JPH10336470A - Image reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image reader capable of reducing the capacity of a delay memory. SOLUTION: This reader is provided with a read means of color image data by a three-line image sensor, a line correction processing means 21 that corrects an inter-line distance of the three-line image sensor, an in-area attribute discrimination processing means 22 that divides a read image M×N areas and discriminates the attribute of the image in each divided area as a character/ a photograph/a dot or the like, and a filter processing means 24 that applies two-dimension filter processing to the read image. Then the line correction processing means 21 and the in-area attribute discrimination processing means 22 receives the input image signal at the same time and conduct parallel processing on the circuit configuration.

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 capable of reading a color original image used in a color facsimile, a color digital copying machine, a color scanner and the like.

[0002]

2. Description of the Related Art In recent years, there have been increasing opportunities to handle color images such as color scanners, color digital copying machines, color facsimile machines and color electronic filing machines. In these apparatuses, scanning is performed in the sub-scanning direction by a three-line image sensor as image input means, image data of a document is color-separated into RGB output by the three-line image sensor, and sequentially read line by line. An image reading device capable of reading the image has become mainstream.

In order to read a color image, an apparatus of the type that separates the color while changing the light primary color and reads it with a one-line sensor, or performs color reading with a color filter for each pixel in the light receiving portion of the one-line sensor. There are devices of the type, but in terms of reading speed and resolution, a device of the type in which a color filter is placed on each line sensor of a three-line image sensor and performs color separation has better performance.

Hereinafter, a conventional image reading apparatus using a three-line image sensor equipped with such a color separation filter will be described with reference to the drawings. FIG. 9 is a configuration diagram of a conventional image reading apparatus. As shown in FIG. 9, the image reading apparatus includes a document placing glass 2 on which a document 1 is placed, and a white reference plate 4 that supplies white reference data for shading correction to an image processing unit before starting image reading. And a carriage 5 that scans the original 1 and sequentially reads the original 1 line by line. The carriage 5 is connected on a support shaft 6, and the movement of the carriage 5
Is limited to only the reading direction. When the carriage is moved, the rotation of the drive motor 7 is transmitted as a linear driving force of the carriage 5 through the drive wire 8, the drive pulley 9, and the driven pulley 10. Note that a tension generating member 11 such as a spring is connected to apply a tension to the drive wire 8.

[0005] The carriage 5 includes a light source 1 for irradiating the original 1.
2 and an aperture 13 for regulating the reflected light from the original 1
A reflection mirror 14, a shading correction plate 15 for equalizing the illuminance distribution on the three-line image sensor 17,
An image forming lens 16 for forming an image of the original 1 on a three-line image sensor 17 and a three-line image sensor 17 for reading a reflected image from the original 1 and converting the image into an electric signal are provided.

FIG. 10 is a block diagram of a conventional electric circuit. As shown in FIG. 10, the analog signal processing means includes an analog processing means 18 including a sampling circuit for amplifying an image signal obtained from the three-line image sensor 17 and correcting a variation of each pixel of the three-line image sensor 17, and the like. /
A D converter 19, a shading correction processing means 20,
Line correction processing means 21, color correction processing means 23, area attribute discrimination processing means 22, filter processing means 24
And a buffer memory 2 for temporarily storing read data
5 and an interface 26 for exchanging data with an external device. In an actual device, a timing generation circuit for controlling the operations of the three-line image sensor 17 and the A / D converter 19 and a C for controlling all electric circuits are provided.
RO in which control programs for PU and CPU are stored
M, RAM for work of CPU, I / O port for monitoring on / off of various control objects and state of control objects, drive circuit for turning on light source 12, drive circuit for carriage drive motor 13 Are also required, but are omitted and not shown.

FIG. 11 is a block diagram of a conventional electric circuit, and shows a schematic structure of a three-line image sensor 17. The three-line image sensor 17 is divided into three lines at a certain interval, and R, G, and B color filters are deposited on the three sensors by an organic filter or the like. Here, it is assumed that reading is performed with the R sensor at the top.

