain two frame image datas, respectively as compression before destination image data and compression after by matching image data.

Particularly, a pair I/Q baseband signal frame data is obtained:

S _u=I _u+iQ _u

S _c=I _c+iQ _c

Wherein, S _ufor the destination image data before compression, S _cfor compression after by matching image data; I _uand Q _uand I _cand Q _cbe respectively destination image data and by the signal parameter of matching image data.

The calculating of each frame displacement result needs to utilize two frame I/Q baseband signal data, and gained displacement result refers to the space relative displacement between two frame signals.Above-mentioned two frame baseband signal data can be two continuous frames data, and can be two frame data having a frame interval, frame period quantity be preset by system, or determines according to the selection of user.Using two frame data of certain intervals, effectively can adjusting the amount of displacement between two frame data for calculating, make the final strain pattern better quality obtained.

Step 420. finds the match point of each node of destination image data

In destination image data step 410 obtained, the node of the first row is as impact point, finds corresponding match point respectively in by matching image data; For the node of other each row in destination image data, using the highest node of lastrow each node matching degree as initial guide point, then find each node of one's own profession by the match point in matching image data based on initial guide point.

Even if it will be apparent to those skilled in the art that through down-sampled, the sample rate of baseband signal is still higher, and displacement is general less, and the displacement difference between neighbouring sample point is very small.In order to make to reduce amount of calculation, divide the position of displacement detecting estimation point (node) in advance, such as, image can be divided at equal intervals some continuous and nonoverlapping piece, be that node carries out mating and displacement detecting with block, effectively can avoid or reduce redundant computation amount.Be illustrated in figure 5 the view data gridding schematic diagram in the embodiment of the present application, divide n altogether capable, stain (i.e. grid node place) is the position possible carrying out Displacement Estimation, in figure, black line represents baseband signal data or envelope data, stress and strain model with the data acquisition sampling point position in target image frame in two frame signals for benchmark.

Longitudinally, from the data of the most shallow degree of depth, get a little after the data acquisition sampling point (or every certain depth) of some, this position needs the node or the impact point that carry out Displacement Estimation, and this longitudinal separation quantity is preset by system.

Laterally, from center probe scan-data line, get a little after the sample line (or every one fixed width) of some, this position needs the node or the impact point that carry out Displacement Estimation, and this lateral separation quantity is preset by system.

After grid division, the amount of calculation obtained needed for strain pattern will greatly reduce, particularly when horizontal and vertical interval is very large time.But interval is too large to be had a certain impact to image quality, the spatial resolution of Displacement Estimation result and final image can be affected.

Destination image data node is being node maximally related with this node by the match point in matching image data, by calculating the maximally related node with this node in relevance algorithms upon compression image, and it can be used as the match point of impact point.

Step 430. calculates cross-correlation phase place

After the match point finding destination image data interior joint, the cross-correlation phase place of this node and match point can be calculated.Phase calculation uses I, Q data.

Take out by target image with by the parameter I of relevant position in matching image and Q, then trying to achieve phase place is:

Wherein, (i, j) relative coordinate or the position of back end is represented, same (i, j) relative position represented in two frame data is identical, such as, suppose that (0,0) represents the point in the Nuclear Data lower left corner in the target image, so (0,0) also represents the point in the Nuclear Data lower left corner in by matching image.

In above formula, the computational methods of n are relevant with the final mean annual increment movement estimated result of lastrow node, suppose that the final vertical displacement result of lastrow node is u _y, then:

$n=\mathrm{round}\left(\frac{u_y}{T_c/2}\right)$

Wherein, round represents round (also can adopt the mode rounding up or round downwards).

In addition, the range of results calculated due to arctan function exists between, also need according to its above formula the positive and negative situation of Middle molecule denominator by calibration of the output results in-π ~ π scope.By the PHASE DISTRIBUTION rule of trigonometric function, the symbol of above-mentioned molecule with correspondence, the symbol of denominator with corresponding.

Step 440. displacement calculating result

After above-mentioned calculating, the displacement result of present node (or Displacement Estimation point) is:

Wherein, T _cfor the signal period, with signal center's angular frequency _ccorresponding.NT _cthe introducing of/2, compensate for the aliasing of phase calculation, thus makes this algorithm both be applicable to the situation of thin tail sheep, is also applicable to the situation of Large travel range.

