JP3523045B2 - Image reading device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image input device such as a scanner or a facsimile, and more specifically, it optically reads an image on a reading medium and converts the image into an analog electric signal, and the analog signal is digitally converted. The present invention relates to an image input device for further converting into a signal and outputting the signal.

[0002]

2. Description of the Related Art Conventionally, there is an image input device shown in FIG. 12 as an image reading circuit (patent application number: Sho 63-296557). According to this image reading apparatus, a peak holding circuit for holding the peak value of the read value based on the one-line reading of the input document is provided, and this peak value is used as the reference value for the binarization threshold generation.

[0003]

However, in the prior art, when there is a large variation in the read value caused by variations in the sensitivity of the sensor, variations in the light source, etc., the white peak value of one line is held and this value is used as the reference value. However, since there is a large variation in the input level, there is a problem in that the threshold value cannot be stably generated and the image binarization process is not performed correctly.

Therefore, an object of the present invention is to provide an image input device which can perform stable binarization and can simplify the user's operation.

[0005]

In an image reading apparatus for optically reading an image on a reading medium and converting it into an electric signal, an input data holding unit for temporarily holding input data of one line, and an input data holding unit The object of the present invention is achieved by an image reading apparatus that includes a density histogram distribution generation unit that generates a density histogram distribution of each unit and a threshold generation unit that generates a binarization threshold value from the density histogram distribution.

In particular, the present invention is characterized by including a correction unit for correcting shading of input data, and holding the corrected input data in the input data holding unit. In addition, the threshold value generation unit uses the density histogram distribution generated by the density histogram distribution generation unit to determine the white peak value (W
P), the black peak value (BP) of the black portion of the character is calculated, and the threshold value is
It is characterized in that the threshold value is generated under the condition of × (WP-BP) + BP (0 <α <1).

Further, the input data holding unit is characterized in that it holds only the input data at a limited place on one line designated in advance. Furthermore, the input data holding unit reads one line of the target paper in advance before reading and holds it in the input data holding unit.
More specifically, the input data holding unit holds the read data at the head of the read target in the input data holding unit,
The read data of the last one line to be read is held in the input data holding unit, or the input data at the time of line reading is held in the read N lines (N: an integer of 1 or more) at a ratio of one line. The density of the input data is distributed between the density of the background of the paper and the density of the character portion (ink). Generally, the character area includes a space between characters, so that the density of the background of the paper is distributed at a constant rate. In addition, if the optical resolution of the scanner is high, the distribution of character lines also includes the density level of ink at a constant rate. Therefore, the background level and the black level of the character portion can be detected from the density histogram distribution.

Even in the case of input data of a sensor in which sensitivity variations due to the influence of a sensor or illumination occur, a white peak value (white peak: WP) of the input level of the background (white) of the paper is obtained from the density histogram distribution of the input data. , And the black peak value (black peak value: BP) of the input level of the character portion are held as reference values. By generating a threshold value from this reference value, the determined threshold value does not exceed the white and black levels, and correct binarization can always be performed.

As a method for correcting the static sensitivity variation of the sensor, the illumination, etc., there is also a shading correction process for correcting the black-and-white input level which is read in advance. By performing the shading correction processing, there is an effect of expanding the input dynamic range over the entire sensor area. First,
The input data after the shading correction processing is retained. A reference value for threshold generation is obtained from this input data, and the threshold is generated. In this method, the pseudo-halftone input using the dither using the expanded dynamic range (defined by combining a plurality of binarized data in a certain area) can always be processed with the correct threshold value.

The threshold generated by the threshold generator is: threshold = α × (white peak value−black peak value) + black peak value
By generating with (0 <α <1), it is possible to generate an optimum threshold value that does not exceed the white and black levels. In particular, the case of α = 0.5 corresponds to the case where the maximum margin is set for the white and black levels.

When the paper to be read is shorter than the sensor length of the image reading unit (for example, when reading a business card with an A6 size scanner), the density histogram distribution includes white peak values and black peaks including those other than the target paper. I will ask for a value. Except for the density range to be read, by generating a threshold value using the white peak value and the black peak value from the density histogram distribution of the input data at a limited location on one line specified in advance (for example, business card size) Even when the background of is also input at the same time, the binarization threshold can be generated.

It takes a lot of processing time to generate the density histogram from the input data and obtain the black-and-white peak value by software. Therefore, these series of processes are different from the actual reading process. Apart from the original reading operation, one line of the paper is read only once, and the optimal threshold for binarization is determined. By using the same binarization threshold value for the same type of paper after that,
The binarization process can be performed without causing a delay in the reading process.