The operation of the image reading apparatus constructed as described above will be described below with reference to FIGS. First, when an external host (not shown) issues a command to read the document 1, the drive motor 7 is rotated,
The carriage 5 connected by the drive pulley 9 and the drive wire 8 is moved to the position of the white reference plate 4. When detecting that the position has been reached, the CPU stops the carriage 5 and turns on the light source 12.

After the light source 12 is turned on, the reading operation of the white reference plate 4 arranged at that position by the three-line image sensor 17 is started. R of 3-line image sensor 17
The GB output signal is amplified by the analog processing means 18 and input to the A / D converter 19. The A / D converted data is used for shading correction described later.
Stored in a memory (not shown). After the reading of the white reference plate 4 is completed, the CPU moves the carriage 5 again at a constant speed.

When the CPU detects that the carriage 5 has reached the reading start point of the original 1, the CPU starts the image reading operation of the three-line image sensor 17 again. The light beam from the light source 12 is applied to the reading section of the original 1, and the reflected image of the reading section of the original 1 enters the carriage 5. This reflected image changes the optical path with the reflecting mirror 14, and
After passing through a shading correction plate 15 for uniformizing the illuminance distribution, an image is formed on a three-line image sensor 17 by a lens 16 and captured as data.

Thereafter, the R of the three-line image sensor 17
The A / D converter 19 converts the GB output into digital data. A shading correction operation is performed on the image data based on the previously stored white reference data. The image data after the shading correction is corrected by the line correction processing unit 21 to correct the RGB delay in order to correct the data delay caused by the shift of the RGB reading position of the three-line image sensor 17.

Next, after the color correction of the RGB data is performed by the color correction processing means 23, the read data is divided into M × N areas by the in-area attribute discriminating processing means 22, and the attribute of the image is set in the area. (Characters / photographs / halftone dots, etc.), and based on the result of the determination, the filter processing means 24 performs adaptive filter processing. Then, the data is sequentially taken into the buffer memory 25 and output through the interface 26.

The carriage 5 reaches the end of reading the original 1.
When the is moved, the reading is terminated, the light source 12 is turned off, and the drive motor 13 is driven to move the carriage 5 to the reading start portion of the document 1 to finish the operation. With the above operation, the two-dimensional color image of the document 1 can be read two-dimensionally.

Here, the line correction processing means 21, the area attribute discrimination processing means 22, and the filter processing means 24 will be described in detail. FIG. 12 is a block diagram of a conventional line correction processing. The line correction processing means 21 corrects the input delay of the RGB data. For this reason, a delay memory for delaying the data is required. 27a,
27b and 27c are RGB data delay memories, respectively. Now, it is assumed that the sensor interval of the three-line image sensor 17 is eight lines apart in units of 600 dpi. Also,
The number of pixels of the three-line image sensor 17 is 5000 pixels.

If the data of the other R and G sensors are to be delayed at the data input time of the B sensor, which is the last end of reading, R is delayed by 16 lines, so that a delay memory capacity of 80 KB is required. The delay of eight lines requires a delay memory capacity of 40 KB. Since B does not require a delay, a delay memory is not required.

FIG. 13 is a block diagram of a conventional area attribute discrimination process. The G data of the RGB data that has passed through the color correction processing unit 23 is input to the in-region attribute determination processing unit 22. This is because the Green signal is more sensitive than the other colors in terms of resolution. The data of the G signal is thinned out by the thinning processing means 32. This is to increase the size of the area determined with a small amount of data.
Now, it is assumed that the thinning process is performed every other pixel in both the main scanning and the sub-scanning. Therefore, after the thinning-out processing, 300 dpi
become.

Next, a two-dimensional window is formed by the MTF processing window forming means 33 and the MTF window memory 28. Here, it is assumed that the MTF is improved by the MTF processing means 34 with a window size of 3 × 3. Based on the data after the MTF processing, an M × N window is formed by the in-region attribute discrimination window forming processing means 35 and the in-region attribute discrimination window memory 29, and the in-region attribute discrimination processing means 36 forms characters / photos / halftone dots. Attributes are determined.

