JPH041867A - Picture quality evaluating method - Google Patents

Abstract

PURPOSE:To evaluate blur itself regardless of the level of a picture signal by extracting an edge part of an input image, obtaining the average picture element value of the edge part, and using the average picture element value as an index for the picture quality of input picture elements. CONSTITUTION:An image D of a peripheral part (edge part) obtained when an image B is magnified into an image C is taken out by an exclusive OR of the images B and C or by subtracting the image B from the image C. Then a sum IB of the picture element values of an input image A included in the area of the binarized image B is obtained. The value of IB is divided by an area SB of the image B (pattern part) to obtain the average density MB (average picture element value) of the pattern part. Then a sum ID of the picture element values of the image A included in the area of the image D of an edge part is obtained, and the value ID is divided by an area SD of the image D to obtain the average density MD of a peripheral part. Then an average density ratio R is obtained between a pattern part B and an edge part D. The picture quality is decided by the value of the ratio R. Thus the picture quality is quantitatively evaluated regardless of the level variance of an input picture signal.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image quality evaluation method that focuses on blurring in the peripheral area of an image.

[Conventional technology]

One method for inspecting image quality is to determine the degree of blur in an input image. Since it is difficult to quantify the degree of blur, it is generally determined visually. As a device that automatically performs the determination, one is known that detects the maximum value and minimum value of the image signal and compares it with a preset reference level to determine the quality.

[Problems to be Solved by the Invention] If it is attempted to judge the quality of the image based on the maximum and minimum values of the image signal, accurate judgment cannot be made because the image signal level varies depending on the intensity of the illumination to the imaged object.

In view of this problem, it is an object of the present invention to provide an image quality evaluation method that can evaluate blur itself without being influenced by the image signal level.

[Means to solve the problem]

The image quality evaluation method of the present invention is characterized in that an edge portion of an input image is extracted, an average pixel value of the edge portion is determined, and this average pixel value is used as an index of the image quality of the input pixel.

[Effect]

Since the average pixel value of the edge portion indicates the blurring of the peripheral portion of the image, this is used as an index for image quality evaluation.

〔Example〕

FIG. 1 shows a flowchart of the processing procedure of the image quality inspection method according to the present invention, and FIG. 2 shows a cross section of an image in the pixel value direction along the processing procedure.

First, in step S1, input image A is captured (Fig. 2A)
. The input image is the output of an image input device such as a television camera. Note that FIG. 2A shows an image input object 21 shown in a plan view.
The pixel value (image density) on the line 22 that crosses the is shown.

Next, in step S2, input image A is binarized using an appropriate threshold value to obtain image B (FIG. 2B). This image B
is a reduced image obtained by reducing the peripheral part of the input image A by binarization.

Next, in step S3, an enlargement process (expansion process) is performed on the binarized image B to obtain an enlarged image C (FIG. 2C).

Next, in step S4, the peripheral area (
The image of the edge portion) is extracted by exclusive OR of images B and C or by subtracting B from image C.

Next, the process proceeds to step S5, where the sum IB of the pixel values (signal intensities) of the input image A belonging to the area of the binarized image B (pattern image with no gradation) is calculated, and the value of IB is converted to The average density MB (average pixel value) of the pattern portion is determined by dividing by the area SB of the pattern portion.

MB=IB/SB Note that the area SB may be the number of pixels of the image B. The sum IB of the pixel values of the pattern part is on line 2, as shown in FIG.
This is the sum of the area of the entire area of image B for the area of the shaded portion 23 on 2.

Next, in step S6, the sum ID of the pixel values of the input image A belonging to the area of the image in the peripheral part (edge part) is calculated, and the value ID is divided by the area SD of the image to calculate the average density MD of the peripheral part. (average pixel value).

MD=ID/SD Note that the sum 10 of the peripheral area is the sum of the entire area of the image with respect to the area of the hatched area 24 on the line 22, as shown in FIG. 2 ID.

Next, in step S7, the pattern part (B) and the peripheral part (D)
Find the ratio R of the average concentration.

R=MD/MB This ratio R is a numerical value that quantifies image quality, and in an ideal image without blur, R should be 0, and the larger R is, the more blurry the image is.