At the time of reading the processing of generating a density histogram from the input data and obtaining the black and white peak value (of the area)
At the beginning of. The optimum binarization process for the area to be read can be performed. In many cases, the same type of paper is continuously read to read the paper. In particular, when the reading area is smaller than the document size, it can be expected that both the background and characters are included at the end of reading. Therefore, the input data is read at the end of reading. For the next reading, an optimum binarization process can be performed by generating a density histogram from the held input data and performing a process for obtaining a black and white peak value.

The base of the paper may have locally different white properties. By generating the density histogram distribution from the input data and sequentially performing the process of obtaining the black-and-white peak value at the same time as reading the data, the optimum binarization process can always be performed.

[0015]

1 is a drawing showing the configuration of an image reading apparatus according to an embodiment of the present invention. The image reading apparatus according to the present embodiment binarizes an image reading unit that optically reads an image on a medium on which an image to be read is recorded and converts the image into an electric signal, and an image signal output from the image reading unit. Image signal processing unit (binarization unit), input data holding unit that temporarily holds input data from the image reading unit during line reading, and density histogram distribution generation unit that creates a density histogram distribution from the input data And a threshold generation unit that generates a binarization threshold value for the image signal processing unit from the white peak value of the background value of the density histogram distribution and the black peak value of the character.

The image reading section comprises a sensor section and an A / D conversion section. As the sensor unit, a line sensor including a CCD or the like is used in this embodiment. Since the sensor type is not directly related to the present invention, any other type of sensor can be used.
The sensor unit outputs an analog electric signal having a value according to the density of the image read from the medium.

The A / D conversion unit converts the analog electric signal output from the sensor unit into a digital signal. Here, if an 8-bit digital signal is handled,
If the read result is white, 255 is displayed; if it is black, 0 is A /
It is output from the D converter. FIG. 2 is a drawing showing the flow of the binarization processing according to this embodiment. As shown in FIG. 2, the image reading apparatus of the present embodiment executes a threshold generation process and a binarization process of the read image.

First, in order to generate a threshold value, the image of the background (white portion) portion including the characters of the medium is read for one line, data is input, and this is stored in the line data holding unit. Also, the input data is passed to the density histogram distribution generation unit. The density histogram distribution generation unit generates a density histogram distribution from the density value of each pixel of the input data. In the threshold generation unit, based on the density histogram generated by the density histogram distribution generation unit, a white peak value indicating white of the background and a black peak indicating black of the character are obtained, and a threshold value is generated from this black and white peak value, It is sent to the image processing unit (binarization unit). 1
When the read target area of the line is all black or all white, the white peak value and the black peak value cannot be obtained only by reading one line. In this case, the threshold value is generated using the white maximum value (255) and the black minimum value (0) prepared in advance,
Perform the following processing.

Next, in the image reading process, the image processing unit uses the threshold value of the threshold value generating unit to binarize the digital image signal from the image reading unit. Since the threshold value generated by the threshold value generation unit is generated from the input data in which the background (white part) and the character part (black part) of the medium are read,
The threshold value is optimal for the original, and the binarization is performed correctly. FIG. 3 shows an outline of a pen-type scanner device, which is an example incorporating the image reading device shown in FIG. 1, and an internal configuration diagram thereof. The pen-type scanner device according to the present embodiment employs a mode in which the image reading unit is brought into close contact with the document to read the document. By moving the pen-type scanner over the read document and scanning the document in this way, and binarizing the obtained input data based on the determined threshold value, the binarized image data is output. There is. FIG. 4 illustrates the control and data flow at the time of image reading in the pen-type scanner shown in FIG. In this example, an example of reading data for four lines is shown. In FIG. 4, the encoder pulse corresponds to the movement amount signal output from the movement amount detection unit. Further, the line sensor activation signal is a signal generated every one driving cycle of the line sensor, and based on this, the signal transfer from the line sensor is started. Figure 3
In, it corresponds to the drive signal. The line sensor image signal is an image signal output from the line sensor based on the line sensor activation signal. The image processed signal is an image signal that has been binarized.

The data transfer interrupt to the apparatus will be described later in detail, but is performed from the image reading apparatus to the information processing apparatus when the image data is transferred from the image reading apparatus to the external information processing apparatus. The control of the reading operation and the flow of data will be described below with reference to FIGS. 1 to 4. The image reading unit includes a line sensor, a light source (LED or cold cathode tube), and a lens, and performs image reading in close contact with the reading medium. During reading, the light source is turned on and the line sensor starts reading with a start signal, thereby converting the amount of reflected light at each position on the sensor into an electric signal and outputting it as an analog signal. The movement amount detection unit converts the movement amount on the reading medium into a pulse signal by the encoder and outputs the pulse signal. Thereby, the relative movement amount of the image reading device with respect to the medium is detected. The image signal processing unit digitizes the image signal from the image reading unit by an A / D conversion circuit and performs binarized image processing with a threshold value determined from the density histogram distribution.
The control unit controls the entire reading and transfers the data to the information processing device. The start signal to the line sensor is always generated during the reading period of the medium, and the line data (image data) that has undergone image processing is stored in a buffer inside the image reading device.