FIG. 14 is a diagram showing image attributes in a conventional area. The determination window is 8 × 8 pixels. FIG. 14A shows the case of a character, which is MT
The edge component of the F-processed data is analyzed, and it can be determined that the character is a character by connecting the edges linearly. FIG.
(B) shows the case of a photograph, in which the edge component of the data subjected to the MTF processing is analyzed, and it can be determined that the photograph is a photograph because there is no edge. FIG. 14C shows a case of a halftone dot. The edge component of the data subjected to the MTF processing is analyzed, and it can be determined that the halftone dot exists because the edge exists in a granular form.

By the way, in the in-region attribute discriminating processing means 22, the MTF processing window forming processing means 33 and the in-region attribute discriminating window forming processing means 35 delay eight lines in units of 300 dpi in the sub-scanning direction. 6
When converted to 00 dpi units, the delay is 16 lines. In order to match the RGB read data to the discrimination delay, the RGB
The read data itself is transferred to the data delay processing unit 3 of FIG.
7 and the delay memory 38. The required delay memory capacity is as large as 240 KB.

FIG. 15 is a block diagram of a conventional filtering process.
FIG. 16 is a conventional circuit block diagram. Finally, an internal block of the filter processing unit 24 will be described. Based on the RGB data delayed by the data delay processing unit 37, a window is formed by the filter window formation processing unit 39 and the filter window memory 30. Based on this window, the filter processing means 40
Is filtered. According to the attribute discrimination result of the in-region attribute discrimination processing means 22, a high-pass filter shown in FIG. 6A is applied to a character, a weak high-pass filter shown in FIG. In the case (1), a low-pass filter shown in (c) is applied.

[0021]

However, in the conventional image reading apparatus, it is necessary to delay the read data for the attribute discrimination processing within the area, so that a large-capacity delay memory is required.

Accordingly, an object of the present invention is to provide an image reading apparatus capable of reducing the capacity of a delay memory.

[0023]

SUMMARY OF THE INVENTION The present invention provides a means for reading color image data by a three-line image sensor,
Line correction processing means for correcting the distance between lines of the three-line image sensor, and the read image is divided into M × N regions, and image attributes such as characters / photos / halftone dots are determined in each of the divided regions. The image processing apparatus includes an in-region attribute determining unit and a filtering unit that applies a two-dimensional filter to the read image. With this configuration, an image reading device that can reduce the capacity of the delay memory can be realized.

[0024]

According to the first aspect of the present invention, there is provided a means for reading color image data by a three-line image sensor, a line correction processing means for correcting a distance between lines of the three-line image sensor, and a method for reading a read image by M × An area attribute discriminating unit for discriminating image attributes such as characters, photographs, and halftone dots in each of the divided regions; and a filter unit for applying a two-dimensional filter to the read image. ing. With this configuration, an appropriate filter process can be performed on the read image, and the read image quality can be improved.

According to the second aspect of the present invention, the line correction processing means and the in-area attribute discriminating means input the input image signal at the same time and perform parallel processing on the circuit configuration. As a result, it is possible to reduce the delay associated with the in-region attribute determination by the number of delay lines required for line correction.

According to the third aspect of the present invention, the type of the filter of the filter means and the coefficient of the filter are controlled in accordance with the result of the determination by the attribute determination means in the area. As a result, it is possible to perform an appropriate filtering process on the read image, and to improve the read image quality.

According to the fourth aspect of the present invention, the filter means is capable of selecting a filter type and a filter coefficient set from outside, independently of the result of the determination by the in-region attribute determining means. This makes it possible to sharpen or soften the entire read image as intended by the user.

According to the fifth aspect of the present invention, the head reading color of the three-line image sensor is made to match the color of the input image signal of the area attribute discriminating means. It is possible to reduce the delay due to the attribute determination in the area by the number of sensor lines where the maximum delay occurs among the delays required for the line correction.

