JPH064705A - Picture processor - Google Patents

Abstract

(57) [Summary] [Purpose] A character frame can be detected without limiting the print color of the input form, and even if there is a line segment break in the character frame printed in the halftone print color, the broken line segment (EN) An image processing device that can be connected to accurately recognize a character frame. A quantization circuit converts image data read by a reading scanning unit 41 into a multi-valued quantization pattern divided into an integer × integer number of cells, and the connectivity of the quantization pattern in the x direction or the y direction. The density processing unit 45 determines the density value of each pixel by using the density value table 44 that emphasizes.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus capable of accurately reading characters written on a form having a character frame.

[0002]

2. Description of the Related Art In a conventional image processing apparatus for optically reading an image, a character frame is printed on a form in a dropout color that is not detected by the image processing apparatus in order to accurately cut out characters on the form one by one. Was. That is,
A character frame of a dropout color that is not detected by the image processing apparatus is printed in advance on the form, and when the form is filled, characters are written with reference to the character frame printed with the dropout color. Since the character frame of the dropout color that is not detected by the image processing apparatus is also visible to the person, the writer can write characters with reference to the character frame printed in the dropout color.
It is preferable that the character frame of the dropout color that can be handled in such an image processing apparatus be rich in the color types because restrictions on printing and management of the form are reduced. Therefore, in order to increase the types of dropout colors that the image processing device can handle,
For example, an image processing device as shown in FIG. 8 has been proposed.

That is, FIG. 8 is a block diagram showing a conceptual configuration of a first conventional image processing apparatus, for example, the image processing apparatus described in Japanese Patent Laid-Open No. 3-214381. Further, FIG. 9 is a diagram for explaining the character reading principle of the image processing apparatus shown in FIG.

In FIG. 8, the first conventional image processing apparatus includes a reading unit 11 for reading characters and numbers written on a plurality of input forms of different printing colors and converting them into an electric signal. A light source selection unit 12 that selects one of a plurality of light sources, a print determination unit 13 that determines the presence or absence of data of a print pattern portion from the data of the input form read by the reading unit 11, and the need for reading For the non-printing pattern, the input form 14 is printed in a printing color of one of the color systems of a plurality of light sources and a plurality of light sources, for example, a red light source 16 and a blue light source 17. It is composed of a light source unit 15.

The operation of the first conventional image processing apparatus configured as described above will be described with reference to FIG. 9 showing the character reading principle of the image processing apparatus.

In FIG. 9, the print pattern 21 of the character frame showing the character reading area of the input form 14 is the input form 1
4 is printed in the same color system as one of the light sources of the reading unit 11. The reading result 22 indicates that the character frame portion is not read and the character described in the character reading area is read because the printing color of the character frame and the color system of the light source are the same. The reading result 23 indicates that, as a result of reading the input form 14 using a light source having a different color and printing system of the character frame, the character frame portion is also read together with the characters written on the input form 14.

The above-mentioned reading process will be described together with the specific operation of the first image processing apparatus shown in FIG. That is, as shown in FIG. 8, the first conventional image processing apparatus includes a light source unit 15 including a plurality of light sources such as a red light source 16 and a blue light source 17, and the print determination unit 13 includes a reading unit 11. The character frame data is determined based on the data converted into the electric signal, and the light source selection unit 12 controls to switch the plurality of light sources.

FIG. 10 is a diagram showing scanning and photoelectric conversion output signals for a character and a character frame in the reading process by the first image processing apparatus. When the scanning line 30 scans the red frame 31 and the character 32 as shown in FIG.
As shown in (b), the blue light source 17 outputs an output signal (1) including a character and a character frame, and the red light source 16 outputs a character output signal (2). By calculating the difference between these signals, the output signal (3) corresponding to the character frame can be obtained, and the position of the character frame can be accurately detected. Therefore, it is understood that the first conventional image processing apparatus can highly accurately cut out a character from the output signal (2) of the character portion.

