JPH06282646A - Picture processing method - Google Patents

Abstract

PURPOSE:To accurately and satisfactorily extract the contour of an object in an original picture when original picture data is converted into binary data by dividing the original picture into plural local areas and emphasizing the contrast for the original picture data of each pixel of each local area. CONSTITUTION:When the original picture data is converted into binary data, the original picture data are divided into plural local areas 1a, 1b.... The data processing emphasizing the contrast of the original picture data in each local area is performed. Then, the negative picture data are prepared for each pixel of the original picture data. Then, the blurring negative picture data consisting of the average negative picture data of pixel in the prescribed area including the pixel for each pixel are prepared. The multi-value contour picture data are obtained by adding the original picture data and the blurring negative picture data on each pixel with the prescribed factor multiplied. The linear density conversion is performed on the data in the range of the prescribed gradation value. As each local area is reduced, the degree of extracting the contour section becomes high.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used when an original image having light and shade such as a photograph is captured as binary digital information and then various image processing is performed, and particularly, original image data. When converting to binary data,
The present invention relates to an image processing method suitable for favorably extracting the contour of an object in the original image.

[0002]

2. Description of the Related Art Generally, since the color density is multi-valued information formed by a plurality of gradation values, when this is taken in as digital information, the amount of memory is reduced,
In order to facilitate the subsequent data processing, it is performed to classify it into one of black and white, convert it into binary information, and take it in.

When converting multi-valued information into binary information in this way, it is necessary to incorporate the features of the original image required for image processing into the binary information.

In particular, it is necessary to clearly divide the contour of the object in the original image into binary information. This is because this contour line is used for various purposes in image processing. For example, (1) it is used as a cutout line of an object in an image, (2) it is used as an outer frame when transforming the shape of the cutout image, and (3) brightness, contrast, and color of the cutout image. It is used as an outer frame when adjusting image quality such as, and (4) used as a feature amount when recording an image.

Therefore, in the past, in order to extract the contour of an object in the original image, the original image data consisting of the density value held by each pixel of the original image and the negative image data for this original image data are mutually exchanged. The pixel positions of are shifted in a certain direction and overlapped, the gradation values of the original image and the negative image are added, and multivalued image data to be converted into binary data is created. The gray scale value of each pixel and the predetermined reference gray scale value are compared to classify them into binary black and white.

[0006]

However, in the above-mentioned conventional method, the contour of the object in the original image could not be accurately extracted.

This is because the contour of the object generally appears as a portion where the gradation value of the density between adjacent pixels is large, but the original image data and the negative image data are shifted and overlapped as in the conventional method. This is because the change in the gradation value of the contour portion is canceled out, and the contour portion may not be extracted. Further, conventionally, the degree of extraction of the contour portion is low because the processing of creating multi-valued image data at once is performed on the entire area of the original image.

The present invention has been made in view of these points, and when converting original image data into binary data,
An object of the present invention is to provide an image processing method capable of accurately and satisfactorily extracting the contour of an object in the original image.

[0009]

In order to achieve the above object, the image processing method of the present invention provides original image data in which each of a plurality of pixels is composed of a plurality of gradation values, Data is either black and white binary 2
When converting into the value data, the original image data is divided into a plurality of local regions, data processing for increasing the contrast of the original image data in each local region is performed, and then each pixel of the original image data is negatively processed. Create image data,
After that, blur negative image data composed of average negative image data of pixels in a predetermined area including the pixel is created for each pixel, and the original image data and the blur negative image data are multiplied by predetermined coefficients for each pixel and added. Multi-valued contour image data is obtained, and the multi-valued contour image data is linearly density-converted within a range of a predetermined gradation value to be stored, whereby the multi-valued contour image is provided for conversion into the binary data. It is characterized by creating data.

[0010]

According to the present invention, the original image is divided into a plurality of local regions, the contrast is emphasized with respect to the original image data of each pixel in each local region, and then the blurred image data of the pixel is superimposed. Furthermore, since the multi-level contour image data used for conversion into binary data is created by performing linear density conversion of the data within a range of a predetermined gradation value, the object in the original image It is possible to obtain image data in which the contour is accurately and favorably extracted.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT An embodiment of the present invention will be described below with reference to FIGS.

1 to 5 show an embodiment of the present invention.

In this embodiment, the apparatus shown in FIG.
The image data is formed according to the flowchart shown in FIG.

In the present embodiment, as shown in FIG. 1, a case will be described in which an original image 1 having a shade of color such as a color photograph is captured as binary image data.

An apparatus for creating image data of the original image 1 is a computer 2 having a built-in CPU and various memories (both not shown) and the like, and a keyboard and a mouse for inputting various instructions to the computer 2. And the like, an output device 4 such as a CRT, and a density reader 5 for reading the gradation value of the color density of each pixel of the original image 1.

Next, the method of the present invention will be described.

