JP3548320B2 - Image reading device - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading apparatus applicable to a digital copying machine, an electronic file device, and the like, which can capture high-quality images.
[Prior art]
Conventionally, there is one in which data of one scanning line is read simultaneously by a plurality of CCDs and averaged to improve the S / N ratio (Japanese Patent Application Laid-Open No. Hei 4-337679). In addition, there is a device in which gamma correction is performed before A / D conversion and the low-level portion is extended to maintain the gradation of the low-level portion (JP-A-5-284354). In addition, there is one in which the S / N ratio is improved by performing arithmetic processing on pixel data of a read scanning line (Japanese Utility Model Application Laid-Open No. 5-76170).
[0002]
[Problems to be solved by the invention]
Due to the nature of the image reading device, image data in a dark portion is unstable and noise is conspicuous, and gradation is not reproduced clearly. Therefore, conventionally, in order to reduce noise, power or cost is consumed by increasing the light amount of a light source, cooling a reading sensor (CCD), performing complicated and large-scale signal processing, and the like. Further, in order to obtain stable read data, the read operation is delayed. With the recent increase in speed, many CCD outputs are divided into even-numbered pixels and odd-numbered pixels, but vertical stripes in dark areas due to variations in registers are generated. In addition, flare light in the sub-scanning direction such as a black line in a bright portion may cause a line to be thin. In the above-mentioned conventional means, regardless of the image data of the read original, even if there are many bright parts, the CCD is driven by a low-speed clock so as to obtain a sufficient signal or performs integrated averaging over the entire original. For example, it is difficult to increase the reading speed. Therefore, the present invention takes advantage of the fact that it is difficult for human eyes to perceive image quality degradation even when the resolution is reduced in dark areas, and it is important to determine whether or not there is a continuous dark area in the image of the original to be read. In addition to determining whether there are black streaks in the scanning direction, if there is a continuous dark area, the black line in the bright area should be clear without lowering the reading speed, clear gradation and no vertical streaks. An object of the present invention is to provide an image reading apparatus capable of reliably reading an image.
[0003]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, in the first invention, light reflected or transmitted by scanning a document surface is condensed by an imaging lens and made incident on a photoelectric conversion element to obtain image data based on an electric signal. In a reading device, a document image discriminating means for discriminating a dark portion by detecting a plurality of continuous pixels having a density level near a black reference level for each reading scan line of a document image, and a discrimination result of the document image discriminating means. A continuous averaging for successively outputting the number of continuous pixels determined as a dark area by integrating and averaging continuous image data determined for each predetermined number of pixel data, and replacing the integrated averaged data with the pixel data determined to be the dark area. And an image reading device.
In the second invention, the document image discriminating means is provided with a comparator for comparing the pixel data with a preset density level, and counting the number of times the same comparison result of the density level by the comparator is repeated, and counting the number of times by a predetermined number. If there is, a counter for issuing a control signal to the integrating and averaging means is provided.
According to a third aspect of the present invention, in an image reading apparatus, light reflected or transmitted by scanning a document surface is condensed by an imaging lens and made incident on a photoelectric conversion element to obtain image data by an electric signal. A data holding unit for holding pixel data for each of a plurality of read scanning lines; and a plurality of pixel data for a plurality of read scan lines held by the data holding unit, which are continuous in a sub-scanning direction at a density level near a black reference level. A document image discriminating means for discriminating a dark portion by detecting pixels of the plurality of scanning lines held in the data holding means, the pixel data of the plurality of scanning lines are discriminated as a dark portion by the discrimination result of the document image discriminating means. Continuous image data is integrated and averaged for each of a predetermined number of pixels at the same scanning position, and the integrated averaged data is determined as the dark portion. To and a cumulative averaging means for outputting the number of pixels consecutive in place of the pixel data to be continuous, which is to constitute a digital image reading apparatus.