Therefore, the image quality can be determined based on the R value.

The enlarging process in step S3 in FIG. 1 in the above-described embodiment can be performed, for example, by arithmetic processing as shown in FIG. 3. That is, a 3×3 matrix pixel array a-i is extracted from the original image B, and its maximum pixel value is replaced with the pixel value at the center e' of the new matrix a'-i'.

e' = m a x (a, bSc---i
) This process is performed on the entire two-dimensional plane including image B while shifting the position of the matrix pixel by pixel, and by repeating this process two to three times, an enlarged image is obtained.

In the above example, the average density of the pattern portion CB) was determined, but the threshold level when binarizing the binarized image B may be used as the average density of the pattern portion.

Further, if the input image A has a background, the image after removing the background may be regarded as the input image A in FIG. 1 and the processes of steps 81 to S7 may be performed.

Further, in the above embodiment, the pattern part (B) and the peripheral part (D
) is used as an index for image quality evaluation, but this has the effect of removing the fluctuation factor when the level of the input image A is affected by illumination or the like.

Therefore, when the influence of illumination and the like is small, it is possible to use the average density MD- itself of the peripheral area (edge area) as an index for image quality evaluation.

FIG. 4 shows an example of the field of application of the image quality inspection method of the present invention. FIG. 4 is a template for markings formed on plate glass by sandblasting. This stamp is made by placing a template with letters and symbols on the plate glass and then blowing sand onto it. Conventionally, it was visually judged whether the stamp was formed correctly without blurring or chipping, but by using the image quality evaluation method of the present invention, the image of the stamp can be processed according to the steps shown in Figure 1. By this,
Automatic determination becomes possible.

〔Effect of the invention〕

As described above, the invention of claim 1 evaluates the image quality using the average pixel value of the edge portion of the input image, and since the average pixel value indicates the degree of blur of the edge portion, the input image Quantitative image quality evaluation can be performed while reducing the influence of signal level fluctuations.

According to the second aspect of the invention, a reduced image is obtained by binarizing the input image and edge portions are extracted, so that the processing procedure is simple and high-speed processing is possible.

According to the third aspect of the invention, since the image quality is evaluated based on the ratio of the average pixel value of the input image corresponding to the reduced image portion to the average pixel value of the edge portion, it is possible to completely remove the level fluctuation factors of the input image. The image quality of the target image can be evaluated without being affected by the illumination during image capture.

In the invention of claim 4, the reduced image by binarization is enlarged, and the edge portion of the input image is extracted based on the difference between the enlarged image and the reduced image, so the processing is suitable for digital images, and the processing The procedure is simple.

[Brief explanation of the drawing]

FIG. 1 is a flowchart of an image processing procedure showing an embodiment of the image quality evaluation method of the present invention, FIG. 2 is a diagram showing the two-dimensional shape of the input image and a cross section in the pixel value direction, and FIG. FIG. 4, which is a diagram showing enlargement processing, is a diagram showing an example of an image evaluated by the present invention. In addition, in the symbols used in the drawings, A --- 1 --- 1 --- Input image B --- .・・−
Binarized image C-・-・・−・・・・・−・−・・
−・・Enlarged image D−・−・−・−・−・・−・−−
111 --- This is the peripheral part (edge part).

Claims (1)

[Claims] 1. An image quality evaluation characterized in that an edge portion of an input image is extracted, an average pixel value of the edge portion is determined, and this average pixel value is used as an index of the image quality of the input image. Method. 2. Binarize the input image to obtain a reduced image on the two-dimensional surface of the image, extract the difference between the input image and the reduced image as the edge portion, and extract the input corresponding to this edge portion. 2. The image quality evaluation method according to claim 1, wherein the average pixel value is obtained by dividing the sum of pixel values by the area of an edge portion. 3. The image quality evaluation method according to claim 2, wherein a ratio between an average pixel value of the input image corresponding to the reduced image portion and an average pixel value of the edge portion is used as an index for the image quality evaluation. 4. A claim characterized in that an enlarged image is formed by enlarging the reduced image obtained by the binarization on a two-dimensional image, and a difference between the enlarged image and the reduced image is used as the edge portion. Item 3: Image quality evaluation method.