According to the present invention, the delay memory used in the line correction processing means, the MTF window memory used in the area attribute discriminating means, and the window of the M × N area used in the area attribute discriminating means. The memory is also used. This makes it possible to reduce the delay associated with the intra-region attribute discrimination by the number of lines required for the line correction, and to use both the MTF window memory inside the intra-region attribute discrimination and the intra-region attribute discrimination window memory. Therefore, there is no need to provide a new one.

Hereinafter, the first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a configuration diagram of an image reading device according to a first embodiment of the present invention, FIG. 2 is a block diagram of an electric circuit according to the first embodiment of the present invention, and FIG. 3 is an explanatory diagram of the same data processing points. Here, in FIG. 1 and FIG. 2, the same components as those in FIG. 9 and FIG.
The description is omitted by attaching the same reference numerals.

Next, only differences from the prior art will be described. As shown in FIG. 2, RGB after shading correction
The data is input to the line correction processing means 21 on the one hand,
On the other hand, only the G signal is used for the in-area attribute determination
To enter. The positional relationship of the data at this time is shown in FIG. 3 (an explanatory diagram of the data processing point according to the first embodiment of the present invention). In FIG. 3, the mesh of the thin line represents a pixel expressed at a resolution of 600 dpi.

The mesh of a thick frame expressed in units of 2 × 2 pixels on this mesh is processed by the attribute determination processing means 22 in the area.
Means that 300 dpi has been achieved by the thinning-out processing in the pre-processing. The window size of 8 × 8 pixels in units of 300 dpi indicates an 8 × 8 area for performing image attribute determination. Since the image attribute determination is made based on the G signal, the in-region attribute determination processing means 22 outputs
When the signal is input, the attribute determination is performed.
Now, assuming that the determination is completed when one line in the sub-scanning direction and two pixels in the main scanning direction have passed from the 8 × 8 area.
The determination is completed when the data of the points R, G, and B shown in FIG.

FIG. 4 is a circuit block diagram according to the first embodiment of the present invention. FIG. 4 shows a detailed diagram of the line correction processing unit 21, the in-region attribute determination processing unit 22, the color correction processing unit 23, and the filter processing unit 24. In FIG.
Reference numeral 28 denotes an MTF window memory for performing MTF processing. This is the same size as the conventional example. Reference numeral 29 denotes a window memory for performing in-region attribute determination processing. This is also the same size as the conventional example.

Here, as described above with reference to FIG. 3, after the result of the attribute determination in the area is obtained, in order to align the result with the output data of the line correction, the R signal shown in FIG. The line delay memory 27a requires 24 lines because the line interval between the point of the current input image (the point indicated by R in FIG. 3) and the top line of the thick frame mesh is required. This results in a capacity of 120 KB. 1
20 KB is a value larger than 80 KB shown in the conventional example.

Line delay memory 27b for G signal
Is necessary for the line interval between the point of the current input image (the point indicated by G in FIG. 3) and the top line of the thick frame mesh, so that 16 lines are required. This is 80KB
Is the capacity. 80 KB is a value larger than 40 KB shown in the conventional example. Line delay memory 2 for B signal
7c is necessary for the line interval between the point of the current input image (the point indicated by B in FIG. 3) and the top line of the thick frame mesh, so 8 lines are required. This is 40K
B. 40 KB is a value larger than 0 KB shown in the conventional example. Thus, the line delay memory 27
Although a, 27b, and 27c are larger by 40 KB each for the RGB than the conventional example, since the data delay memory 38 (240 KB) shown in FIG. 13 is not required, the memory is totally reduced.

Finally, the filter processing means 24 applies a high-pass filter shown in FIG. 3A for characters and a weak high-pass filter shown in FIG. Is applied, and in the case of a halftone dot, a low-pass filter shown in (c) is applied. As described above, an appropriate filtering process can be performed on the read image, and the read image quality can be improved.