[0004]
BEST MODE FOR CARRYING OUT THE INVENTION
An embodiment of the present invention will be described with reference to the drawings. In FIG. 1, an image reading device 20 includes a contact glass 2 on which a document 1 is placed, a mirror 3 for guiding reflected light from the document 1 to a predetermined position, and a light source 4 for irradiating light for scanning the document 1 in the main direction. A motor 5 for moving the contact glass 2 for sub-scanning the original 1, an imaging lens 6 for condensing the reflected light guided by the mirror 3 to form an image, and photoelectrically converting the formed image. A CCD 7, a sample-and-hold circuit 8 for removing a transfer clock component of the CCD 7, an A / D conversion circuit 9 for converting an image signal into a digital signal, and an arithmetic operation by reading out data stored in advance in a correction memory for each pixel and calculating the data. A dark output correction circuit 10 for correcting the dark output of each pixel of the CCD 7 and a pre-stored data read out from a correction memory and operated for each pixel to provide an illumination system, A shading correction circuit 11 for correcting variations in brightness in the system and variations in sensitivity of the CCD 7, a line density conversion circuit 12 for performing density conversion in the main scanning direction as necessary, and outputting signals to an output device or a storage device Interface 13, a light source drive circuit 14 for driving the light source 4, a motor drive control circuit 15 for driving the motor 5, and a light accumulation time which is set so as to obtain a sufficient output of the CCD 7 even in a multi-valued image. And a document image for discriminating a plurality of (n) pixels (dark portions) having a black density level (hereinafter referred to as a pixel level) of a pixel near a black reference level for each scanning line for reading a document image According to the determination result of the determination unit 17 and the document image determination unit 17, the image signal of the read scanning line is multiplied for each predetermined number (n) of pixels. Averaged, the same number (n pieces) of the new pixel signal averaged density level and a cumulative averaging means 18 for outputting.
[0005]
The image reading device 20 irradiates the original 1 placed on the contact glass 2 with illumination light from the light source 4, guides the reflected light to the CCD 7 through the mirror 3 and the imaging lens 6, and places the reflected light on the surface of the CCD 7. An original image is formed. The light source 4 is driven by a light source driving circuit 14 for main scanning, and the contact glass 2 is moved in the X direction by a motor 5 for sub-scanning. The output signal of the CCD 7 is removed from the transfer clock component of the sensor by the sample and hold circuit 8, and the output is converted to a digital signal by the A / D conversion circuit 9. The A / D conversion output is obtained by reading out the data stored in advance in the correction memory for each pixel in the dark output correction circuit 10 and calculating the dark output for each pixel of the CCD 7. A variation in brightness due to the imaging system, a variation in sensitivity of the CCD 7, and the like are corrected. In this case, similarly to the correction of the dark output, data stored in advance is read out from the correction memory and calculated for each pixel. The line density conversion circuit 12 performs density conversion in the main scanning direction as necessary, and sends an output signal to an output device or a storage device through the interface 13.
[0006]
As shown in FIG. 2, the output signal (i) of each line of the CCD 7 includes an output DS (E) of a pixel having only a transfer section without a photosensitive section and an output DS (S) of a pixel having a light-shielded section. And the following signal of the read document image data D. The output signal (i) of the CCD 7 is converted into an analog signal (ii) by removing the transfer clock component by the sample and hold circuit 8. The output signal (ii) of the sample-and-hold circuit 8 is detected and held by another sample-and-hold circuit (not shown) as a signal DS (E), held as the upper limit reference voltage of the A / D conversion circuit 9, and a white reference plate. The dynamic range of the A / D conversion circuit 9 is determined by detecting and holding the peak voltage of the output when the data is read by a sample and hold circuit (not shown) and setting the reference voltage at the lower limit of the A / D conversion circuit 9.
[0007]
The original image discriminating means 17 first compares the output signal of the sample and hold circuit 8 amplified by an amplifier (not shown) with a preset reference voltage to determine the dark level of the original (the dark level having a black density near the black reference level). It is determined whether or not the pixel is within the range. The reference voltage value is a value slightly larger than the black reference level (several digits in digital value, for example, about 5 digits when the white level is 256 digits), and is experimentally determined for each system. Is determined. As shown in FIG. 3, the comparator 17a of the document image discriminating means 17 compares the magnitude of the image data signal with reference to a preset reference voltage value Vref. The comparison result is latched by the clock CLK by the latch circuit 17b. The clock CLK has a phase delayed from the hold pulse of the sample hold circuit 8 in the delay circuit 17d, and has a cycle synchronized with the pixel data signal. The latch circuit 17b detects a dark-level pixel data signal by latching the pixel data signal, and outputs "H". The counter 17c receives the output signal of the latch circuit 17b, generates an "H" output signal Cout when n consecutive "H" signals, and determines that the image is a dark portion. In the present embodiment, a case where the dark level continues for three or more pixels is determined as a dark part. For example, as shown in FIG. 4, when the image data signal changes like a thin line A, the output of the latch circuit 17b becomes "H" for one clock cycle, and the output Cout of the counter 17c becomes "L". Become. When the image data signal changes as indicated by the thick line B, the output of the latch circuit 17b becomes "H" for three clock cycles from t1 to t3, and the counter output Cout becomes "H". While the dark portion continues, the output Cout of the counter 17c keeps “H”. Here, in the counter 17c, the number of continuous pixels required to determine a dark portion is set in advance.
