JPH11243488A - Image reading device - Google Patents

Abstract

(57) [Problem] To provide an image reading apparatus which prevents so-called black crush caused by continuation of black pixels as much as possible. SOLUTION: When a predetermined number of black pixels continue in a binary image signal binarized by a character processing unit 6,
The black pixel correction unit 9 determines a predetermined number of white pixels in the pixel direction ahead of the last black pixel of the continuous black pixel group according to the white level reference value created by the white level reference value creation unit 7. Change to

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading apparatus which reads an image on a document by an image sensor, binarizes a multi-valued image signal output from the image sensor into white or black, and outputs the resulting signal. More specifically, the present invention relates to an image reading apparatus that prevents as much as possible so-called black crush due to continuation of black pixels.

[0002]

2. Description of the Related Art Heretofore, as this type of image reading apparatus, there has been known a flat bed type in which a book type original is also read.

FIG. 11 schematically shows a conventional flat-bed type image reading apparatus, and FIG. 12 shows an image sensor unit using a contact image sensor as an example.

In FIG. 12, in this conventional image reading apparatus, a light source D irradiates a document A, and a reflected light corresponding to the density of the document A is focused on a light receiving portion F of the image sensor 1 by a lens E. The image information is obtained by converting the electric signal into an electric signal by a photoelectric conversion element built in 1. The image information is generally called an image signal, and is represented as a numerical value such as 6 bits or 8 bits via an amplifier 2 and an A / D converter 3 as shown in FIG. At this time, if it is 8 bits, the range from black to white can be expressed in 256 tones from 0 to 255.

[0005] The image signal thus quantified is subjected to a shading correction section 4 for electrically correcting various non-uniformities such as uneven light amount of the light source D for irradiating the original A and uneven sensitivity of the image sensor 1. Be uniformed.

[0006] The image signal thus obtained is almost always used when the image is reproduced by a recording device such as a printer, and each pixel of the recording device is turned on or off (white or black). Since only two gradations are represented, it is inconvenient for the above-mentioned 8-bit image signal, for example.

For this reason, the 8-bit multi-valued image signal as described above is binary-processed by an image signal processing circuit for converting to white or black, and then output as a binary image signal from the image reading device.

The image signal processing circuit includes a character processing section 6 for reading an original A composed of only white and black characters and the like, and a photographic processing for reading an original containing halftones such as gray. It comprises a unit 5.

Here, in the conventional character processing unit 6, specifically, as shown in FIGS. 14 and 15, a threshold value is set to 50% of the maximum value of the multi-level image signal level represented by the multi-level. And
A simple binarization process for determining each pixel such that the input image signal level is equal to or higher than the above threshold value is white and the input image signal level is lower than black is performed.

For this reason, in the conventional image reading apparatus, for example, when a book type original A shown in FIG. 11 is to be read, a certain area (near a seam of the book) of the original A is
When the document A is irradiated from the light source D in order to move away from the original predetermined reading position (the contact surface of the stage glass), the amount of light reaching the document A decreases, and the reflected light reaching the image sensor 1 decreases. The image signal level input to the character processing unit 6 is low, and in the worst case, even when the input image signal level is less than the threshold value and an image such as a character exists at the reading position, the All the output binary image signals have a black level.

Further, even if the input image signal level across the threshold value is input to the character processing unit 6, since the original A is away from the predetermined position, the image forming position of the lens E is shifted and the focus is not obtained. The image reading device outputs a binary image signal obtained by processing a blurred image signal whose white is undetermined into white or black.

[0012]

As described above, in the conventional image reading apparatus, only when the document is at a predetermined reading position, an image signal corresponding to the density of an image in the document and a thin line having a fixed focus are used. An image signal can be obtained.

Conventionally, however, there are two types of image reading apparatuses of this type: a sheet-through type for sheet-type documents, and a flat-bed type for book-type documents. In the case of the sheet-through type for the former sheet type document, by accurately controlling the reading position of the document by a document conveying mechanism such as a guide roller, an image signal corresponding to the density of an image in the document can be obtained. There is no particular problem because an image signal such as a thin line with a fixed focus can be obtained. Is not always at the scanning position, so that an image signal corresponding to the density of the image in the original and an image signal such as a thin line with a fixed focus can be obtained accurately. It can not be.