When the movement pulse (encoder pulse) from the movement amount detection portion is detected, the control portion raises an interrupt to the information processing device to transfer the image-processed line data stored in the buffer. The movement pulse is
During normal operation, only one line sensor is output in one driving cycle (line sensor image signal 1 in FIG. 4). When the binarization processing is performed, the image reading device issues a data transfer interrupt to the information processing device, and the image data is transferred to the information processing device.

However, when the movement pulse is detected twice or more within one driving cycle of the line sensor, the control unit judges that the movement of the scanner exceeds the reading performance of the scanner, and the movement portion corresponds to one movement pulse. Transfer data only.
In the example of FIG. 4, it corresponds to the line sensor image signal 2. Here, two encoder pulses appear between the line sensor activation signals. On the other hand, when the movement pulse is not detected within one driving cycle of the line sensor, it means that the scanner is not moving. In FIG. 4, it corresponds to the portion of the line sensor image signal 3. Here, no encoder pulse appears between the line sensor activation signals. In this case, the image-processed line data stored in the buffer is discarded and the read data (line sensor image signal 4) of the next line is held.

The I / F section is an electrical interface to an information processing apparatus connected to the image reading apparatus, and performs interrupts and data access to the information processing apparatus. The binarized data transferred to the information processing device is displayed on the display unit of the information processing device. Therefore, the operator can confirm the image data being read from the display unit. FIG. 5 is a diagram showing a block diagram of the image reading apparatus according to the second embodiment. The image reading apparatus shown in FIG. 5 is an image reading unit that optically reads an image on a reading medium and converts it into an electric signal, and an image signal that binarizes the image signal output from the image reading unit after shading correction. A processing unit (binarization unit), an input data holding unit that holds input data after shading correction during line reading, a density histogram generation unit that generates a density histogram distribution from the data of the input data holding unit, and a density histogram A threshold generation unit that generates a binarization threshold value for the image signal processing unit from the white peak value of the background value and the black peak value of the character is provided.

The image reading section is composed of a sensor section and an A / D conversion section, which digitizes and outputs the image signal to be read (for example, in the case of 8 bits, white is 255 and black is 0). The output of the image reading unit is converted by the shading correction unit into an output from which the influence of surface unevenness is removed.
The background value read in advance is saved as a white shading correction value, and the input image signal that has been input is always converted and output according to the following equation (1).

Input image signal: x (i), corrected signal: y (i) White shading correction value: ww (i) y (i) = 255x (i) / ww (i). ． ． (1) The image signal converted by the above equation is binarized as in the first embodiment. Therefore, one line of shading-corrected input data is held in the input data holding unit. On the other hand, a threshold value is generated from the input data as in the case of the first embodiment. Although the white shading correction processing is shown here, the black shading correction may be performed in the same manner. FIG. 6 is a diagram for explaining an image reading operation according to the third embodiment of the present invention, and is a graph showing the relationship between the input waveform of the background, the density histogram distribution, and the determined threshold value. The graph on the left side of the drawing shows the input value of the read signal for one line including the background of the read document and the black portion corresponding to the character. The horizontal axis is the pixel in the lateral direction of the medium (pixel)
And the vertical axis represents the density value. Further, the graph on the right side of the drawing shows the density histogram distribution of pixels. The horizontal axis in the figure represents the density value, and the vertical axis represents the frequency of occurrence of the density value.

Here, even if there is a variation in the sensitivity of the sensor, a white peak value (WP) indicating the background and a black peak value (BP) indicating the characters appear in the density histogram distribution. Therefore, if a threshold value is generated inside the white peak value and the black peak value, the read image can be binarized. That is, if a threshold value is generated at a position as shown in the figure, a signal near the black peak value is correctly output as black and a signal near the white peak value is correctly output as white. The threshold value is calculated by the following formula.