First, the gradation value of the color density of each pixel of the original image 1 is read using the density reader 5 and stored in the memory in the computer 2 as the original image data. The reading of the gradation value may be performed by a known method.

For example, first, the resolution to be read is determined. That is, the pixel size as the resolution and the maximum value of the color density, that is, the maximum gradation value (Cmax) that the image data can take.
To decide. Then, the original image 1 is scanned in the X and Y directions according to the pixel size, and the gradation value I (n, m) of the color density of each pixel (n, m) is measured,
The whole is once stored in the memory in the computer 2 as original image data. Here, n is the X coordinate of each pixel, and m is the Y coordinate of each pixel.

Next, multivalued contour image data used for conversion into binary data is created. This method will be described with reference to the flowchart shown in FIG.

First, in step ST11, as shown in FIG. 3, the original image 1 is converted into a plurality of local regions 1a, 1b,
1c ... The size of each local region is specified by the input device 3 according to the set resolution and the like, and the smaller the size, the higher the degree of extraction of the contour portion.

Next, in step ST12, it is determined whether or not the contour extraction processing has been completed for all local regions. If this determination is YES, the process is terminated, and if NO, the process proceeds to step ST13.

In step ST13, data processing for increasing the contrast of the original image data is performed as a preprocessing for contour extraction.

That is, the original image data I of the original image 1
When the change amount of (n, m) is small and the contrast is relatively weak, the original image data I (n, m) is subjected to a preprocessing for enhancing the contrast, and the original image data IB after the preprocessing is performed. (N, m), that is, IB (n, m) = T {I (n, m)} = T (i).

In this case, the gradation frequency distribution of the color density of the original image 1 is hp (i), and the target gradation frequency distribution is hq (i) = (total number of pixels of original image 1 / Cmax). And T (i) is Σhp (j) in the range of 0 ≦ j ≦ i.
Is obtained as a gradation value that minimizes the absolute value of the difference between Σhq (j) in the range of 0 ≦ j ≦ T (i).

Next, in step ST14, contour extraction processing is performed.

This contour extraction processing is performed according to the flowchart shown in FIG.

First, in steps ST21 and ST22, the method of creating multivalued contour image data shown in the following steps ST23 to ST27 is applied to all pixels.

Then, in step ST23, the negative image data N for each pixel (n, m) of the original image 1
(N, m), that is, N (n, m) = Cmax-I (n, m) is created.

Next, in step ST24, blurred negative image data U consisting of average negative image data of pixels in a predetermined area including the pixel for each pixel (n, m).
Create (n, m).

This will be described with reference to FIG. 5. The size of the blur matrix B, which is a predetermined region including the pixel to which each pixel (n, m) is blurred (the blur matrix size in the X and Y directions is usx. , Usy)
Is determined, and the average negative image data of all pixels in the blur matrix B is calculated to obtain the blurred negative image data U (n,
m), that is, U (n, m) = ΣN (n−i, m−j) / (usx × u
sy) is created.

Here, by setting the values of usx, usy, i, j according to the contents of the image processing, each pixel (n, m) is set to the center of the blur matrix B as shown in FIG. Or you should bias it. Further, when obtaining the blurred image data of the edge of the local area, it is preferable to use the image data of the adjacent local areas.

Next, in step ST25, a predetermined coefficient (WI, WI) is added to the original image data I (n, m) and the blurred negative image data U (n, m) for each pixel (n, m). WN) is multiplied and added to obtain multilevel contour image data L (n, m), that is, L (n, m) = WI · I (n, m) + WN · U (n,
m) is calculated.

Each coefficient (WI, WN) may be freely set within the range of 0 to 1 according to the contents of image processing.

Next, in steps ST26 and ST27, the multivalued contour image data L (n, m) is stored within a predetermined gradation value range, that is, within the maximum gradation value (Cmax) that the image data can have. By doing so, the multi-valued contour image data L2 (n, m) to be used for conversion into binary data is further described. In step ST26, the multi-valued contour image data L (n, m) is converted into image data. Of the maximum gradation value (Cmax) that can be taken is determined, and if NO is not exceeded, L2 (n, m) = L (n, m) is set and the process returns to step ST12.

When the above determination is YES, L2 (n, m) = Cmax is set and the process returns to step ST22.

Thus, according to this embodiment, the original image 1
The original image data I (n, m) of each pixel (n, m) is superposed with the blurred negative image data U (n, m) of the pixel to be used for conversion into binary data. Since the multi-valued contour image data L2 (n, m) is created, the data is not shifted and overlapped as in the conventional case. Therefore, the contour of the object in the original image 1 is accurately and favorably extracted. Image data can be obtained.

Next, in step ST15, the multivalued contour image data L2 (n, m) is further post-processed.