That is, when an image such as a character is present in a portion where a floating occurs, the distance from the light source is farther than that of a document at a predetermined reading position.
The image signal level obtained in a white base portion that occupies a large part of the original document is reduced, and thin lines constituting characters and the like are blurred and thicker due to the shift of the focal position.

When such an image signal is input to a conventional character processing unit and subjected to simple binarization processing of white or black,
For example, as shown in FIG. 15, when the input image signal level from the white area, which is the background of the document, becomes less than 50% of the input image signal level obtained from the background of the document at the predetermined reading position, all are determined to be black. Will be processed.

In this method, a threshold value is set at 50% of the maximum image signal level input by the conventional character processing unit determination method, and a pixel having an image signal level higher than the threshold value is determined as white. This is for determining a pixel having an image signal level lower than black as black. Here, the threshold value is defined as a maximum (100%) of an image signal level when a document is located at a predetermined reading position and complete white is read in that state, and is, for example, 50% of the maximum value. This is a predetermined setting value.

The threshold value remains fixed until reading of at least one document is completed, regardless of which portion of the document is read. That is, in the conventional image reading apparatus, when a document is locally separated from a predetermined position such as a book type, it is difficult to reproduce an image of that portion. Also, even if the white area of the background barely exceeds the threshold value and is determined to be a white pixel, a thin line constituting a character or the like will be processed as a thick line or blur due to blur due to a shift in the focus position,
With an image signal output from a conventional image reading apparatus, an image in a document cannot be accurately reproduced in all areas.

Therefore, the present invention solves the above problem,
It is an object of the present invention to provide an image reading apparatus that prevents so-called black crushing due to a continuation of black pixels as much as possible.

[0019]

According to a first aspect of the present invention, an image on a document is read by an image sensor and a multi-valued image signal output from the image sensor is converted to white or black. In an image reading apparatus that outputs a binary image, if a predetermined number of black pixels of the image signal after the binarization process are continuous, a pixel ahead of the last black pixel of the continuous black pixel group Black pixel correction means for changing the number of white pixels in the direction to white pixels by a preset number of pixels is provided.

According to a second aspect of the present invention, in the first aspect of the present invention, the black pixel correcting means is configured to perform the continuous black pixel group only for a black pixel group exceeding a preset minimum required number of components. It is characterized in that white pixels are changed from the last black pixel in the forward pixel direction by a predetermined number of pixels.

According to a third aspect of the present invention, in the first or second aspect of the present invention, the black pixel correction means sets the number of pixels to be changed to white pixels in accordance with the maximum value of the image signal level detected on the read line. It is characterized by setting.

According to a fourth aspect of the present invention, in the third aspect of the present invention, the black pixel correcting means divides a pixel row in the read line into a plurality of blocks, and outputs an image signal in the divided block. The number of pixels to be changed to white pixels is set according to the maximum value of the level.

According to a fifth aspect of the present invention, in the third aspect of the present invention, the black pixel correction means includes a high level of a maximum value of the image signal level of the read line and a maximum value of the image signal level of the previous line. The number of pixels to be changed to white pixels is set in accordance with the level of the image signal.

[0024]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing a circuit configuration from image reading to binarization processing in an image reading apparatus to which the present invention is applied. Portions having the same functions as the conventional circuit shown in FIG. The same reference numerals are given.

As shown in FIG. 1, the image reading apparatus includes a white level reference value creating section 7 in addition to a character processing section 6 and a photograph processing section 5 following the conventional shading correction section 4. A correction pixel number setting unit 8 is provided after the white level reference value creation unit 7, and a black pixel correction unit 9 is provided after the correction pixel number setting unit 8 and the character processing unit 6. .

Here, when a predetermined number of black pixels continue in the binary image signal binarized by the document processing unit 6, the white level reference value creation unit 7 determines the last of the continuous black pixel groups. A white level reference value serving as a threshold value for setting the number of pixels to be changed from a black pixel at the tail to a white pixel in the forward pixel direction is created.

