JP2006229728A - Image reading device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image reading apparatus capable of performing a correction process optimal for a joint correction process.
When image information is recognized and the image is a character image, an arbitrary weighting coefficient 210 to 213 is selected and multiplied in the overlap portion, and the addition processing is performed as it is. On the other hand, when the image is a halftone dot, a weighting coefficient that is not subjected to correction processing is selected and multiplied in the overlap portion, and when there is a difference in the MTF characteristics in the individual data, the MTF characteristics The data on the better side is subjected to a smoothing filter process and an addition process is performed to obtain corrected data.
[Selection] Figure 2

Description

  The present invention relates to joint correction processing of an image reading apparatus that reads a document image by dividing a plurality of image sensors in a staggered manner in a copying machine, a scanner, or the like that performs digital image processing.

In an image reading apparatus that reads a document image, in order to guarantee reading characteristics, shading correction, which is a generally known technique, is performed to correct variations in main scanning direction data. In an image reading apparatus in which a plurality of image sensors are arranged in a staggered manner, the output characteristics of each image sensor vary, and it is difficult to obtain a uniform density output over the entire surface. Further, the upstream side image sensor arranged in a zigzag pattern (the side that reads the original document first) temporarily stores the image data in a memory, and matches the data line of the downstream side image sensor that reads the original document later. It is adjusted. However, line deviation in the sub-scanning direction may occur in the data of the downstream image sensor and the upstream image sensor due to uneven feeding during document conveyance.
In addition, when a halftone image is read due to variations in MTF characteristics of individual image sensors or degradation of the MTF characteristics due to defocusing of the image sensor due to document flutter during document transport unique to a sheet-through type document transport device. There is a problem in that it appears as a density difference at the joint between adjacent image sensors.
In order to solve such problems, techniques described in the following patent documents have been proposed.
JP 2003-46736 A

  However, the technique disclosed in Patent Document 1 is good for images such as characters and lines and images having a uniform halftone density. However, in the case of a halftone image composed of halftone dots, image information is used. It is not a combined correction. For this reason, when the lines are deviated by several lines in the sub-scanning direction, the peak density of the halftone dots at the joint portion is not preserved, and the density may be lowered or blurred.

Accordingly, the first object of the present invention is to reduce the deviation in the sub-scanning direction and the density variation between the sensors in the character and halftone density when performing the joint correction process, and the density of the halftone dot part. An object of the present invention is to provide an image reading apparatus capable of preventing a decrease or fading and further mitigating when there is an MTF difference between image sensors.
A second object of the present invention is to provide an image reading apparatus capable of performing correction processing optimal for joint correction processing.
A third object of the present invention is to provide an image reading apparatus capable of performing joint correction processing in which a weighting coefficient suitable for image information can be selected.
A fourth object of the present invention is to associate a halftone dot detection result with a weighting coefficient, so that when the halftone dot detection result is a halftone dot, a coefficient that is not subjected to correction processing (no correction) is selected as the weighting coefficient. It is another object of the present invention to provide an image reading apparatus capable of performing a seam correction process for preventing density reduction and fading of a seam halftone dot.
The fifth object of the present invention is that when the halftone dot detection result is a halftone dot, the MTF characteristic of each adjacent image sensor and the MTF difference due to the document floating are taken into consideration, and the MTF characteristic is good. It is an object of the present invention to provide an image reading apparatus capable of performing joint correction processing so that the MTF difference at the joint halftone dot portion can be reduced by performing smoothing filter processing on the side image data.

  According to the first aspect of the present invention, in the image reading apparatus, a plurality of image sensors are arranged in a staggered manner, and reading portions of adjacent image sensors are arranged so as to overlap each other in the main scanning direction by a predetermined number of pixels. Joint correction means for acquiring image data of the overlap portion from each image sensor and correcting the image data of the joint portion of each image sensor; and whether the image data of the overlap portion is a halftone image or a non-halftone image Halftone dot detection means for recognizing whether the image is a dot image, and MTF comparison means for comparing the MTF characteristics of the image areas of the overlapping portions, wherein the joint correction means is determined by the halftone dot detection means. The first object is achieved by performing joint correction processing according to the comparison result of the MTF comparison means.