The mode of above-mentioned displacement result all in units of the sampling time represents, also can be converted into physical length unit to represent, the two is one to one.

In the present embodiment, at step 420 which, adopt when searching the match point of each node in target image in by matching image and find each node of one's own profession by the method for the match point in matching image data based on initial guide point, please refer to Fig. 6, comprise the following steps:

Step 421. searches the match point of the first row impact point

Using the node of the first row in destination image data as impact point, find in by matching image data corresponding match point when every two field picture is divided into the capable grid of m row n respectively, each lattice is a node, the employing Block-matching when mating impact point.。Match point is maximally related node with impact point, also can obtain the shift offset of impact point and match point, i.e. the position coordinates variable quantity of impact point and match point simultaneously.

After the match point of the first row impact point is determined, perform step 430 on the one hand, calculate cross-correlation phase place according to impact point and match point, perform following steps on the other hand, search the match point of the second row in destination image data and each node of each row below thereof.

Step 422. determines the row initial guide point of N-th row

From the second row, adopting the method for pilot point when searching the match point of impact point, suppose the match point searching each node of N-th row, N be successively from 2 to n integer, n is the line number of a two field picture division.Now, during N-1 is capable, each node has found match point, therefore first in the matching result of the capable each node of N-1, find out the row initial guide point of the highest node of matching degree as N-th row, each node searching N-th row based on row initial guide point is by the match point in matching image data.In a kind of instantiation, the quality factor of node (QF) can be adopted to pass judgment on the matching degree of this node, and quality factor is higher, and matching degree is higher.

In one embodiment, quality factor can be passed through S _uand S _cthe correlation coefficient of Using statistics calculates, and absolute difference summation (Sum-Absolute Difference, SAD), the difference of two squares summation (Sum-Square Difference, SSD) etc. also can be adopted to try to achieve, and a kind of preferred account form is:

$\mathrm{QF}=\frac{\left|\mathrm{\Σ}{S}_u\stackrel{\‾}{S_c}\right|}{\sqrt{\mathrm{\Σ}{\left|S_u\right|}^2\mathrm{\Σ}{\left|S_c\right|}^2}}$

Wherein represent that plural number gets conjugation.The value of quality factor QF, between [0,1], actual can be quantized to other numerical rangies according to custom when using and judges, such as, be multiplied by 100 and round, and makes quality factor become integer between 0-100.

In other instantiation, other mode also can be adopted to pass judgment on the matching degree of this node, such as, adopt the mode of mean strain size or node space position to pass judgment on the matching degree of this node.

Using node the highest for matching degree in capable for N-1 as initial guide point, thus determine the first object point of N-th row in destination image data.Alleged first object point refers in all nodes of N-th row, and first in order to search for the node of match point in by matching image data as impact point.

Step 423. determines the first object point of N-th row

Determine N-1 capable in the highest node of matching degree as initial guide point after, in N-th row, usually choose the point close with initial guide point as first object point.Preferred mode is, to choose in N-th row with the node of initial guide point same column as first object point, because the consecutive points of same column the most easily influence each other, if the coordinate supposing initial guide point is (t-1, x), then first object point can be selected (t, x).The selection of first object point also can be the node with initial guide point adjacent column in N-th row, but the drawback of this mode is, initial guide point is when first or terminal column, likely cause N-th row cannot select first object point: such as, if select the node of the previous column of initial guide point as first object point, when initial pilot point is positioned at first, first object point just cannot be selected in N-th row; If select the node of the rear string of initial guide point as first object point, when initial pilot point is positioned at terminal column, first object point just cannot be selected in N-th row.

Step 424. determines the region of search of first object point

The guiding thinking of the application is, is guide the region of search determining one's own profession first object point according to the shift offset of the highest node of lastrow matching degree.Therefore, the shift offset of row initial guide point should first be obtained.In the present embodiment, obtain vertical misalignment amount and the transversal displacement of row initial guide point respectively, in other embodiments, also only can obtain the vertical misalignment amount of initial guide point.Suppose that the coordinate of first object point is for (t, x), the side-play amount of the vertical and horizontal of initial guide point is respectively u ₀and x ₀, then the initial offset setting first object point is u ₀and x ₀(vertical and horizontal), namely first object point is being (t+u by the core searching position of the match point in matching image data ₀, x+x ₀).