However, in the image processing apparatus shown in FIG. 8, the printing color of the character frame is limited to the dropout color unique to the image processing apparatus as described above (for example, the red light source 16 and the blue light source 17). There is a problem that the degree of freedom in designing the input form becomes small, and in order to increase the number of types of dropout colors that the image processing device can handle, the print color of the input form must be increased. Since it is necessary to provide a light source of a color corresponding to, there is a problem that the configuration is complicated and the hardware is increased. Therefore, as another means for detecting the character frame, for example, a method of performing contour tracing of an image as shown in FIG. 12 has been proposed.

That is, FIG. 11 is a view for explaining a reading process of a second conventional image processing apparatus, for example, the image processing apparatus disclosed in Japanese Patent Laid-Open No. 59-119484.
It is a figure which shows the form image containing a character and a character frame, and the character image which extracted the character frame. FIG. 12 is a diagram showing the contours of the images of the characters and the character frames to be read and the contour tracing results of only the character frame portion in the reading process by the second conventional image processing apparatus.

In FIG. 11A, the character frame 34 and the character 33 are
Duplicate images of are shown. The character frame 34 forms a straight band parallel to the four sides of the input form, and the approximate position can be known in advance even when the cutting accuracy, expansion and contraction of the form are taken into consideration. Here, the problem is the character frame 34.
By tracing and detecting the edge of, the overlapping portion 35 with the character 33 is extracted, and as shown in FIG. 11B, a character image in which the character portion and the character frame overlapping portion 35 are complemented is obtained.

In the image contour tracing method of the second image processing apparatus, the contours of the binarized image are sequentially detected at a predetermined pitch step by step in the linear direction of the character frame. At that time, the displacement of the contour position in the direction perpendicular to the linear direction of the character frame is stored for each step, and the difference between the displacement newly obtained after the next step and the previously stored known displacement is stored. Was calculated, and the contour position was corrected based on this result by comparing with a predetermined threshold value so that the position of the character frame can be accurately determined. That is, the result of tracing the contour of the binarized image including the character and the character frame is the contour 36 of the image of the character and the character frame shown in FIG. 12, and the contour tracing result 37 of only the character frame portion is detected. It is intended to obtain the overlapping portion 38 of the character frame and the character.

[0013]

Since the second conventional image processing apparatus has the above-described configuration, the dropout color character frame of the input form is not detected in the character frame detection by the image contour tracking method. The print color must be sufficiently detectable in consideration of the characteristics peculiar to the image processing device, and if the line break of the character frame occurs, the character frame cannot be detected or the character frame is detected. There was a problem such as making a mistake in itself.

The present invention has been made to solve the above problems, and detects a character frame without limiting the printing color of an input form and a character frame printed in a halftone printing color. It is an object of the present invention to provide an image processing apparatus capable of accurately cutting out a character frame by connecting broken line segments even if the line segment is broken.

[0015]

In order to achieve the above object, an image processing apparatus according to the present invention irradiates a document surface on which an image is drawn with light along a predetermined scanning line so that the surface of the document is scanned. An image processing device that forms an image of reflected light on a photosensitive element of an image reading unit to optically read an image, and converts an image data into a multi-valued quantization pattern divided into integer × integer cells. A combination circuit, a combination table that emphasizes the connectivity in the x direction or the y direction, and a density processing unit that determines the density value of each pixel using the connection table, and accurately connects a character frame with a line segment breakage. It is something that is done.

[0016]

Therefore, according to the image processing apparatus of the present invention, the quantization circuit converts the image data read by the reading scanning unit into a multi-valued quantization pattern divided into an integer × integer number of cells, and the quantum Since the density processing unit determines the density value of each pixel by using the combination table that emphasizes the connectivity in the x-direction or the y-direction of the digitized pattern, there is a break in the line segment of the character frame printed in the halftone print color. Even if the broken line segments are connected, the character frame can be accurately recognized.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of functional blocks of the image processing apparatus according to this embodiment.