Threshold = α × (W P-BP) + BP (1 <α <1) Note that α is a coefficient for determining the threshold, and an arbitrary value can be applied. The value of α is
It may be appropriately selected according to the desired image quality of the output image. Thus, the white peak value (W
P) and black peak value (BP)
It can be valued. FIG. 7 shows an image reading apparatus according to the fourth embodiment. When the width of the paper to be read is shorter than the reading area of the image reading unit, even if the line input data value is obtained, the non-target portion (outside the document) is also read. Therefore, data unrelated to the medium is input, and the density histogram distribution of the paper cannot be correctly obtained. Therefore, in the present embodiment, in the effective area designating section, it is used as an input data value of only data at a limited location on a line designated in advance,
Generate a density histogram distribution. To specify the effective area,
For example, the width of the medium to be read may be specified. The image reading device recognizes a portion to be treated as an effective signal in the read signal based on the designation of the effective area, generates a density histogram only for that portion, and generates a threshold value (effective area in FIG. 7). As a result, the optimum binarization threshold value can be generated even for a sheet having a line length shorter than that of the image reading unit. FIG. 8 is a view showing an image reading device according to a fifth embodiment of the present invention. 9 to 11 are flowcharts showing the procedure of the reading process in the image reading apparatus shown in FIG.

The above drawings show the relationship between the line input data holding for threshold value determination and the actual reading processing. There are four types of line inputs for determining the threshold value. 1) Separate processing type In addition to reading the actual image, one line of paper is input and the threshold value is determined in advance from the density histogram distribution,
For subsequent reading, binarization processing is performed using the determined threshold value. 2) One line of the leading line of the pre-reading type reading start is input, the threshold value is determined from the density histogram distribution obtained as a result of reading, and at the time of reading, the binarization process is performed with the threshold value determined at the beginning. 3) Input one end line of post-reading type reading and determine the threshold value from the density histogram distribution. For the subsequent reading of the original, the binarization processing is performed with the threshold value determined at the end of the previous reading. 4) Simultaneously with image input of one line of sequential type, a threshold value is determined from the density histogram distribution for each line, and binarization processing is always performed with the latest threshold value using the determined threshold value.

By adopting such a configuration,
The hardware configuration and the configuration according to the priority of the processing time can be selected.

[0030]

As described above, according to the present invention, in the image input device, when the variation of the read value caused by the variation of the sensitivity of the sensor or the variation of the light source is large, the input value of the line of the sensor is large. , And a black-and-white peak value representing the white of the background and the black of the character is obtained from the density histogram distribution. By determining the threshold value with this reference value, it is possible to always perform correct binarization, but it is possible to provide an image input device.

[Brief description of drawings]

FIG. 1 is a configuration of an image reading apparatus according to an embodiment of the present invention,

FIG. 2 is a diagram showing an image reading processing procedure according to an embodiment,

FIG. 3 is a specific configuration example of an image reading device,

FIG. 4 is a timing chart of image reading,

FIG. 5 is a configuration of an image reading device according to a second embodiment,

FIG. 6 is a drawing showing a relationship between a read image signal and a density histogram,

FIG. 7 is a diagram illustrating an example in which the effective area is narrower than the reading range of the image reading device;

FIG. 8 is a view for explaining image reading according to another embodiment,

FIG. 9 is a drawing showing a binarization processing procedure according to another embodiment,

FIG. 10 is a drawing showing still another binarization processing procedure;

FIG. 11 is a drawing showing still another binarization processing procedure;

FIG. 12 is a drawing showing a conventional binarization process.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Tsuguo Noda               4-1, Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa               No. 1 within Fujitsu Limited                (56) Reference JP-A-1-258558 (JP, A)                 Japanese Patent Laid-Open No. 4-368058 (JP, A)                 JP-A-5-63973 (JP, A)                 JP 63-246975 (JP, A)                 JP-A-4-284778 (JP, A)                 JP-A-2-283165 (JP, A)

Claims (3)

(57) [Claims] 1. A hand-held image data input section for scanning by moving on a reading medium to optically read an image on the reading medium, and before reading for binarizing the input data. An input data holding unit that temporarily holds the input data of the first line of the image including the characters on the reading medium in advance , and a character that is temporarily stored in the input data holding unit Picture
A density histogram distribution generation unit that generates a density histogram distribution from the input data of the first line of the image, a threshold generation unit that generates a binarization threshold from the density histogram distribution, and a threshold value generation unit when reading the medium. Was done 2
An image reading apparatus, comprising: an image processing unit that performs binarization processing of input data obtained from a medium based on a binarization threshold value. 2. The image reading apparatus further comprises a correction unit that performs shading correction of input data, and holds the corrected input data in the input data holding unit. Image reading device. 3. In the image reading apparatus, the threshold value generation unit uses the density histogram distribution generated by the density histogram distribution generation unit to determine the white peak value (WP) of the white portion of the paper and the black peak value (BP) of the black portion of the character. The image reading apparatus according to claim 1, wherein the threshold is generated and the threshold is generated under the condition of α × (WP−BP) + BP (0 <α <1).