[0008]
The integrated averaging means 18 receives the output Cout of the document image discriminating means 17. The integration and averaging means 18 performs integration and averaging of the continuous n pieces of pixel data determined as dark areas based on the output Cout of the document image determination means 17 and outputs n pieces of new pixel data. In the integrating and averaging means 18, as shown in FIG. 5, the input / output timing of the signal is based on the clock CLK. The image data signals D1, D2,... For each pixel in one scanning line are continuously taken in as the data signal a by the integrating and averaging means 18 in synchronization with the clock CLK. Based on the image data signals D1, D2,..., When the number of integrations n = 3, if the number of dark-level pixels becomes three in a row, a control signal b is generated, and at the same time, a period of two pixels continues. Is generated. An integration process is performed based on the control signals b and c. When the integration is not performed, that is, when the dark portion is not determined by the document image determination unit 17, the control signals b and c are not generated, and the output of the A / D conversion circuit 9 is output by the integration averaging unit 1. 8 Is output directly to the dark output correction circuit 10 as it is. On the other hand, when integration is performed (in FIG. 5, pixels D1 to D3 are dark levels), pixel data D1 is stored in delay element 18a and storage element 18b of integration and averaging means 18 as shown in FIG. Written. The delay element 18a sequentially takes in the pixel data, and the storage element 18b stores the pixel data of the comparator 17a of the original image discriminating means 17. output Captures pixel data only during the period of “H”. At the input timing of the pixel data D2, the pixel data D1 is read from the storage element 18b, is simultaneously added to the pixel data D2, and is written into the storage element 18b again. Then, when the third pixel data D3 is determined to be a dark portion and the signal b becomes "H", the pixel data D3 is integrated into D1 + D2, written into the storage element 18b, and thereafter averaged by the averaging circuit 18c. The new pixel data ND1 is latched by the latch circuit 18d and output to the dark output correction circuit 10 side. Here, when the data D4 is not determined to be a dark portion and continuous dark portion data is interrupted, ND1 is continuously and repeatedly output as the second and third pixel data by the control signal c. Further, when the four pixel data from the pixel data D5 to D8 are successively determined to be a dark portion, the pixel data D5 + D6 of the storage element 18a are integrated and written to the storage element 1 when the third pixel data D7 is read. , The newly averaged data ND2 is output, and when the data D8 is read, the data D6 + D7 is integrated and written into the storage element 18b, and the newly averaged ND3 continues as the second and third pixel data. Output repeatedly. Thereafter, the above operation is repeated. The pixel data D4 that is not determined as a dark portion is selected by the selection unit 18e after being delayed by, for example, two pixels by the delay element 18a, and is output as it is. At this time, the storage element 18b does not take in the pixel data. In the present embodiment, the condition number n for determining the dark portion is set to 3, but the same processing is performed when any other number is set. Although the data bus (the number of bits) increases during integration, the value to be integrated is very small (near the black level), so that there is no need to prepare extra in advance.
[0009]
Another embodiment is shown in FIG. In the present embodiment, the integrating and averaging means 27 is provided with line storage elements 27a, 27b, and 27c for a plurality of lines (here, three lines). Is determined, and integration averaging is performed. That is, the first line data L1 written to the line storage element 27a is written to the line storage element 27b simultaneously with the reading of the next second line data L2, and the storage element is simultaneously written with the next reading of the third line data L3. 27c. Therefore, when the reading of the three line data is completed, the line data L3 is stored in the line storage element 27a, the line data L2 is stored in the line storage element 27b, and the line data L1 is stored in the line storage element 27c. At the same time, if a plurality of (for example, three as in the left embodiment) continuous pixel data of the dark level in the sub-scanning direction is determined, the line data read by the CCD 7 is determined to be a dark portion and integrated. The averaging circuit 27d sequentially obtains integrated averaged data for three lines.
