JPS63303474A - Digital image processing method - Google Patents

Abstract

PURPOSE:To realize sharpness corresponding to the chevron pattern of a light- and-shade level using the maximum value in the neighborhood of each image element, by applying a specific equation on the density value of an input image. CONSTITUTION:To a processor 10, the input image F={fi, j} is inputted. In such a case, fi, j represents the density value in the image element (i, j), and (i) and (j) represent natural numbers. the processor 10 performs a processing shown in equation I on the input F, and outputs an output in which natural sharpness corresponding to the chevron pattern of the input image is applied to a display medium 12. Also, it performs the processing by equation II for the result of the equation I, then, performs the compression of a dynamic range. In such a case Ni, j represents the set of values fi, j in the neighborhood of a point (i, j), and Max(Ni, j) the maximum value of the Ni, j, and (alpha) and (beta) represent parameters possible to be changed from a console 14.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a digital image processing method that can simultaneously compress, normalize, and sharpen the level dynamic range of an input image.

(Conventional technology) For simplicity, the case of a one-dimensional image will be explained.

Differential processing has been widely used as a conventional sharpening technique. This method converts the output concentration value sequence (y,) to the human concentration value sequence (xo) to Yn"Xn-172 (Xn-1"Xn+1)
This is a calculation method using (+)Jl=positive integer. Figure 2 shows the human power (X,), and Figure 3 shows the result (y,,) of equation (1).

(Problem to be solved by the invention) The above-mentioned conventional method is effective when 1 is appropriately selected for the shape of the mountain created by the 'c4 light level (() in the figure). a) (8)) If i is inappropriate, the processing will be unnatural (part (d) in the figure).

Moreover, the effect of compressing the dynamic range of the level is insufficient if only this method is used.

It is an object of the present invention to overcome the above-mentioned drawbacks and provide a digital code processing method that compresses the dynamic range.

(Means for Solving the Problems) G [The features of the present invention to achieve the purpose are that the input image F.(fij) (fij is the density value at pixel (i, j), i=1,2 ..., k, j=1.z, ・-,
L, k, L are natural numbers), the following formula (1) or (
The objective is to obtain an output image by applying equations 1) and 2).

0≦α〈l Ni,+ is the set M of values fij in the vicinity of point (i, j), , X(N,) is the maximum value of Nij βΔ0(1-α)M, , (Nij) lk L O≦β, A=−Σ Σhij kLi=l j=1 (Operation) By applying equation (1) in L, the maximum value in the vicinity of each pixel point is used. Q: Is it possible to create a natural sharpening that corresponds to the shape of the mountain at the 7A level? Here, the neighboring point N
The range of ij and the value of α can be adjusted by the operator or the device to obtain the desired results. Furthermore, k, by applying equation (2) to the result of equation (1), each density value is normalized by the sum of the average value of the entire area and the maximum value in the local vicinity, and the dynamic range is calculated. can be compressed.

(Example) An example of an image processing system is shown in FIG. The digital processor IO can be a normal programmed computer or a processor for -Y, and the display medium 2 can be a graphic display, ALSA paper, etc. Here, the parameters α, β+NiJ can be changed as appropriate using human power R from the console 14.

An input image F=(f, J) is input to the processor 10 manually. Here fij is the density value at pixel (i, j)
=1,2,-,k,j=1,2,...,L, k and L
is a natural number.

Processor IO and 1 human power? Then, the process shown in equation (1) is performed to output an output to the display medium 12 that has been naturally sharpened according to the shape of the mountain in the input image.

([1J-αM,,
αku1 Nij is a point, the set M of values f1J in the vicinity of j),
X(N, ,) is the maximum value of Nij.Then, the result of equation (1) is processed by equation (2) to compress the dynamic range.

βΔ10(l−α)M,, (Nij) lk L O≦β, A=−Σ Σh+j i+Li=1 jl The denominator in equation (2) is the average value of the concentration over the entire area and the maximum value in the local vicinity. is the sum of the values obtained by multiplying each of them by an appropriate coefficient, and each density value [H, is normalized by this sum.

The variables in formulas (1) and (2) are
From there, the operator or human labor is used to obtain the desired results. Therefore, multiple output images can be obtained according to the combination of variables.

For example, α=0. If N and J are the entire area and β-0, the original image will be displayed as is, and if Nij is only the adjacent element and α is sufficiently close to 1, only the maximum point will be displayed. Furthermore, by changing β, the compression characteristics of the dynamic range can be controlled.

The results of processing the human effort in Figure 2 using the present invention are depicted in Figure 4. Regarding (a), (b), and (c), the results are the same as in conventional differential processing, and the wide mountain at the base (:) is not unnaturally crushed and is faithfully sharpened. I understand that. In addition, in Fig. 4, α-0,
There are three points in the vicinity of 75, β=1, N, J, two points before and after, and the own point, and since it is a -dimensional display, it is indicated by n instead of the subscript (i, j).

As explained above, according to the present invention, both mountains with narrow bases (gradation level) and mountains with wide bases are naturally sharpened. In other words, in differential processing, a mountain whose base width is not appropriate is deformed unnaturally;
This method does not cause such a problem.

Moreover, in the system example shown in FIG. - It is also an advantage in terms of hardware and software that various output images such as original image display and local maximum point display can be produced in one unit by simply changing parameters.

(Effects of the Invention) In the method of the present invention, since the maximum value in the vicinity is used, natural sharpening is achieved that is faithful to the original image. Further, by changing the parameters α, β, and N+j, various variations such as dynamic range compression and contrast enhancement can be easily obtained. Therefore, it can be applied to all systems that use display processing.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a digital image processing system according to the present invention, FIG. 2 is an example of an input image, FIG. 3 is a diagram showing the result of processing the input image of FIG. 2 using a conventional technique, and FIG.
The figure is a diagram showing the result of processing the input image of FIG. 2 according to the present invention. 10: Arithmetic unit, 12: Display medium, 14: Console, I9, B, C, 22 gray level.

Claims (2)

[Claims] (1) Input image F = (f_i_j) (f_i_j is the density value at pixel (i, j), i = 1, 2, ..., k,
j = 1, 2, ..., L, k, L are natural numbers), the output image H = (h_i_j) is obtained by performing the following processing and natural sharpening according to the shape of the mountain in the input image. Image processing method characterized by obtaining; ▲Mathematical formulas, chemical formulas, tables, etc.▼(1) 0≦α<1 N_i_j is the value f_i_j near the point (i, j)
The set M_a_x(N_i_j) is the maximum value of N_i_j. (2) Input image F = (f_i_j) (f_i_j is the density value at pixel (i, j), i = 1, 2, ..., k,
j=1, 2, ..., L, k, L is a natural number), (1
) process to obtain an output image H = (h_i_j) that has been naturally sharpened according to the shape of the mountain in the input image,
An image processing method characterized by applying equation (2) to H to obtain an output G (g_i_j) with a compressed dynamic range; ▲Mathematical formulas, chemical formulas, tables, etc.▼(1) 0≦α< 1 N_i_j is the value f_i_j near the point (i, j)
The set M_a_x(N_i_j) is the maximum value of N_i_j g_i_j=h_i_j/[βA+(1-α)M_a_
x(N_i_j)](2)▲There are mathematical formulas, chemical formulas, tables, etc.▼