After determining that first object point initial displacement side-play amount obtains the core searching position of first object point, just region of search can be determined.By in matching image frame data, centered by this core searching position (or for benchmark), near taking out or around a blocks of data calculates.

As shown in Figure 7, suppose that first object point is being (t+u by the core searching position in matching image frame data ₀, x+x ₀), the size of this blocks of data is preset by system, such as sets x left in the transverse direction of core searching position ₁, set x to the right ₂; Longitudinally upwards set u ₁, set u downwards ₂.After setting, first object point by the region of search in matching image frame data can by longitudinal region between and determine between transverse region, wherein, be [t+u between longitudinal region ₀-u ₁, t+u ₀+ u ₂], be [x+x between transverse region ₀-x ₁, x+x ₀+ x ₂], as the region of Fig. 7 black line institute frame.

According to clinical experience, destination organization is darker, and its coefficient of elasticity is less, namely more soft, and the length travel side-play amount showed can be larger.Therefore the present embodiment also discloses a kind of determination scheme of region of search: wherein same such scheme between transverse region, and for longitudinal direction, upwards no longer setting search scope, namely sets x left in the transverse direction of core searching position ₁, set x to the right ₂; Longitudinally set u downwards ₃.After setting, first object point by the region of search in matching image frame data is being: be [t+u between longitudinal region ₀, t+u ₀+ u ₃], be [x+x between transverse region ₀-x ₁, x+x ₀+ x ₂].

Region of search setting is less, and amount of calculation is less.

Step 425. searches the match point of first object point

The point searching out dependency in region of search maximum is match point, based on the thinking of Block-matching (block-matching), in by matching image data, searches for the position maximum with first object point dependency.The distinguishing rule that in search, dependency is maximum can adopt SAD method, NCC method etc.The position that SAD is minimum or the maximum position of NCC are the maximum position of dependency.Also other similar distinguishing rules can be adopted.

Can step 430 and 440 be carried out after finding the match point of first object point, calculate cross-correlation phase place and displacement result.

Step 426. searches the pilot point of other node of N-th row

After the match point of first object point finding N-th row, successively with the node of N-th row first object point colleague both sides for impact point, search the match point of impact point in by matching image data.For selected impact point, the concrete grammar searching its match point in by matching image data comprises:

Calculating in the node of shift offset around selected impact point selects node that matching degree is the highest as pilot point.Other node near impact point is referred to around impact point, it can be the node of lastrow, also can be the node of one's own profession, such as successively with the point on the left of first object point for impact point mates, then the right side of impact point, above, search the highest node of matching degree in top-right node as pilot point.The condition that the node being selected as pilot point should meet is: calculate shift offset.In other embodiments, also can priority be set, as setting matching degree threshold value, when the matching degree of a certain node do not reach require time, even if this node is nearest from impact point, also can directly abandon this node, and select from impact point farther but the node that matching degree is high as pilot point.

Step 427. determines the region of search of other impact points of N-th row, mates

After determining pilot point, obtain the shift offset of pilot point, and using the initial offset of the shift offset of pilot point as impact point.By in matching image data, with the position after the position of impact point skew initial offset for core searching position, region of search is determined based on core searching position, in region of search, impact point in destination image data is mated, and obtain the match point of the N-th row impact point in destination image data.Concrete grammar can refer to such scheme, does not repeat them here.

Can step 430 and 440 be carried out after finding the match point of selected node, calculate cross-correlation phase place and displacement result.

Step 428. judges whether that all nodes of N-th row have all mated, and if so, then performs step 429, otherwise continues to perform step 426.

Step 429., when the displacement result of all nodes of N-th row all calculates complete, needs to judge whether N-th row is last column, if so, then terminates; Otherwise carry out the calculating of next line.

In the present embodiment, when carrying out displacement detecting to the node of the second row and back row thereof, all select the point the highest with the matching degree of this near nodal as pilot point owing to mating at every turn, when the pilot point of fixing therefore can be avoided to make a mistake, mistake is extended on follow-up node, decreases the accumulative effect of mistake in computation.