[0010]
The raw line data from the storage element 27c is compared with the reference value Vref near the black reference level for each pixel by the comparator 27e. The integrated averaged data is compared with a reference value Vref2 by a comparator 27f. From these comparison results, the switching output circuit 27g outputs the raw pixel data when the pixel level of the raw line data is higher than the reference value Vref of the comparator 27e, and lowers the raw pixel data when the pixel level of the raw line data is higher than the reference value Vref of the comparator 27e. In this case, the line integrated averaged data is output as pixel data of the first line of the line group determined to be a dark part.
[0011]
For example, as shown in FIG. 8, in the line of interest, arbitrary pixel levels are D11 <Vref, Dm1 <Vref, Dn1> Vref, and the integrated average data of D11, D12, and D13 are D1a, and Dm1, Dm2, as well. Assume that Dm3 is Dma, and Dn1, Dn2, and Dn3 are Dna, and the comparison results are Dla <Vref2, Dma> Vref2, and Dna <Vref2. At this time, the pixel data D11, D12, D13 at the same position in the main scanning direction is determined as a dark part, and an averaged Dla is output as the data of the line of interest. Even if Dn3 is determined to be a dark part, the pixel data Dn1 is output as it is because the raw pixel data Dn1 is larger than the reference value Vref. Further, since the pixel data Dm is not determined to be a dark portion, Dm1 is output as it is. Thus, if the document image discriminating means is applied to the image reading apparatus shown in FIG. 1, the variation of the output image data in the dark portions in the main and sub scanning directions is reduced.
[0012]
FIG. 9 shows another embodiment. In the following drawings, the same components as those shown in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted. In this image reading apparatus, a CCD drive control means 19 is provided in place of the integration averaging means 18 and a preliminary scan is performed before a main (fine) scan, and the original image discriminating means 17 is used during the preliminary scan. The dark part of the image is determined. When the dark portion is determined, a signal is output to the CCD drive control means 19, and the CCD drive control means 19 reads the image by lowering the drive clock of the CCD drive control means 16 to a predetermined frequency.
[0013]
FIG. 10 shows an example of the CCD drive control signal (part). a is a transfer pulse for transferring the charge of the CCD to the transfer register, and the period Tint is the light accumulation time. b is a drive clock CLK for reading out the charge of the transfer register. Here, assuming that the number of pixels of the CCD is h, the number of clocks requiring a period of Tint is at least h × CLK. c is a drive clock whose frequency has been reduced by the CCD drive control means 19.
[0014]
In a normal case where a dark portion is not determined, driving is performed using signals a and b in FIG. When an actual device has a plurality of reset pulses and a plurality of transfer registers, a drive clock is required for each of the transfer registers. However, the description is omitted here. When the original image discriminating means 17 reads a line determined to be a dark portion, the CCD drive control means 19 generates a clock c having a reduced driving clock frequency and reads out image data. The frequency of CLK is set so that all data of CCD effective pixels can be read during the period of Tint.
[0015]
FIG. 11 shows an image reading apparatus according to another embodiment. The image reading apparatus 20 includes a CCD drive control unit 21 and a bit operation unit 22 which are different from those described above. When the original image discriminating means 17 discriminates a dark portion of the image during the preliminary scan, a signal is output to the CCD drive control means 21, and the CCD drive control means 21 reads the image by lowering the CCD drive clock to a predetermined frequency. . At this time, as shown in FIG. 10, the CCD drive control means 21 generates a drive clock e whose frequency is further reduced from that of the previous embodiment and a transfer pulse d whose cycle is changed. At this time, control is performed so that the transfer pulse period Tint2 = Tint × 2k (k is a positive integer). Here, k = 1. Then, the bit operation unit 22 divides the image signal obtained by the drive signals d and e by 2k (k = 1 in this case) in order to normalize the image signal into data obtained under the initial drive conditions a and b.