Further, in accordance with the white level reference value output from the white level reference value generating section 7, the correction pixel number setting section 8 sets a pixel direction ahead of the last black pixel in the continuous black pixel group. Set the number of pixels to be changed to white pixels.

The black pixel correction section 9 changes the number of correction pixels set by the correction pixel number setting section 8 from the last black pixel in the continuous black pixel group to a white pixel in the forward pixel direction.

FIG. 2 is a block diagram showing the overall configuration of the white level reference value creating section 7 shown in FIG.

As shown in FIG. 2, the white level reference value creating section 7 includes a block dividing section 71, a peak detection holding section 7
2. Peak value storage unit 73 for each block, memory control unit 7
4, a shift register 75, and a comparison / selection unit 76.

Here, the block dividing section 71 counts a data clock which is a clock for each pixel, and sets a PK (Phase Clock) for every predetermined number of pixels (blocks), for example, every 16 pixels.
[0] is created and output.

The peak detection and holding section 72 detects and holds the maximum value of the image signal level for each block (white level reference value for each block) in accordance with PK [0] output from the block dividing section 71. I do.

The peak value storage unit 73 for each block stores
Under the control of the memory control unit 74 described later, the writing operation of the maximum value of the image signal level for each block detected by the peak detection holding unit 72 and the maximum value of the image signal level for each block of the previous line already written A value read operation is performed.

The memory control unit 74 counts the PK [0] output from the block division unit 71 and sequentially increments the address, and stores the image signal level of each block in the block-by-block peak value storage unit 73. Control the maximum value write and read operations.

The shift register 75 matches the maximum value of each block of the preceding line output from the block-by-block peak value storage 73 with the timing of the multi-level image signal input on the reading line.

The comparison and selection section 76 compares the maximum value of each block of the previous line output from the block-by-block peak value storage section 73 with the image signal level of the multi-level image signal input on the reading line. The higher level is selected and output as the white level reference value.

Next, a series of operations of the white level reference value creating section 7 will be described.

As shown in FIG. 2, in the white level reference value creating section 7, first, the block dividing section 71 is initialized by a line start indicating the start of one-line reading, and then the data as a clock for each pixel is obtained. The clock is counted, and PK [0] is created for every predetermined number of pixels, for example, every 16 pixels (block), and output to the peak detection holding unit 72. FIG. 3 shows a circuit configuration in the peak detection holding unit 72.

In the peak detection and holding section 72, the input PK [0] is fed back from the first register 72a which holds the maximum value of the image signal level in the block (the image signal level in the block). Maximum) and AN
The signal is input to the D gate 72b.

Further, the comparison / selection section 72c compares the data output from the AND gate 72b with the multi-valued image signal (one pixel), selects the higher-level signal, and selects the higher level first register 72a.
To the next multi-valued image signal (one pixel).

Therefore, after the first peak detection, the first register 72a compares the PK [0] with the multi-valued image signal, so that the multi-valued image signal is selected and held.

By repeating this operation until the next PK [0] is newly input from the block dividing section 71, the maximum value of the image signal level in the block is detected and the first PK [0] is detected.
The data is held in the register 72a.

Here, the comparison / selection section 72c compares the last pixel in the block with the data fed back from the first register 72a (the maximum value of the image signal level up to the previous pixel) to determine the higher level. And the peak value held in the first register at the next data clock, that is, the maximum value of the image signal level in the block, is stored in the second register 72d at the next PK [0] timing, and Is output to the block-by-block peak value storage unit 73 as the white level reference value of. FIG. 4 shows a series of operation timing charts.

Here, the block-specific peak value storage 73
Is composed of a storage element such as a memory,
4 and the write operation of the white level reference value for each block from the peak detection and holding unit 72 in accordance with the control signal such as the address, read, and write, and the white level reference value for each block in the previous line already stored. Perform a read operation.

At this time, the memory control unit 74 counts PK [0] input from the block division unit 71 and sequentially increments the address.