In FIG. 1, the image processing apparatus according to the present embodiment irradiates the surface of a document on which an image is drawn with light along a predetermined scanning line, and forms light reflected from the surface of the document on a photosensitive element. A scanning unit 41 for optically reading an image and converting an optical signal into an electric signal, an A / D converter 42 for quantizing the read electric signal, and a multi-valued buffer for storing the quantized electric signal 43, a density value table 44 that emphasizes connectivity in the x direction or the y direction, and a density pattern processing unit 45 that determines the density value of each pixel using the density value table 44.
Based on the binarization processing unit 46 that binarizes the density pattern determined by the density pattern processing unit 45, the format data and the character frame data 47 stored in advance in the memory, and the format data and the character frame data 47. For example, a character frame detection unit 48 for detecting the position of the character frame printed on the document surface by a histogram in the x direction or the y direction or by contour tracing, and a character portion cut out from the character frame position with high accuracy. Based on the character cutout unit 49, the recognition dictionary 50 stored in advance in the memory, and the standard pattern stored in the recognition dictionary 50, for example, recognition processing by a matching method is performed, and the result is used as the read result of the input image. It is composed of a recognition unit 51 for outputting.

Next, the operation of the image processing apparatus of the present embodiment configured as described above will be described with reference to FIGS. 2 is a diagram showing an example of the quantization pattern of the image processing apparatus according to the present embodiment, FIG. 3 is a diagram showing an example of the density value table, and FIG. 4 is a diagram showing the reading pattern. It is a figure which shows one Example, and FIG. 5 is a figure which shows one Example of a read image.

In FIG. 1, the reading scanning section 41 of the image processing apparatus of the present embodiment irradiates the document surface on which the image of the input form is drawn with light along a predetermined scanning line, and reflects light from the document surface. Image on the photosensitive element to read the image optically,
Converts optical signals into electrical signals. The electric signal read by the reading scanning unit 41 is quantized by the A / D conversion unit 42,
The quantized electric signal is stored in the multilevel buffer 43. The multi-valued buffer 43 that stores the quantized electric signal can be considered as an integer × integer number of cells corresponding to each pixel of the read image. FIG. 2 shows an embodiment of such a multi-valued buffer 43 which is matrixed.

As a concrete example of the quantization pattern, FIG. 2 shows a pixel using a 3 × 3 pattern, and the following description will be given using a 3 × 3 pattern. In FIG. 2, P, P0, P1, P2 and P3 indicate the densities corresponding to the respective pixels, and the densities are 0 to 7 (0: white to 7:
The value of (black) is taken.

Now, it is assumed that the value of "P" is obtained by the equation (1) and that value is used as the density at the point P thereafter. P = f (P0 + P2, P1 + P3) (1) where f (x, y) is a function with two arguments, and the first argument is the pixel on the left and right of the target pixel. Is the sum of the concentrations,
The second argument is the sum of the densities of the pixels above and below the target pixel and is "0".
It takes a value of ~ 14 ".

For example, the function "f (x, y)" is defined as a density value table as shown in FIG. 3 which emphasizes connectivity in the x direction or the y direction. Further, an image as shown in FIG. 5 is considered as a read image of the image read by the image processing apparatus of this embodiment. In this case, assuming that the read quantization pattern of the image processing apparatus of this embodiment is as shown in FIG. 4A (the read values are P0 = 3 and P1 =
0, P2 = 4, P3 = 0), and "P" is as shown in equation (2). P = f (3 + 4,0 + 0) = f (7,0) ... (2) This value becomes P = f (7,0) = 4 from the density value table of FIG. The density pattern is as shown in b).

Here, the quantization method of this embodiment is compared with the conventional simple smoothing method for averaging the density of each pixel.
As a “P” calculation formula based on the simple smoothing method, for example, formula (3) is defined. P = (P0 + P1 + P2 + P3) / 4 (3) In this case, if the read pattern of the image processing apparatus is the quantization pattern shown in FIG. )become that way. P = (3 + 0 + 4 + 0) / 4 = 1 ... (4) Here, for example, when the threshold value taken by the binarization processing unit 46 is “3”, the pixel color by the simple smoothing method is “white”. However, the color is “black” in the pixel by the quantization method of the present embodiment. Therefore, in the line segment of FIG. 5 showing the character frame reading image of the image processing apparatus, an image in which the line segment of the character frame 53 is broken as shown in FIG. 5A is the character frame 53 of FIG. 5B. As shown in, it can be seen that, as a result, the left and right line segment breaks are recognized in the connected image.