[0016]
FIG. 12 shows an image reading apparatus according to another embodiment. This image reading apparatus includes a document image discriminating means 23 having a different discriminating method from the previous embodiment, and an inter-line calculating means 24 having at least five line data holding devices in place of the integrating and averaging means 18. . Then, a preliminary scan is performed before the main scan, and at the time of the preliminary scan, the document image determining unit 23 determines a dark portion of the document. When the document image discriminating means 23 discriminates a dark-level data line sandwiched between bright portions in the sub-scanning direction in the document as a dark portion, it stores the address of this data line and stores the data of the data line adjacent to the dark portion. Hold. When the original image discriminating means 23 reaches the data line in the main scan, it outputs a signal to the line-to-line calculating means 24, and the line-to-line calculating means 24 adds a predetermined value to the line data of the bright part adjacent to the dark part in the sub-scanning direction. An image is read by subtracting the value obtained by multiplying the ratio from the data of the data line.
[0017]
In this image reading apparatus, processing is performed for every five data lines for convenience. As shown in FIG. 13, in the line data group a, the data lines L1, L2, L4, and L5 are light portions and only L3 is a dark portion. From the line data group a, a signal b is generated by an integrator (not shown) having a time constant of about 1/3 of the cycle of the data line. The signal b is binarized by a reference value Vh near the white reference level and a reference value Vl near the black reference level, and then latched by a synchronization signal with the data line to generate a signal c. The signal c is supplied to counters 23a and 23b as an enable signal, as shown in FIG. A gate signal d for five line periods is generated at the same time as the signal c, and is supplied as an enable signal for the latch circuits 23c, 23d, 23e, and 23f. Each of the counters 23a and 23b counts a pulse (not shown) generated substantially at the center of the line period synchronized with the signal a, and outputs a latch signal d when the counter 23a counts two pulses and the counter 23b counts one pulse. The outputs e, f, and g of the latch circuits 23c, 23d, and 23e determine the dark portion of one line sandwiched between the bright portions when the gate 23g becomes "H". Here, when two or more dark portions are continuous, the gate 23g is set to "L" by the latch circuit 23f and is not discriminated.
[0018]
FIG. 15 shows the white reference plate 25. The white reference plate 25 is read each time the reading operation is started, the black line 25a in the white reference plate is read, and the flare light ratio from white (bright portion) is obtained from this data. The ratio is to be multiplied by the bright part to obtain a value to be subtracted from the sandwiched dark part. The obtained ratio is stored as flare correction data for each pixel data, and can be read out each time line-to-line arithmetic processing is required.
[0019]
The method of obtaining the ratio by which the bright portion is multiplied is as follows. First, the image data when continuous white (bright portion) of the white reference plate is read is W, and when the black reference plate separately provided in the reading device is read. Assuming that the image data is B and the image data when reading the black line provided on the white reference plate is Bf, the ratio Per becomes Per = (Bf-B) / W.
[0020]
If the image data is smaller than the black reference level as a result of subtracting the data multiplied by the ratio, the image data is set to the black reference level. The subtracted image data is compared with the black reference level for each pixel data. If pixel data> black reference level, the pixel data is output as it is, and if pixel data <black reference level, the black reference level is substituted. Output to
[0021]
A document image determining unit according to another embodiment will be described. The document image discriminating means discriminates a line having a dark portion after a bright portion of the document is continuous or a line of the last dark portion when a bright portion is continuous after the dark portion. When the line of the dark part is determined, the document image determining means stores the address of the line on which the determination is made and holds the data of the line adjacent to the dark part. In the main scan, when the original image discriminating means 23 reaches the corresponding line, it outputs a signal to the inter-line calculating means 24, and the inter-line calculating means 24 adds a predetermined ratio to the line data held by the original image discriminating means 23 during the pre-scan. Is subtracted from the data of the corresponding line having the dark portion, and the image is read.
[0022]
Further, a reference value for discriminating a bright portion and a dark portion of the document image discriminating means can be set from outside.
[0023]
【The invention's effect】
As described above, in the first invention, the document image discriminating means for discriminating a dark portion by detecting a plurality of continuous pixels having a density level near the black reference level for each reading scan line of the document image, Based on the determination result of the document image determining unit, continuous image data determined as a dark portion is integrated and averaged for each predetermined number of pixel data, and the integrated averaged data is replaced with the pixel data determined as the dark portion. The provision of the integrating and averaging means for outputting a continuous number of pixels makes it possible to automatically determine the dark portion of the document and reduce noise in the dark portion of the document image.