According to a second aspect of the present invention, in the first aspect of the present invention, the joint correction means multiplies the weighting coefficient multiplying means for multiplying both image data from each adjacent image sensor by a weighting coefficient corresponding to the main scanning direction. And a filter processing means for performing a smoothing filter correction process on both image data from each adjacent image sensor, and an adding means for adding both image data from each adjacent image sensor. Achieve the purpose of 2.
According to a third aspect of the present invention, in the second aspect of the present invention, the weighting coefficient multiplication means has a plurality of weighting coefficients to achieve the third object.

According to a fourth aspect of the invention, in the third aspect of the invention, the weighting coefficient multiplying means achieves the fourth object by selecting a weighting coefficient based on a determination result of the halftone dot detecting means. To do.
According to a fifth aspect of the present invention, in the second aspect of the present invention, the filter processing means determines that the determination result of the halftone dot detection means is a halftone dot, and determines the filter based on the comparison result of the MTF comparison means. The fifth object is achieved by selectively outputting whether to output the processed data or the through data.

  According to the first aspect of the present invention, the image recognition (halftone dot detection) and the MTF comparison are performed in the overlap portion when the correction processing is performed on the joint portion. And the density variation between the sensors can be reduced, the density of the halftone image can be prevented from being lowered or blurred, and further, when there is an MTF difference between the image sensors, it can be reduced. Can perform joint correction processing suitable for the image information.

According to the second aspect of the present invention, since the processing method is optimal for the joint correction processing, good correction data can be obtained.
According to the third aspect of the invention, since a plurality of weighting coefficients for joint correction processing are prepared, correction processing can be performed in association with the halftone detection result.

According to the fourth aspect of the present invention, since the weighting coefficient is selected based on the determination result of the halftone dot detecting means, the correction process can be performed with the weighting coefficient suitable for the document image.
According to the fifth aspect of the present invention, the output of the filtered data or the output of the through data is selectively output based on the comparison result of the MTF comparing means, so that the joint is a halftone original. If there is a difference in MTF characteristics, the difference can be relaxed.

Hereinafter, a preferred embodiment of the present invention will be described in detail with reference to FIGS.
FIG. 1 is a block diagram showing an outline of an image reading apparatus according to an embodiment of the present invention. The image reading apparatus 100 includes CSI (image sensors) 1 to 5, A / D (analog / digital) converters 101 to 105, memories 111 to 114, a line composition processing block 120, and a memory 130.
In the image reading apparatus 100, the five image sensors CIS1 to CIS5 are arranged so as to overlap each other and read in the main scanning direction. In addition, the sensors CIS1, CIS3, and CIS5 are arranged in a staggered manner with intervals at the upstream side in the sub-scanning direction and the sensors CIS2 and CIS4 at the downstream side.

Image data output from each of the image sensors CIS1 to CIS5 is converted into digital signals by A / D (analog / digital) converters 101 to 105, and the data of CIS1 and CIS3 to CIS5 are temporarily set for delay adjustment in the sub-scanning direction. Accumulated in the memories 111-114. Since CIS2 is on the most downstream side, data is transferred without being stored in the memory. On the other hand, the CIS 4 is on the same line as the CIS 2 but is stored in the memory 104 because it is arranged several lines upstream for easy adjustment.
The desired line delayed data for each CIS 1-5 is sent to the line synthesis processing block 120. Here, the correction processing of the joint portion and the one-line processing are performed on the data of each of the CISs 1 to 5 flowing in parallel and are passed to the subsequent processing.

Next, the line composition processing block will be described with reference to FIG.
As shown in FIG. 2, the line synthesis processing block 120 includes a joint correction processing circuit 201, a halftone detection circuit 202, and an MTF comparison circuit 203.
The image data of the overlapping portions of the CSI 1 to 5 (image sensors) are simultaneously input to the joint correction processing circuit 201, and the weighting coefficients 210 to 213 determined by the result A from the halftone dot detection circuit 202 are set and the multiplication circuit. Multiplied by 220 · 221. When the result A from the halftone dot detection circuit 202 is a character (0), the weighting coefficients 210 and 211 are selected.
If the result A from the halftone dot detection circuit 202 is a halftone character (1), the weighting coefficients 212 and 213 are selected. Details of the weighting factors 210 to 213 will be described later.