In this block-by-block peak storage section 73, before writing a new block-based white level reference value, a read operation of the already stored block-based white level reference value of the previous line is performed. As for the address at that time, +1 is added to the address obtained by counting PK [0]. This is to prevent the white / black determination of the read pixel from being performed for one block based on the white level reference value for each block and the white level reference value.

The read operation is performed in the portion of PK [0], and is held until the next read data is input to a third register (not shown) in the peak value storage unit 73 in the vicinity of the stable data. Is done.

The signal output from the third register is
It becomes a white level reference value for each block one line before the read line.

In the write operation, the address is set to PK.
The white level reference value for each block input from the peak detection holding unit 72 is written to the designated address. FIG. 5 shows a series of operation timing charts.

The white level reference value for each block in the previous line output from the third register in the block-by-block peak storage unit 73 is transmitted via the shift register 75 to
The multi-level image signal whose timing is matched with the white level reference value of each block of the previous line is input to the comparison and selection unit 76, and the image signal level is compared for each pixel, and the higher level is selected.

The data output from the comparing and selecting section 76 is a white level reference value which is a reference for determining whether a read pixel is white or black in the present invention.

According to this configuration, the white level reference value serving as a reference in the intermediate level image signal processing is such that the line preceding the read line is divided into a plurality of blocks and the maximum value in the block is obtained. If the block length is set according to the required processing accuracy, a nearly accurate white level reference value can be easily obtained.

Further, before entering the character processing operation, the white level reference value is compared with the black-and-white determination pixel, and the higher level is set as the subsequent white level reference value. For example, the white level reference value will be detected. Even if there is no white area in the block, if an image signal of white level is input at the time of actual black and white determination, the image signal of white level will be immediately replaced with the subsequent white level reference value, and always accurate black and white Judgment becomes possible.

FIG. 6 is a graph showing a data waveform obtained by actually measuring the state of the white level reference value at that time for the multi-valued image signal for one line.

In FIG. 6, a fine downward waveform shown in the multi-valued image signal indicates a case where a thin line such as a character image is read, and most of the waveform is obtained by reading a white area which is a base of the original. The fact that the level decreases near the center indicates a phenomenon that is often seen when the document is separated from a predetermined reading position.

As can be seen from this figure, it can be understood that the white level reference value well follows the image signal level of the white area of the multi-valued image signal.

FIG. 7 is a block diagram showing the overall configuration of the correction pixel number setting unit 8 shown in FIG. 1. In this correction pixel number setting unit 8, the white level output from the white level reference value generation unit 7 (not shown) is used. Depending on the percentage of the maximum value of the image signal level when the white area is read,
The number of correction pixels to be changed is set.

As shown in FIG. 7, in the correction pixel number setting unit 8, the white level reference value is 75% or more of the maximum value of the image signal level when the white area is read, 74 to 50%,
It is determined which of these four levels the range belongs to, such as 9 to 25% and less than 25%, and codes [A] and [B] indicating the ranges to which each belongs are created. These codes [A] and [B]
Is an image signal selection unit 93 in the black pixel correction unit 9 described later.
Becomes a select signal for selecting the binary image signal after the correction in step (1).

In this embodiment, the white level reference value is
If it is 75% or more of the maximum level value, the code [A = 0, B
= 0], 74 to 50%, the code [A = 1, B =
0], 49 to 25%, the code [A = 0, B =
1], less than 25%, indicates a state where codes [A = 1, B = 1] are set.

FIG. 8 is a block diagram showing the overall configuration of the black pixel correction unit 9 shown in FIG.
Code [A] output from the correction pixel number setting unit 8
According to [B], black pixels are corrected by a desired number of pixels. FIG. 9 shows one-pixel correction unit 9 shown in FIG.
FIG. 10 is a circuit diagram showing details of the circuit No. 1 and FIG. 10 is a diagram showing a time chart of this circuit.

In FIG. 8, a character processing section 6 (not shown) converts an input multivalued image signal into a binary image signal, and the image signal exceeds a predetermined number of pixels set in advance in the image signal. Only the continuous black pixel group is input to a one-pixel correction unit 91 (first stage) that changes one black pixel at the end of the black pixel group to a white pixel.

