CN110827215A - ERT image reconstruction artifact removing method based on fuzzy clustering - Google Patents

Abstract

The invention relates to an ERT image reconstruction artifact removing method based on fuzzy clustering, which comprises the following steps: 1) respectively calculating the gray level vector of the measured field by utilizing at least two image reconstruction algorithms; 2) combining the gray vectors in the step 1) to form a gray matrix as a characteristic vector in a clustering algorithm; 3) applying a fuzzy clustering algorithm to the gray matrix, and performing clustering analysis on the gray matrix; 4) finding out cluster clusters respectively representing a target, a background and an artifact according to the statistical characteristics expressed by the cluster clusters; 5) cluster clusters representing artifacts are classified into cluster clusters representing background.

Description

ERT image reconstruction artifact removing method based on fuzzy clustering

Technical Field

The invention belongs to the field of electrical tomography image reconstruction algorithms, and particularly relates to an ERT image reconstruction artifact removing method based on fuzzy clustering.

Background

An Electrical Tomography (ET) technique is a technique for reconstructing the internal Electrical property distribution of a measured object field, and the technique adopts simple hardware equipment structure and low cost, and the detection process has the characteristics of no radiation and non-invasiveness. The technology has wide application prospect in the fields of medical clinical monitoring, industrial measurement and the like.

In Electrical Resistance Tomography (ERT), current excitation is applied to a measured object field by using array electrodes, boundary voltage data are measured, and conductivity distribution in a measured field domain is reconstructed. Taking a 16-electrode electrical resistance tomography sensor as an example, a method of adjacent current excitation and adjacent voltage measurement is adopted. At each excitation, the excitation electrode was removed for a total of 13 measurements. Such as: the voltage signals of the electrode pairs 3, 4, 5, 6, …, 15 and 16 are measured by taking the No. 1 and No. 2 electrodes as excitation electrodes. There are 16 excitations in total, i.e. 13 × 16 — 208 measurements. In order to achieve both temporal resolution and spatial resolution, the sensitive field to be measured is generally divided into 812 pixels. The problem of reconstructing the electrical resistance tomography image is to solve the gray value of the 812 pixel points.

The spatial resolution of the reconstructed image is poor due to the "soft-field" effect of ERT and the inherent ill-qualification of the inverse problem. The traditional linear back projection algorithm (LBP) is to perform back projection of voltage data according to the relationship between the boundary voltage value and the projection area of the measured field, but due to the limitation of hardware equipment, the back projection data is far less than the gray value to be solved, so that the reconstructed image has larger artifacts, the boundary of the target and the background is fuzzy, and the spatial resolution is poor.

Disclosure of Invention

The invention aims to overcome the defects in the prior art, and provides an ERT image reconstruction artifact removing method based on fuzzy clustering, which can effectively reduce the influence of artifacts on image resolution. All pixel points of the reconstructed image can be classified into 3 types: objects, background, and artifacts. The artifacts influence the imaging resolution to a great extent, and the quality of the reconstructed image can be improved to a certain extent by effectively processing the artifacts. Considering that the result obtained by the above algorithm is a gray value, i.e. numerical data, the gray value can be divided by using a clustering algorithm, and then the gray value is further processed according to the division result, so as to obtain a more ideal imaging effect.

The technical scheme of the invention is as follows:

an ERT image reconstruction artifact removing method based on fuzzy clustering comprises the following steps:

1) respectively calculating the gray level vector of the measured field by utilizing at least two image reconstruction algorithms;

2) combining the gray vectors in the step 1) to form a gray matrix as a characteristic vector in a clustering algorithm;

3) applying a fuzzy clustering algorithm to the gray matrix, and performing clustering analysis on the gray matrix;

4) finding out cluster clusters respectively representing a target, a background and an artifact according to the statistical characteristics expressed by the cluster clusters;

5) cluster clusters representing artifacts are classified into cluster clusters representing background.