The post-multiplication data is sent to selectors 240 and 241 after smoothing filter processing 230 and 231 and through processing. The selectors 240 and 241 select and output whether to output through or based on the result A from the halftone dot detection circuit 202 and the result B from the MTF comparison circuit 203.
For example, when the halftone detection result A is a character (0), the selection signals of the selectors 240 and 241 always output 0 from the AND circuits 250 and 251. Through output.
When the halftone detection result A is a halftone dot (1), the selection signal of the selectors 240 and 241 is selected from the output of the AND circuits 250 and 251 or the output from the signal from the MTF comparison circuit 203.

When the MTF characteristic is D1 = D2, since the outputs of the AND circuits 250 and 251 are 0, through data is output for both the data from D1 and the data from D2.
When D1> D2, the output from the AND circuit 250 is 1, so the data from D1 is the data after the smoothing filter processing 230, and the output from the AND circuit 251 is 0, so the data from D2 is Through data is output. When D1 <D2, the output of the AND circuit 250 is 0, so the data from D1 is output as through data, and the output of the AND circuit 251 is 1, so the data from D2 is the data after the smoothing filter processing 231. Data will be output.

The outputs from the selectors 240 and 241 are subjected to addition processing by the addition circuit 260 and are output as corrected data.
Summarizing what has been described above, when the image information is recognized and the image is a character image, an arbitrary weighting factor 210 to 213 is selected and multiplied in the overlap portion, and the addition process is performed as it is. On the other hand, when the image is a halftone dot, a weighting coefficient is selected and multiplied so that correction processing is not performed on the overlapped portion. If there is a difference in the MTF characteristics in the individual data, the MTF characteristic Smooth data is subjected to smoothing filter processing on the good side data, and addition processing is performed to obtain corrected data.

Next, a specific example when the determination result from the halftone dot detection circuit 202 is a character will be described.
In FIG. 3, description will be given by taking the joint of CIS1 and CIS2 as an example. The same processing is performed at the joint of CIS2 and CIS3, CIS3 and CIS4, and CIS4 and CIS5.
If the determination result from the halftone dot detection circuit 202 is a character, the left and right overlapping portions to be joined are provided with 128 pixels, and the data from the CIS 1 in units of 16 pixels is 7/8, 6/8 in order from the left side. ,... 1/8 weighting coefficient (210 in FIG. 2) is selected and processed by the multiplication circuit.
In the data from CIS2, on the contrary, 7/8, 6/8,..., 1/8 weighting coefficient (211 in FIG. 2) is selected from the right side of the overlapping portion, and is processed by the multiplication circuit. .

  Since the determination result from the halftone dot detection result is a character, the data after the smoothing filter processing is not selected, and the density after the weighting calculation for each of CIS 1 and 2 is directly input to the addition circuit and added, and the overlap portion of the joint Image density. With this method, the image density at the time of reading can be stored even after the joint correction processing.

  Further, FIG. 4 shows the image density after the joint correction process when there is a density difference between CIS1 and CIS2. As shown in FIG. 4, even when there is a density difference, it is possible to absorb the density step by gradually connecting the density of the overlap portion in a stepwise manner.

Next, a specific example when the determination result from the halftone dot detection circuit 202 is a halftone dot will be described.
As shown in FIG. 5, in the data from CIS1, a weighting coefficient of 1 (210 in FIG. 2) is set from the left side of the overlap portion to the overlap center. On the right side from the overlap center, a weighting coefficient of 0 (212 in FIG. 2) is set and is calculated by the multiplication circuit.
On the other hand, in the data from CIS2, conversely, a weighting factor of 1 (211 in FIG. 2) is set from the right side of the overlap portion to the overlap center, and a weighting factor of 0 (in FIG. 2) from the overlap center to the left side. 213) is set and calculated by the multiplication circuit.