Then, the one-pixel correction unit 91 (first stage)
The image signal after the correction is input to the next-stage one-pixel correction unit (second stage) while being shifted by one cycle of the data clock, and the last black pixel is changed to a white pixel. .
In other words, here, the first-stage one-pixel correction unit (first-stage) 91
This means that two pixels from the end are corrected.

Similarly, the image signal from the one-pixel correction unit (second stage) 91 is input to the last one-pixel correction unit (third stage) 91, and finally the binary signal input to the first stage is output. The image signal is changed from black pixels to white pixels for the last three pixels.

The one-pixel correction unit 91 sets the number of pixels considered necessary for forming a character in advance,
For example, if the set number is two pixels, and if the number of consecutive black pixels is three or more, correction is performed to change the last pixel from a black pixel to a white pixel, while the number of consecutive black pixels is two or less. In this case, the operation is performed so as to output to the next stage without performing the correction. By doing so, it is possible to prevent the black pixels from being reduced more than necessary.

The uncorrected binary image signal and the binary image signals output from the first-stage and second-stage one-pixel correction units 91 are finally converted into data by the final-stage one-pixel correction unit 91. Are transferred to the shift register 92 in order to match the timings of the above, and then output to the image signal selection unit 93.

The image signal selector 93 outputs the output from the one-pixel corrector 91 described above, ie, no correction and no correction, based on a select signal based on the codes [A] and [B] from the correction pixel number setting unit 8 (not shown). One of the four types after correction is selected and output as a final corrected binary image signal.

FIG. 9 is a diagram showing a detailed circuit configuration of the one-pixel correction unit 91 shown in FIG. 8, and FIG. 10 is a diagram showing a time chart thereof.

As shown in FIG. 9, the input 2
The value image signal is first supplied to a flip-flop 1 (hereinafter, “F /
F ". ) Is shifted by one cycle of the data clock and output.

Then, the edge of the monochrome pixel is detected by the exclusive OR of the output data and the data before the shift.

Next, the edge signal is separated by a NAND gate into edges on the front side and the rear side as viewed from the black pixel, and glitches generated when passing through each gate are separated by F / F.
It is removed by F2 and F / F3.

The leading edge of the output of the F / F2 sets the next stage F / F4 to [1] at its rising edge, and the output of the F / F4 is output for one cycle of the data clock by the F / F5. Is shifted.

The reset of F / F4 and F / F5 is
This is performed by the rear edge signal which is the output of the F / F3.

The output of the F / F 6 is obtained by shifting the output from the F / F 5 by a half cycle of the data clock, and
The timing is matched with the rear edge signal obtained by the gate.

The output of the F / F 6 indicates when the number of black pixels is three or more consecutively by [1]. By gating the rear edge signal with the output, the output of the F / F 6 is changed from the black pixel. A pixel to be changed to a white pixel can be specified.

The final OR gate at the final stage
A value image signal is created and output from the one-pixel correction unit.

By performing such processing, the input binary image signal before correction “black pixel 5 → white pixel 1 → black pixel 1 →
The white pixel 1 → black pixel 3 → white pixel 2 → black pixel 2 is corrected to “black pixel 4 → white pixel 2 → black pixel 1 → white pixel 1 → black pixel 2 → white pixel 3 → black pixel 2” It can be understood that only when three or more black pixels continue, the last black pixel of the continuous black pixels is changed to a white pixel.

According to such a configuration, when a predetermined number of black pixels of the image signal after the binarization process are continuous, the black pixels in the continuous black pixel group are preset in the forward pixel direction from the last black pixel. Because the number of pixels has been changed to white pixels,
For example, the phenomenon that the thin line becomes thicker than the original image when reading the image in the original at the reading position that exceeds the depth of focus of the lens, and the phenomenon such as the collapse of the characters of the white pixels sandwiched between the black areas beforehand. Can be prevented.

Further, the number of pixels to be changed from black pixels to white pixels is automatically set according to the reading position of the document, and the number of pixels to be set increases as the lens resolution decreases. At this time, black pixels are changed to white pixels in a portion excluding the minimum black pixel necessary for forming a character, so that it becomes difficult to read the character and no blurring occurs.