In the image processing apparatus according to the above-described embodiment, the density value table 44 has been described as a density value table for emphasizing the connectivity in the x direction or the y direction. Fragmentation is likely to occur, but
When it is difficult to break the line segment in the y direction, or when it is necessary to emphasize only the connectivity in the x direction in the post-processing,
For example, as shown in FIG. 6, the density value is determined by using a density value table that emphasizes only the connectivity in the x direction, the subsequent processing is performed, and the line segment breaks in the character frame are combined. The effect of can be expected.

In the image processing apparatus of the above embodiment, the density value table 44 has been described as a density value table for emphasizing the connectivity in the x direction or the y direction.
The density value may be determined using a density value table that emphasizes only the connectivity in the directions.

In the image processing apparatus of the above embodiment, the function f (x, y) has been described as a function using only pixels in the x direction and pixels in the y direction. P4, P5, P6, P in FIG. 7 showing one embodiment of the second quantization pattern of the image processing apparatus according to the example
It may be a function that considers 7).

Further, in the image processing apparatus of the above embodiment,
Although the density value determination processing using a 3 × 3 quantization pattern has been described, the matrix may be an integer × an integer number of squares with a larger matrix.

[0029]

As described above, according to the image processing apparatus of the present invention, the quantization circuit divides the image data read by the reading / scanning unit into integer × integer cells and the multivalued quantization pattern. The density processing unit is configured to determine the density value of each pixel by using the density value table that emphasizes the connectivity of the quantization pattern in the x direction or the y direction. Even if a line segment of the printed character frame is broken, it is possible to connect the broken line segments and accurately recognize the character frame. Therefore, it is possible to eliminate the need for printing using a special dropout color when creating an input form and reduce the cost for creating a form, as well as the required quality of the character frame. Can be alleviated to a certain extent, so that there is an effect that it is possible to increase the types of forms that can be handled by the image processing apparatus.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of functional blocks of an image processing apparatus according to an embodiment.

FIG. 2 is a diagram showing an example of a first quantization pattern of the image processing apparatus according to the present embodiment.

FIG. 3 is a diagram showing an example of a first density value table of the image processing apparatus according to the present embodiment.

FIG. 4 is a diagram showing a quantization pattern of read image data in the image processing apparatus according to the present embodiment.

FIG. 5 is a diagram showing a read image of image data in the image processing apparatus according to the present embodiment.

FIG. 6 is a diagram showing an example of a second density value table of the image processing apparatus according to the present embodiment.

FIG. 7 is a diagram showing an example of a second quantization pattern of the image processing apparatus according to the present embodiment.

FIG. 8 is a block diagram showing a conceptual configuration of functional blocks of a conventional first image processing apparatus.

FIG. 9 is a diagram for explaining a character reading principle of a conventional first image processing apparatus.

FIG. 10 is a diagram showing scanning and photoelectric conversion output signals for a character and a character frame in a reading process by the first image processing apparatus.

FIG. 11 is a diagram illustrating a form image including characters and a character frame and a character image in which the character frame is extracted, for explaining the reading process by the second conventional image processing apparatus.

FIG. 12 is a diagram showing contours of an image of a character and a character frame to be read and a contour tracking result of only a character frame portion in a reading process by a second conventional image processing apparatus.

[Explanation of symbols]

 41 Reading / scanning unit 42 A / D conversion unit 43 Multi-valued buffer 44 Density value table 45 Density processing unit 46 Binarization processing unit 47 Format data and character frame data 48 Character frame detection unit 49 Character cutout unit 50 Recognition dictionary 51 Recognition unit

Claims (1)

[Claims] 1. An image is optically read by irradiating the surface of the original on which an image is drawn with light along a predetermined scanning line, forming the reflected light from the original surface on a photosensitive element of the image reading means, and reading the image optically. In an image processing device, a quantization circuit for converting image data into a multi-valued quantization pattern divided into integer × integer boxes, a connection table for emphasizing connectivity in the x direction or the y direction, and a connection table are used. An image processing apparatus comprising a density processing unit that determines a density value of each pixel by connecting the character frames with broken line segments.