Further, the type of filter and the coefficient of the filter set from outside can be selected independently of the determination result of the in-region attribute determining means 22, so that the whole read image can be sharpened as intended by the user. It can be made soft or soft.

Hereinafter, a second embodiment of the present invention will be described with reference to FIGS. Here, FIG. 5 is a schematic diagram of a three-line image sensor according to the second embodiment of the present invention, FIG. 6 is an explanatory diagram of data processing points according to the second embodiment of the present invention,
FIG. 7 is a circuit block diagram of Embodiment 2 of the present invention.
FIG. 5 shows a configuration of a three-line image sensor 17 characteristic of the second embodiment. FIG. 11 shown in the conventional embodiment.
What is different is the order of the color filters of the line sensor. In the second embodiment, the color to be read at the head is
The color is matched with the color used for the attribute determination processing means 22 in the area. Specifically, the three-line image sensor 1
The head reading color of No. 7 is G. FIG. 6 shows the positional relationship of data according to the configuration of the three-line image sensor shown in FIG.

In FIG. 6, the mesh of the thin line is 600d.
Represents a pixel expressed at pi resolution. The thick frame mesh expressed in units of 2 × 2 pixels on top of this mesh indicates that it has been converted to 300 dpi by the thinning process which is the preprocessing of the intra-region attribute determination processing means 22. This 300d
A window size of 8 × 8 pixels in pi units indicates an 8 × 8 area in which image attribute determination is performed. If a G signal is input to the region attribute determination processing means 22 beyond this region, attribute determination has been performed. Now, assuming that the determination is completed when one line in the sub-scanning direction and two pixels in the main scanning direction have passed from the 8 × 8 region, R, G, and R shown in FIG.
The determination is completed when the data of the point B is input to the line correction processing means 21.

FIG. 7 is a detailed diagram of the line correction processing means 21, the in-area attribute determination processing means 22, the color correction processing means 23, and the filter processing means 24. Reference numeral 28 denotes a window memory for performing MTF processing. This is the same size as the conventional example. Reference numeral 29 denotes a window memory for performing in-region attribute determination processing. This is also the same size as the conventional example.

Here, as described with reference to FIG. 6, after the result of the attribute determination in the area is obtained, in order to align the result with the output data of the line correction, the G signal shown in FIG. The line delay memory 27c needs 16 points because the line interval between the point of the current input image (the point indicated by G in FIG. 6) and the top line of the thick mesh is required. This results in a capacity of 80 KB. 80
KB is the same value as 80 KB (the R signal delay memory in the conventional example) shown in the conventional example. The line delay memory 27b for the B signal stores the point of the current input image (in FIG. 6,
(Point indicated by B) and the line interval between the top line of the thick frame mesh and eight lines are required. This results in a capacity of 40 KB. 40 KB is the same value as the conventional example. In the line delay memory 27a for the R signal, the point of the current input image (the point indicated by R in FIG. 6) coincides with the top line of the thick frame mesh, so that the delay memory becomes unnecessary. Is the same as As described above, the delay memory for line correction has the same size as the conventional example, but has the data delay memory 38 shown in FIG.
(240 KB) is not required, as compared with the first embodiment.
Further, the memory can be reduced.

Finally, as described in the first embodiment, the filter processing unit 24 performs a high-pass filter shown in FIG. 7A in the case of a character in accordance with the attribute determination result of the in-region attribute determination processing unit 22, In this case, a weak high-pass filter shown in (b) is applied, and in the case of a halftone dot, a low-pass filter shown in (c) is applied. As described above, an appropriate filtering process can be performed on the read image, and the read image quality can be improved. In addition, the in-area attribute determining means 22
Irrespective of the judgment result of, the filter type and filter coefficient set from outside can be selected,
It is also possible to sharpen or soften the whole read image as intended by the user.

Hereinafter, a third embodiment of the present invention will be described with reference to FIG. Here, FIG. 8 is a circuit block diagram of Embodiment 3 of the present invention. The difference from FIG. 7 used in the description of the second embodiment is that an input signal to the in-region attribute determination processing means 22 is obtained from the data bus of the G delay memory. That is, the window memory 28 for performing the MTF processing and the in-area attribute determination processing means 2
Memory 29 for line 2 and line correction means 2
1, the same delay memory 27c is shared by the same memory.