[0080]

As described above, according to the present invention,
When a predetermined number of black pixels of the image signal after the binarization process are continuous, the black pixels at the end of the continuous black pixel group are changed to white pixels by a predetermined number of pixels in the forward pixel direction. As a result, it is possible to prevent so-called black loss due to the continuation of black pixels as much as possible, thereby correcting the loss of characters that occurs when the resolution of a lens, such as a thin line, exceeds the resolution, and improving the image closer to the original image. A signal can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a circuit configuration from image reading to binarization processing in an image reading apparatus to which the present invention is applied.

FIG. 2 is a block diagram showing a detailed configuration of a white level reference value creation unit shown in FIG.

FIG. 3 is a diagram showing a circuit configuration of a peak detection holding unit in a white level reference value creation unit shown in FIG. 2;

FIG. 4 is a timing chart for explaining the operation of the white level reference value creation unit shown in FIG. 2;

FIG. 5 is a timing chart illustrating the operation of the white level reference value creation unit shown in FIG. 2;

FIG. 6 is a view showing a white level reference value obtained by a white level reference value creation unit shown in FIG. 2 together with an input image signal at that time.

FIG. 7 is a block diagram illustrating a detailed configuration of a correction pixel number setting unit illustrated in FIG. 1;

FIG. 8 is a block diagram illustrating a detailed configuration of a black pixel correction unit illustrated in FIG. 1;

9 is a circuit configuration diagram of a one-pixel correction unit shown in FIG.

FIG. 10 is a timing chart showing the operation of the one-pixel correction unit shown in FIG.

FIG. 11 is a schematic configuration diagram of a flat bet type image reading and sending for reading a general book type original.

12 is an overall configuration diagram showing the sensor unit shown in FIG. 11 as an example of a contact image sensor.

FIG. 13 is a diagram showing a circuit configuration from image reading to binarization processing in a conventional image reading apparatus.

FIG. 14 is a diagram illustrating a configuration of a character processing unit illustrated in FIG. 13;

FIG. 15 is a diagram illustrating the relationship between an image signal level obtained by reading a character processing state in FIG. 13 from a document and a document density.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image sensor 2 Amplifier 3 A / D converter 4 Shading correction part 5 Photo processing part 6 Character processing part 7 White level reference value creation part 71 Block division part 72 Peak detection holding part 72a First register 72b AND gate 72c Comparison selection part 72d Second register 73 Block-by-block peak value storage unit 74 Memory control unit 75 Shift register 76 Comparison / selection unit 8 Correction pixel number setting unit 9 Black pixel correction unit 91 1-pixel correction unit 92 Shift register 93 Image signal selection unit

Claims (5)

[Claims] 1. An image reading apparatus which reads an image on a document by an image sensor, binarizes a multi-valued image signal output from the image sensor into white or black, and outputs the image. When a predetermined number of black pixels of the image signal continue, black pixel correction means for changing from the last black pixel of the continuous black pixel group to white pixels by a predetermined number of pixels in the forward pixel direction is provided. An image reading apparatus, comprising: 2. The method according to claim 1, wherein said black pixel correcting means is set in advance only from a last black pixel of the continuous black pixel group in a pixel direction in front of only the black pixel group exceeding a predetermined minimum required number of components. 2. The image reading apparatus according to claim 1, wherein the number of white pixels is changed to the number of white pixels. 3. The image reading device according to claim 1, wherein the black pixel correction unit sets the number of pixels to be changed to white pixels according to the maximum value of the image signal level detected on the read line. apparatus. 4. The black pixel correction unit divides a pixel column in a read line into a plurality of blocks,
4. The image reading apparatus according to claim 3, wherein the number of pixels to be changed to white pixels is set according to the maximum value of the image signal level in the divided block. 5. The black pixel correction means changes to a white pixel according to the higher image signal level of the maximum value of the image signal level of the read line and the maximum value of the image signal level of the previous line. 4. The image reading apparatus according to claim 3, wherein the number of pixels is set.