Due to the adoption of the technical scheme, the invention has the following advantages:

1. compared with the traditional electrical tomography algorithm, the algorithm has the advantages that a plurality of classical imaging algorithms are fused, the advantages of the classical imaging algorithms are complemented, the gray value is further subjected to cluster analysis on the basis, the influence of artifacts is reduced, the boundary between the target and the background is clearer, and the spatial resolution of imaging is improved.

2. The clustering algorithm adopts a fuzzy clustering algorithm instead of a traditional hard division algorithm, so that the target edge is smoother and closer to the actual edge.

Drawings

FIG. 1 is a diagram of an original object field model in an embodiment of the invention;

FIG. 2 is a result of image reconstruction using the LBP algorithm for the simulation model shown in FIG. 1;

FIG. 3 is a result of image reconstruction using the algorithm of the present invention for the simulation model shown in FIG. 1.

Detailed Description

The invention aims to overcome the defect of low spatial resolution of the conventional image reconstruction algorithm, and provides an ERT image reconstruction artifact removing method based on fuzzy clustering. The fuzzy clustering object is a set of data, and the data are divided by extracting the characteristic vector of the data, so that the intra-class difference is small, and the inter-class difference is large. By carrying out cluster analysis on the gray matrix, the influence of the artifact on the image resolution is reduced.

The method for removing the ERT image reconstruction artifact based on fuzzy clustering according to the present invention is described in detail below with reference to the accompanying drawings and embodiments. The method comprises the following specific steps:

1) fig. 1 shows a flow pattern containing a discrete target medium, and for the measured object field shown in fig. 1, if the regularization (TK) algorithm and LBP algorithm are used to calculate the gray vector, the gray value of the reconstructed image can be shown in formula (1).

G_T＝{g_t1,g_t2,...,g_t812}，G_L＝{g_l1,g_l2,...,g_l812} (1)

Wherein G is_TRepresenting the gray vector, G, obtained by the TK algorithm_LRepresenting the gray vector resulting from the LBP algorithm.

An image reconstructed by the LBP algorithm is shown in fig. 2, an area with a small upper circular-like area represents a target, an area with a largest lower area is a background area, and a gray area between the upper circular-like area and the lower circular-like area is an artifact area.

2) Combining the gray level vectors in 1) to form a gray level matrix as the feature vector in the cluster, i.e. combining the gray level vectors G_T，G_LForming a gray matrix and calculating a gray mean vectorAs shown in equation (2):

3) and (3) carrying out clustering analysis on the gray matrix in the step 2) by using a fuzzy clustering algorithm. The fuzzy clustering algorithm needs to give a clustering number, and as can be seen from the meaning of the inverse problem, the clustering number can be set to 3, that is, data is divided into three types: target G_oBackground G_bAnd an artifact G_tI.e. by

4) And calculating the statistical characteristics of each cluster. The 3 clusters in step 3) have different statistical characteristics: the class with the largest mean is the target, in contrast to the class with the smallest mean being the background, the class with the largest variance represents the artifact. According to the characteristics, the meaning of each cluster in the clustering result can be clarified.

5) Assign clusters representing artifacts to clusters representing background, i.e. to say

The gray value of the middle pixel point is assigned asThe average gray value of the middle pixel point is recorded as

I.e. the final gray vector is:

6) using the formula (4)

Image inversion is performed as shown in fig. 3. The artifact region in fig. 3 is effectively reduced, closer to the original object field distribution.

Claims (1)

1. An ERT image reconstruction artifact removing method based on fuzzy clustering comprises the following steps:

1) and respectively calculating the gray level vector of the measured field by utilizing at least two image reconstruction algorithms.

2) Combining the gray vectors in the step 1) to form a gray matrix as a characteristic vector in a clustering algorithm;

3) applying a fuzzy clustering algorithm to the gray matrix, and performing clustering analysis on the gray matrix;

4) finding out cluster clusters respectively representing a target, a background and an artifact according to the statistical characteristics expressed by the cluster clusters;

5) cluster clusters representing artifacts are classified into cluster clusters representing background.

Images