According to the second aspect, the document image discriminating means can discriminate a continuous dark portion in a scanning line for reading a document image with a simple configuration in real time and at low cost.
In the third invention, a data holding means for holding pixel data for each of a plurality of read scanning lines of a document image, and a plurality of read scan lines held by the data holding means, the pixel data near the black reference level. Document image discriminating means for discriminating a dark portion by detecting a plurality of pixels continuous in the density-level sub-scanning direction; and document image discriminating means for pixel data of a plurality of read scanning lines held in the data holding means. According to the determination result, continuous image data determined to be a dark portion is integrated and averaged for each of a predetermined number of pixels at the same scanning position, and the integrated averaged data is replaced with continuous pixel data determined to be the dark portion. By providing an integrating and averaging means for outputting a continuous number of pixels, it is possible to determine continuous dark portions in the sub-scanning direction in real time and at low cost. For example, it is possible to further reduce noise in the in the sub-scanning direction of the dark part of the original image.
[Brief description of the drawings]
FIG. 1 is a block diagram of an image reading apparatus according to a first embodiment.
FIG. 2 is a time chart of an image data signal.
FIG. 3 is a circuit diagram of a document image determining unit.
FIG. 4 is a time chart of an image data signal.
FIG. 5 is a time chart showing an input / output state of an image data signal in a document image discriminating means.
FIG. 6 is a circuit diagram of an integrating and averaging means.
FIG. 7 is a circuit diagram of a document image determining unit according to a second embodiment.
FIG. 8 is a conceptual diagram of line data.
FIG. 9 is a block diagram of an image reading device according to a third embodiment.
FIG. 10 is a time chart of a driving clock for CCD.
FIG. 11 is a block diagram of an image reading apparatus according to a fourth embodiment.
FIG. 12 is a block diagram of an image reading apparatus according to a fifth embodiment.
FIG. 13 is a time chart showing an input / output state of an image data signal in the document image discriminating means.
FIG. 14 is a circuit diagram of a document image discriminating unit.
FIG. 15 is a plan view of a white reference plate.
[Explanation of symbols]
1 manuscript
7 CCD
8 Sample hold circuit
16 CCD drive circuit
17 Original image discriminating means
18 Means for averaging
18a delay element
18b storage element
17a comparator
18e selection means
19, 21 CCD drive control means
20 Image reading device
23 Document image discriminating means
23a, 23b counter
23c, 23d, 23e, 23f Latch circuit
23g gate
24 Inter-line calculation means
25 White reference plate
25a black line
27 Means for averaging
27a, 27b, 27c Line storage element
27d integrating averaging circuit
27e comparator
27f comparator

Claims (3)

In an image reading apparatus, the light reflected or transmitted by scanning the original surface is condensed by an imaging lens and made incident on a photoelectric conversion element to obtain image data by an electric signal.
Document image discriminating means for discriminating a dark portion by detecting a plurality of continuous pixels having a density level near the black reference level for each reading scan line of the document image;
More determination result of the original image discrimination means, the image data sequential it is judged dark portion, and integration average of every predetermined number of pixel data, instead of the cumulative averaging data, to the determined pixel data with the dark portion And an integrating and averaging means for outputting a continuous number of pixels . The document image discrimination means, comparator and the accumulated averaged if a predetermined number by counting the number of times the same comparison result of concentration levels continuously by the comparator that compares with a preset density level of pixel data 2. An image reading apparatus according to claim 1, wherein said means comprises a counter for issuing a control signal. In an image reading apparatus, the light reflected or transmitted by scanning the original surface is condensed by an imaging lens and made incident on a photoelectric conversion element to obtain image data by an electric signal.
Data holding means for holding pixel data for each of a plurality of read scanning lines of a document image;
Document image discriminating means for discriminating a dark portion by detecting a plurality of pixels continuous in the sub-scanning direction at a density level near the black reference level , for pixel data of a plurality of read scanning lines held by the data holding means ,
With respect to pixel data of a plurality of read scanning lines held in the data holding unit, continuous image data determined as a dark portion is integrated for each of a predetermined number of pixels at the same scanning position based on the determination result of the document image determining unit. and averaging the accumulated and averaged data, the image reading apparatus characterized by comprising an integrating averaging means for outputting the number of pixels consecutive in place of the pixel data to be continuous is determined as the dark space.

Images