Further, if there is no MTF difference between the data from CIS1 and CIS2 based on the information from the MTF comparison circuit 203, the addition process is performed as it is. When the MTF characteristic of the data from CIS1 is good, the data on the CIS1 side is smoothed and added to the data of CIS2. That is, they are joined together.
When the MTF characteristic of the data from CIS2 is good, the data on the CIS2 side is smoothed and added to the data of CIS1.
FIG. 6 shows a specific example when the MTF characteristic of the CIS1 data is good.

Next, FIG. 7 shows another example when a plurality of weighting coefficients are used for the joint correction processing.
In CIS1, weighting of 15/16, 14/16,... 12/16 is applied sequentially from the left side of the overlap portion to the overlap center. Then, weighting of 4/16, 3/16... 1/16 is applied to the right side from the overlap center. On the other hand, in CIS2, on the other hand, weights of 15/16, 14/16,... 12/16 are applied from the right side of the overlapping part to the overlap center, and the left side from the overlap center is 4/16. 3/16 ... 1/16 weighting is applied.

Then, the density after the weighting calculation is added to each of the CIS 1 and 2 to obtain the image density of the joint overlap portion. Even if the weighting coefficient is changed, the image density at the time of reading can be stored after the joint density correction processing.
In this weighting coefficient, the CIS1 image density data is weighted on the left side (CIS1 side) of the overlap portion, and the CIS2 image density data is weighted on the right side (CIS2 side) of the overlap center. Yes.

  Further, the halftone dot detection result is divided into three stages, “(1) character (non-halftone dot) (2) intermediate region (3) halftone dot”, and in the case of (1), the weighting coefficient shown in FIG. Is corrected using the weighting coefficient shown in FIG. 7 in the case of (2), and the weighting coefficient shown in FIG. 5 is used in the case of (3). It is also possible not to perform correction processing.

  In this embodiment, three types of weighting coefficients are shown. In this example, the weighting coefficients are used in units of 1/8 and 1/16. However, the weighting coefficients may be further subdivided. Further, the number of divisions of the overlap portion is 8 blocks in the present embodiment, but this can be further subdivided.

Next, a specific example of halftone dot detection processing will be described. This example shows one example of halftone dot detection processing, and various methods other than this method are conceivable, and the present invention is not limited by these methods.
FIG. 8 shows a detailed view of halftone dot detection processing. In the halftone dot detection process, the image data of the joint is input, and the filter process 800 is first performed. An example of this filter coefficient is shown in FIG. Next, a simple binarization process 801 is performed on the image after the filter process 800. Furthermore, the number of change points (white → black (0 → 1) or black → white (1 → 0)) of the image in the region of interest in the simple binarized image data is greater than or equal to a certain value (arbitrary threshold). If there is, it is determined as a halftone dot, and if it is less than a certain value, it is determined as a non-halftone dot. FIG. 10 illustrates this determination method. In this example, there are two determination results.

Further, in the determination result, two threshold values of change points are provided (Th1 and Th2), (1) less than Th1 (non-halftone dot), (2) more than Th1 and less than Th2 (intermediate region), and (3) more than Th2 ( Three types of determination result outputs can be prepared as halftone dots). Furthermore, a plurality of change point threshold settings may be provided to output a plurality of determination results. In short, the number of lines of halftone dots may be determined.
In this case, the weighting coefficient of the joint correction process described above is changed according to the halftone dot detection result. This process is summarized in the table shown in FIG.
The range for performing halftone dot detection may be the halftone dot detection range for the overlap region, and data on either the CIS upstream side or the CIS downstream side may be used. Alternatively, halftone dot detection may be performed on data on both the CIS upstream side and the CIS downstream side, and the respective detection results may be combined to obtain a final detection result. For example, when there are two halftone dot detection results (halftone dot or non-halftone dot), the final output is a halftone dot if both are determined to be halftone dots by taking the logical product, and the other is a non-halftone dot And

Next, a specific example of the MTF comparison process will be described. Here, an example of the MTF comparison processing is shown, and various methods other than the present method are possible, and the present invention is not restricted by this method.
FIG. 12 shows a detailed view of the MTF comparison processing circuit 203. The MTF comparison processing circuit 203 inputs the image data of the joint portion, and determines the maximum value and the minimum value of the 128 pixel data of the overlap portion on both the upstream side (D1) and the downstream side (D2) of the CIS. 1101, 1104) and a minimum value detection circuit (1100, 1103).