After obtaining the displacement result at all nodes in destination image data (or Displacement Estimation point) place, along the longitudinal gradient is asked to displacement result data, can obtain straining result, i.e. strain value.

In strain post processing, certain error correction can be carried out to the strain result of gained, such as detect abnormal trip point wherein or apparent error point corrects; Space smoothing can also be carried out, to improve image display effect to it; Or use different GTGs or color atlas to map, to strengthen picture contrast to it.The operation that other increase picture quality can also be carried out.

Finally, the strain result of desired zone is carried out output display and becomes strain pattern, this regional organization's elastic difference can be reflected.

After whole frame displacement data has all been calculated, the QF value of whole each node of frame can also be obtained in the lump.By these QF information, can when final image shows for user provide quality information, one is whole frame QF drafting pattern picture, can have a clear understanding of the mass fraction of each node, two is the meansigma methodss that can calculate whole frame QF, is convenient to the quality understanding a width elastic image on the whole.

It will be appreciated by those skilled in the art that, in above-mentioned embodiment, all or part of step of various method can be carried out instruction related hardware by program and completes, this program can be stored in a computer-readable recording medium, and storage medium can comprise: read only memory, random access memory, disk or CD etc.

Above content is in conjunction with concrete embodiment further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations.For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, some simple deduction or replace can also be made.

Claims (22)

1. a displacement detecting method in compressive resilience imaging, is characterized in that comprising:

Obtain two frame image datas, respectively as compression before destination image data and compression after by matching image data;

Using the node of the first row in destination image data as impact point, in by matching image data, find corresponding match point respectively;

According to the impact point of the first row and the displacement result of each node of match point calculating the first row thereof;

Calculate the displacement result of each node of N-th row in destination image data, wherein N be successively from 2 to n integer, n is the line number that a two field picture divides, and comprising:

The highest node of matching degree is found out as row initial guide point in the matching result of the capable each node of N-1;

The each node searching N-th row based on row initial guide point is by the match point in matching image data;

According to each node of N-th row and the displacement result of each node of match point calculating N-th row thereof.

2. the method for claim 1, is characterized in that, every two field picture is divided into the capable grid of m row n, and each lattice is a node, adopts Block-matching when mating impact point.

3. the method for claim 1, is characterized in that, described view data is the ultrasonic echo data ultrasound image data after treatment received.

4. the method according to any one of claim 1-3, is characterized in that, described match point is maximally related node with impact point.

5. method as claimed in claim 4, is characterized in that, adopt the quality factor of node to pass judgment on the matching degree of this node, quality factor is higher, and matching degree is higher.

6. method as claimed in claim 5, it is characterized in that, the quality factor of node is the correlation coefficient of this node and its match point.

7. method as claimed in claim 4, it is characterized in that, the calculating of each modal displacement result comprises: the ultrasonic radio frequency complex signal based on this node and match point thereof calculates the cross-correlation phase place of this node, based on the final mean annual increment movement result of this node of cross-correlation phase calculation.

8. the method according to any one of claim 1-7, is characterized in that, each node searching N-th row based on row initial guide point is being comprised by the match point in matching image data:

The first object point of N-th row in the node determination destination image data that in capable based on N-1, matching degree is the highest;

Obtain the shift offset of initial guide point;

Using the initial offset of the shift offset of row initial guide point as N-th row first object point;

By in matching image data, with the position after the position of N-th row first object point skew initial offset for core searching position, region of search is determined based on core searching position, in region of search, the N-th row first object point in destination image data is mated, and obtain the match point of the N-th row first object point in destination image data;

After having calculated N-th row first object point, successively with the node of N-th row first object point colleague both sides for impact point, search the match point of impact point in by matching image data.

9. method as claimed in claim 8, is characterized in that, N-th row first object point is the close node of node that the matching degree capable with N-1 is the highest.

10. method as claimed in claim 9, is characterized in that, N-th row first object point is the node of the node same column that the matching degree capable with N-1 is the highest.