By doing so, the MTF window memory 28 (10 KB) and the attribute discrimination window memory 29 (32 KB) can be further reduced. However, as shown in FIG. 8, when the G delay memory is also used, access to the memory becomes frequent, so that the image reading speed is limited. That is, in an environment where the image reading speed does not matter, a configuration with the least memory can be provided.

Finally, as described in the first and second embodiments, the filter processing means 24 applies the high-pass filter shown in FIG. In the case of a photograph, a weak high-pass filter shown in (b) is applied, and in the case of a halftone dot, a low-pass filter shown in (c) is applied. As described above, an appropriate filtering process can be performed on the read image, and the read image quality can be improved. In addition, attribute determination means 2 in the area
Regardless of the determination result of 2, the externally set filter type and filter coefficient can be selected from the outside, so that the entire scanned image can be sharpened or softened as intended by the user. Become.

[0046]

According to the present invention, the capacity of the delay memory can be reduced and the read image quality can be improved.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image reading device according to a first embodiment of the present invention;

FIG. 2 is an electric circuit block diagram according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram of a data processing point according to the first embodiment of the present invention.

FIG. 4 is a circuit block diagram according to the first embodiment of the present invention.

FIG. 5 is a schematic diagram of a three-line image sensor according to a second embodiment of the present invention.

FIG. 6 is an explanatory diagram of a data processing point according to the second embodiment of the present invention.

FIG. 7 is a circuit block diagram according to a second embodiment of the present invention.

FIG. 8 is a circuit block diagram according to a third embodiment of the present invention.

FIG. 9 is a configuration diagram of a conventional image reading apparatus.

FIG. 10 is a block diagram of a conventional electric circuit.

FIG. 11 is a block diagram of a conventional electric circuit.

FIG. 12 is a block diagram of a conventional line correction process.

FIG. 13 is a block diagram of a conventional attribute discrimination process in a region.

14A is a diagram illustrating an image attribute in a conventional region. FIG. 14B is a diagram illustrating an image attribute in a conventional region. FIG. 14C is a diagram illustrating an image attribute in a conventional region.

FIG. 15 is a block diagram of a conventional filter processing.

FIG. 16 is a conventional circuit block diagram.

[Explanation of symbols]

 17 three-line image sensor 21 line correction processing means 22 area attribute discrimination processing means 23 color correction processing means 24 filter processing means

Claims (6)

[Claims] 1. A means for reading color image data by a three-line image sensor, a line correction processing means for correcting a distance between lines of the three-line image sensor, and converting the read image into M × N (M and N are natural numbers) Divided into regions,
The image processing apparatus further includes: an in-region image attribute determining unit that determines an attribute of an image such as a character / photo / halftone dot in each of the divided regions; and a filter unit that applies a two-dimensional filter to the read image. Image reading device. 2. The image reading apparatus according to claim 1, wherein said line correction processing means and said in-region image attribute discriminating means input an input image signal at the same time and perform parallel processing on a circuit configuration. apparatus. 3. An image reading apparatus according to claim 1, wherein a type of a filter of said filter means and a coefficient of the filter are controlled in accordance with a result of determination by said in-region image attribute determining means. 4. The apparatus according to claim 1, wherein said filter means is capable of selecting a filter type and a filter coefficient set externally, irrespective of the determination result of said in-region image attribute determining means. An image reading device according to claim 1. 5. The image reading apparatus according to claim 1, wherein a head reading color of said three-line image sensor is matched with a color of an input image signal of said in-region image attribute discriminating means. 6. A delay memory used for said line correction processing means, and an MTF used for said in-region image attribute discriminating means.
2. The image reading apparatus according to claim 1, wherein a window memory and a window memory of the M × N area used by the in-area image attribute discriminating means are shared.