Further, the difference between the maximum value and the minimum value (Δd1 and Δd2) is obtained by the difference calculation circuit (1102, 1105), and the magnitude of the difference is judged by the comparison judgment circuit 1106 to judge the magnitude of the MTF. .
For example, if Δd1> Δd2, it is determined that the data from D1 is better in MTF, and 1 is output. If Δd1 <Δd2, it is determined that the data from D2 has a better MTF, and 2 is output. If Δd1 = Δd2, the MTF is determined to be the same, and 0 is output as the output.
Further, the difference amount between Δd1 and Δd2 may be output. For example, when | Δd1−Δd2 | <= Th10 (arbitrary), 0 is output as the difference amount. When Th20 (arbitrary)> = | Δd1−Δd2 |> Th10 (arbitrary), 1 is output as the difference amount. When | Δd1−Δd2 |> Th20 (arbitrary), 2 is output as the difference amount. Since this is a measure of how much the difference in MTF between adjacent image sensors is, it is possible to change the coefficient of the smoothing filter using this parameter.

1 is a diagram schematically illustrating an image reading apparatus according to an embodiment of the present invention. 1 is a configuration diagram illustrating details of an image reading apparatus according to an embodiment of the present invention. It is the figure which showed the example of the joint of CIS1 and CIS2. It is the figure which showed the image density after a joint correction process in case there is a density difference between CIS1 and CIS2. It is the figure which showed the multiplication circuit of the overlap part in the data from CIS1. It is the figure which showed the specific example in case the MTF characteristic of the data of CIS1 is good. It is the figure which showed the other example at the time of making the coefficient of the weight of a joint correction process into multiple. It is the figure which showed the detail of the halftone detection process. It is the figure which showed one example of the filter coefficient. It is the figure which showed the range which performs a halftone dot detection. It is the figure which showed the relationship between a halftone dot detection result and a joint correction process in a present Example. It is the figure which showed the detail of the MTF comparison process.

Explanation of symbols

1-5 CSI (image sensor)
DESCRIPTION OF SYMBOLS 100 Image reader 101-105 A / D converter 111-114 Memory 120 Line composition processing block 130 Memory 201 Joint correction processing circuit 202 Halftone detection circuit 203 MTF comparison circuit 210-213 Weighting coefficient 220, 221 Multiplication circuit 230, 231 Smoothing filter processing 240, 241 Selector 250, 251 AND circuit 260 Adder circuit

Claims (5)

In an image reading apparatus in which a plurality of image sensors are arranged in a staggered manner, and reading portions of adjacent image sensors are arranged so as to overlap in the main scanning direction by a predetermined number of pixels,
A joint correction unit that obtains image data of an overlap portion from each image sensor and corrects the image data of a joint portion of each image sensor;
Halftone dot detection means for recognizing whether the image data of the overlap portion is a halftone dot image or a non-halftone dot image, and MTF comparison means for comparing the MTF characteristics of the image areas of the overlap portions,
The image reading apparatus, wherein the joint correction unit performs a joint correction process according to a determination result of the halftone dot detection unit and a comparison result of the MTF comparison unit. The joint correction means includes weighting coefficient multiplication means for multiplying both image data from each adjacent image sensor by a weighting coefficient corresponding to the main scanning direction;
Filter processing means for performing smoothing filter correction processing on both image data from each adjacent image sensor;
The image reading apparatus according to claim 1, further comprising an adding unit that adds both of the image data from each adjacent image sensor. The image reading apparatus according to claim 2, wherein the weighting coefficient multiplying unit has a plurality of weighting coefficients. The image reading apparatus according to claim 3, wherein the weighting coefficient multiplying unit selects a weighting coefficient based on a determination result of the halftone dot detecting unit. The filter processing unit determines whether the determination result of the halftone dot detection unit is a halftone dot, and whether to output the filtered data or the through data according to the comparison result of the MTF comparison unit The image reading apparatus according to claim 2, wherein the image reading apparatus outputs the image.

Images