3.1. An Analysis of the Conventional Method
According to the mean value theorem for integration [
26], the errors of Equation (6) are affected not only by the distribution of the temperature but also by the maximum size of the differential element or the pixel in the context of image reconstruction,
, as well as by how the characteristic parameter for a pixel is defined. According to the matrix theory [
27], for a linear system of equations
to have a unique least squares solution, the coefficient matrix
must have full column rank. In particular, there should be no column(s) with all zero members.
It has been a widely adopted practice to virtually divide the measurement zone into small pixels by equally spaced horizontal lines and vertical lines. If a square interrogation area is divided by
horizontal lines and
vertical lines, then
pixels will be counted within this area, named as
n. Accordingly, there will be
columns in matrix
. For a circular interrogation area, however, the interrogation area may not contain all the pixels thus generated, some pixels in the corners will stay outside the circular area. This will be case if the division
exceeds a certain value, since the smaller the pixels, the more the pixels that can be contained between the circle and the square area in the corners. If this happens, there will be zero column(s) in matrix
, implying that there will not be a unique solution for
. Therefore, we name this value the limit of divisions for a unique solution, LDUS for short. It is not difficult to work out the number of the outsider pixels for a given division
. The number of the pixels inside the circular area represented by matrix
corresponding to each
can be given by the
Table 1.
For example, for , where the number of measurements is 66, must be smaller than 9 to limit the number of pixels inside the measurement zone to a valve smaller than the number of data acquired from the measurement, if a unique least squares solution is required for .
If
is chosen to be small to meet the requirement for a unique square solution, then the errors of the solution will be inevitably large, due to the large size of the pixels. However, if
is too large, the requirement on
for a unique least squares solution for
will not be satisfied, resulting in underdetermined problems. The Tikhonov regularization method very often is an effective method to alleviate the underdetermined problems. The essence of the Tikhonov method is to acquire an optimized solution for the following problem:
where λ ≥ 0 is the regularization parameter. The above model first minimizes the error
, while the penalty term
will force a unique solution.
As shown in
Figure 3, a circular measurement zone of a diameter d can be evenly divided into
by
squares. According to the presence of the sound beams, the pixels can be categorized into two sets of indexes,
, and
. For a conventional 12-electrode sensor, the index-set, namely
, will be a nonempty set when
. The number of elements in
will increase with
. For example,
will have 2 elements for
and 18 elements when
.
Theorem 1. If is not an empty set, then the solution of the optimization model is not the solution of the original problem.
Proof of Theorem 1. If Index_2 is not an empty set, then at least one column in matrix will have all-zero members, if this column’s index corresponds to a pixel indexed in . Let vector be the same as except setting all the pixel values to zero if the pixels belong to . Then, using the definitions of the terms in Equation (9), we have , and . Consequently, for any regularization factor , the value of a pixel in the solution of Equation (9) must be zero if the pixel belongs to . This implies that, if is not an empty set, the solution of the optimization problem, i.e., Equation, will not be the solution of the original problem. Therefore, when using Equation (9) for an optimized solution, the size of each pixel must not be smaller than , i.e., must be smaller than 23. □
For the same reason, if is not an empty set, solutions using , in the Equation as the regularization term will not be the solution to the original problem.
3.2. The Neighboring Cells Regularization Model and Solution Scheme
For more refined images, needs to be increased for more pixels. However, as has been discussed above, a large number of will make nonempty, thus exceeding the critical limit for a true solution to the original problem. To solve this problem, we propose a method to maintain as an empty set even when is increased beyond the aforementioned critical limit.
Without loss of generality, suppose the distribution of the medium presents a continuous 2D function inside the measurement area, in which a pixel
is located at
, as described in
Figure 4.
Assuming
has a continuous first derivative, an approximation function can be established using the first-order Taylor expansion as follows.
in which:
The sum of all the is 1. The values of the coefficients are determined because the distances between are unity (or can be normalized to unity), and between , they are . The smaller the distance between two points, the smaller the difference between their function values. When expressed by the function value of adjacent points, if the distance is large, the weight coefficient is small; if the distance is small, the weight coefficient is large.
For cells on the lower boundary,
in which:
For cells on the upper boundary,
in which:
For cells on the left boundary,
in which:
For cells on the right boundary,
in which:
We express as a column vector, i.e., . That means there will be n equations for which are expressed above by the function value of its neighbor. Then, the left part of the Equation (11) (or 14, 17, 20, 23) could be moved to the right part. These n equations can be expressed as , where is the constructed matrix. Here, the elements in the row of in each column are
- (1)
at the position of ,
- (2)
at the position of ,
- (3)
-in the column, which is also the diagonal element of ,
- (4)
0 at the other positions.
Therefore, we can build the following optimization model to solve the problem (
neighboring cells regularization model)
Since is column full rank, the equation has a unique solution. In other words, by adding the regularization term , Equation (26) has a unique solution, and an approximation of the original problem is consequently guaranteed.
As an example, we applied the above method to the reconstruction of a temperature field. The case is a common one—a 2D continuous temperature distribution, for which a minimum solution with a regularization term can be expected. In this new model, the distances among the pixel and neighboring pixels , were taken into account when constructing matrix , which effectively removed the ambiguity of the element values, caused by the all-zero-element in certain columns due to the absence of the probing beam passing the pixels, as described above.