11. methods as claimed in claim 8, is characterized in that, successively with the node of both sides for impact point, the step of searching the match point of impact point in by matching image data comprises:

Calculating in the node of shift offset around impact point selects node that matching degree is the highest as pilot point;

Obtain the shift offset of pilot point;

Using the initial offset of the shift offset of pilot point as impact point;

By in matching image data, with the position after the position of impact point skew initial offset for core searching position, region of search is determined based on core searching position, in region of search, impact point in destination image data is mated, and obtain the match point of the N-th row impact point in destination image data.

12. methods as claimed in claim 8, it is characterized in that, region of search is the region formed after longitudinally offseting downward setting value from core searching position to core searching position, or region of search is the region formed after periphery offset setting value centered by core searching position.

Displacement detector in 13. 1 kinds of compressive resilience imagings, is characterized in that comprising:

Image collection module, for obtaining two frame image datas, respectively as compression before destination image data and compression after by matching image data;

First matching module, for using the node of the first row in destination image data as impact point, find corresponding match point in by matching image data respectively;

Search module, for finding out the highest node of matching degree as row initial guide point in the matching result of the capable each node of N-1, wherein N is the current line needing to search match point, N be successively from 2 to n integer, n is the line number that a two field picture divides;

Second matching module, for each node of searching N-th row based on row initial guide point by the match point in matching image data;

Displacement computing module, for calculating the displacement result of each node of the first row according to the impact point of the first row and match point thereof, and calculates the displacement result of each node of N-th row according to each node of N-th row and match point thereof.

14. devices as claimed in claim 13, is characterized in that, described view data is the ultrasonic echo data ultrasound image data after treatment received.

15. devices as claimed in claim 13, is characterized in that, adopt the quality factor of node to pass judgment on the matching degree of this node, quality factor is higher, and matching degree is higher.

16. devices according to any one of claim 13-15, it is characterized in that, described match point is maximally related node with impact point.

17. devices as claimed in claim 16, is characterized in that, displacement computing module calculates and comprises:

Cross-correlation phase calculation unit, for calculating the cross-correlation phase place of this node based on the ultrasonic radio frequency complex signal of this node and match point thereof;

Displacement computing unit, for the final mean annual increment movement result based on this node of cross-correlation phase calculation.

18. devices according to any one of claim 13-17, it is characterized in that, the second matching module comprises:

First object point determining unit (341), for the first object point of N-th row in the node determination destination image data that matching degree in capable based on N-1 is the highest;

First side-play amount acquiring unit (342), for obtaining the shift offset of initial guide point;

First region of search determining unit (343), for using the initial offset of the shift offset of row initial guide point as N-th row first object point, by in matching image data, with the position after the position of N-th row first object point skew initial offset for core searching position, determine region of search based on core searching position;

First search unit (344), for searching for the match point of the N-th row first object point in destination image data in region of search;

Second impact point determining unit (345), for after having calculated N-th row first object point, successively with N-th row first object point colleague both sides node for impact point;

Pilot point determining unit (346), selects node that matching degree is the highest as pilot point for calculating in the node of shift offset around impact point;

Second side-play amount acquiring unit (347), for obtaining the shift offset of pilot point;

Second region of search determining unit (348), for using the initial offset of the shift offset of pilot point as impact point, by in matching image data, with the position after the position of impact point skew initial offset for core searching position, determine region of search based on core searching position;

Second search unit (349), for mating impact point in destination image data in region of search, and obtains the match point of the N-th row impact point in destination image data.

19. devices as claimed in claim 18, is characterized in that, N-th row first object point is the close node of node that the matching degree capable with N-1 is the highest.

20. devices as claimed in claim 19, is characterized in that, N-th row first object point is the node of the node same column that the matching degree capable with N-1 is the highest.

21. devices as claimed in claim 18, it is characterized in that, region of search is the region formed after longitudinally offseting downward setting value from core searching position to core searching position, or region of search is the region formed after periphery offset setting value centered by core searching position.

22. 1 kinds of supersonic imaging apparatus, is characterized in that comprising:

Probe, for receiving ultrasonic echo to scanning objective emission ultrasound wave;

Signal processor, for processing ultrasonic echo, generates ultrasound image data;

Image processor, for processing ultrasound image data, and generates elastic image, and described image processor comprises:

Displacement detector according to any one of claim 13-21;

The displacement result of each node that deformation based checkout gear detects generates the elastic image generating apparatus of elastic image.