That is, the multivalued contour image data L2 (n,
m) the gradation value within a predetermined gradation value, for example, from the minimum gradation value that the image data can take to the maximum gradation value (Cmax)
Up to the range up to the post-processing multi-valued contour image data LA (n, which is used for linear density conversion and conversion into binary data
m), that is, LA (n, m) = {(Cmax-Cmin) / (L2m
ax-L2min)} * {L2 (n, m) -L2mi
n)} is calculated.

In this post-processing, if the obtained range of L2 (n, m) (L2max-L2min) is narrower than the range of image data (Cmax-Cmin), the range of (L2max-L2min) is set. (Cmax-Cm
in), so that it is the same as the range of (in).

As a result, the range (C
It is possible to obtain the multivalued contour image data LA (n, m) using the entire (max-Cmin).

For the multi-valued contour image data LA (n, m) thus obtained, a threshold value (Cx) for converting into binary of black and white is obtained according to the contents of the image processing. It is set within the range (Cmax-Cmin) to obtain, and the pixels whose multivalued contour image data LA (n, m) is larger than the threshold value (Cx) are white, and the pixels smaller than the threshold value (Cx) are black. Then, binary image data is created as a whole.

After that, the process returns to step ST12, and the processes of steps ST13 to ST15 are continuously performed until the determination becomes YES, that is, until the contour extraction process for all local regions is completed.

Next, the method of the present invention will be described with reference to specific examples shown in FIGS.

In this embodiment, binary image data is extracted from the original image 1 consisting of the color photograph shown in FIG. As the resolution, the original image 1 is decomposed into the number of pixels of X direction × Y direction = 690 × 520, and the size of each local region is set to the pixels of X direction × Y direction = 25 × 25 to set the color density. The gradation value is set to 0 to 255.

The original image data I (n, m) of the Y-coordinate = 68 portion of the original image 1 is as shown in FIG. 7, and is the multivalued contour image data LA used for conversion into binary data.
(N, m) is as shown in FIG. The multilevel contour image data by the conventional method for comparison is as shown in FIG.

Then, when the multivalued contour image data LA (n, m) as shown in FIG. 8 is processed by the known dither method and output by the printer, it becomes as shown in FIG. 10 and as shown in FIG. FIG. 11 shows the multi-valued contour image data processed by the conventional method by the dither method and output by the printer as described above.

As is apparent from FIGS. 7 to 11, according to the present embodiment, the contour portion of the object in the original image 1 is clearly emphasized as compared with the conventional example, and the contour portion has a binary value. It can be seen that the image data can be extracted accurately and surely.

The present invention is not limited to the above-described embodiment, but can be applied to all the monochrome images and other images having gradation values, and can be changed as necessary. be able to.

[0049]

As described above, the image processing method of the present invention divides the original image into a plurality of local regions, emphasizes the contrast of the original image data of each pixel in each local region, and then Since the blurred image data is overlaid and the data is linearly density-converted within a range of a predetermined gradation value, multivalued contour image data used for conversion into binary data is created. When the original image data is converted into binary data, the contour of the object in the original image can be extracted accurately and favorably.

[Brief description of drawings]

FIG. 1 is a schematic view of an apparatus for carrying out an image processing method of the present invention.

FIG. 2 is a flowchart showing an embodiment of the method of the present invention.

FIG. 3 is an enlarged explanatory diagram of an original image.

FIG. 4 is a flowchart showing an embodiment of contour extraction processing of the method of the present invention.

5 is an enlarged view of part A of FIG.

FIG. 6 is an original image diagram to which the method of the present invention is applied.

FIG. 7 is a color density gradation graph showing original image data.

8 is a graph showing multivalued contour image data LA (n, m) used for conversion from the original image data of FIG. 6 into binary data obtained based on the method of the present invention.

FIG. 9 is a graph showing multivalued contour image data by a conventional method for comparison.

FIG. 10 is an image diagram in which multivalued contour image data obtained by applying the method of the present invention to the original image of FIG. 6 is processed by a dither method and output by a printer.

FIG. 11 is an image diagram of an original image in which multivalued contour image data obtained by applying a conventional method is processed by a dither method and output by a printer.

[Procedure amendment]

[Submission date] February 22, 1994

[Procedure Amendment 1]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 6

[Correction method] Change

[Correction content]

[Figure 6]

Claims (1)

[Claims] 1. When converting original image data in which each of a plurality of pixels is composed of a plurality of gradation values into binary data in which the data of each pixel is one of binary black and white , The original image data is divided into a plurality of local regions, subjected to data processing to increase the contrast of the original image data in each local region, then create negative image data for each pixel of the original image data, For each pixel, blur image data consisting of average negative image data of pixels in a predetermined area including the pixel is created, and for each pixel, the original image data and the blur negative image data are respectively multiplied by a predetermined coefficient and added. The multi-valued contour image data is obtained, and the multi-valued contour image data is converted into the binary data by linear density conversion within the range of a predetermined gradation value. An image processing method, characterized in that multi-valued contour image data to be provided is created.