JP2004166151A - Method and device for producing image data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent quality deterioration of an image produced by combining a plurality of image portions sharing a common region in a method, and to provide device for producing image data. <P>SOLUTION: In the method and device for producing image data, a line-detecting means 20 detects light reflected at a linear region S on an original 30 upon irradiation of a light from a linear light source 62, by moving the original 30 in the sub-scanning Y direction with respect to the line-detecting means 20 which is provided with line sensors 10A and 10B, comprising a plurality of light-receiving sections disposed linearly, such that the light-receiving sections Ak and B4 located at the end portions 11A and 11B of the respective line sensors detect the light emitted from the same position P located on the original 30 in duplicate, and one piece of image data is obtained by giving a weighting processing to the image data obtained from the light-receiving sections Ak and B4 which detect the light emitted from the same position in duplicate, and the image data which show overall image information the original 30 is bearing is produced by adding noise components which are reduced by the weighting processing. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image data creation method and apparatus, and more particularly to an image data creation method and apparatus for creating image data representing one image by combining a plurality of image data groups each representing a plurality of image portions sharing the same region. Is.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, there is known an apparatus that reads a document by moving a line detection unit including a line sensor having a large number of light receiving units arranged in a line in the main scanning direction on the document in the sub-scanning direction. A wide line detection unit with a long line sensor is used to read a large document with such an apparatus, but a seamless long line sensor is difficult to realize due to manufacturing reasons. A line detection unit is used in which a plurality of line sensors are arranged in the main scanning direction so that their end portions overlap each other, and function as a single long line sensor as a whole.
[0003]
When a large document is read by the line detecting means configured as described above, the same position on the large document is repeatedly detected by each light receiving unit located at the end of each line sensor. Image data (also referred to as an image data group) representing the image portion sharing the same position in the large document is acquired, and these image data groups are combined to create image data representing the entire large document. .
[0004]
As a method of creating image data representing the entire large-size original by combining image data groups representing each of a plurality of image portions constituting the large-size original, it is detected by a light receiving unit located at an end portion where each line sensor overlaps each other. A plurality of image data representing the same position on the large document is subjected to a weighted addition process for reducing the weight of the image data acquired from the light receiving unit located at the end of the line sensor, and the image data for each position. There is known a method of creating image data representing the entire large-format original using these image data (see, for example, Patent Document 1 and Patent Document 2). Here, the area represented by the image data subjected to the weighted addition process in the image representing the entire large document is a band-like area extending in the sub-scanning direction.
[0005]
[Patent Document 1]
Japanese Patent Laid-Open No. 2002-57860
[0006]
[Patent Document 2]
US Pat. No. 6,348,981
[0007]
[Problems to be solved by the invention]
However, in the method of performing the weighted addition process, a noise component included in the band-shaped image area subjected to the weighted addition process and a noise component included in an image area other than the band-shaped area not subjected to the weighted addition process Due to this difference, streaks appear along the belt-like image area, and there is a problem in that the quality of the entire image representing the large-size original is deteriorated.
[0008]
The streaks appearing along the band-shaped image region are reduced by averaging the noise components included in the image data by the weighted addition process, and as a result, the noise components included in the image data not subjected to the weighted addition process are weighted. A difference from the noise component included in the image data subjected to the addition processing appears as a difference in image texture.
[0009]
Note that this problem is not limited to the case of acquiring image data by moving a plurality of line sensors on a large document as described above, but a plurality of images representing the same area of each of a plurality of image portions sharing the same area. This is a common problem when image data representing one whole image is created by performing weighted addition processing on image data and combining them.
[0010]
The present invention has been made in view of the above circumstances, and provides an image data creation method and apparatus capable of suppressing deterioration in quality of an image created by combining a plurality of image portions sharing the same region. It is intended.
[0011]
[Means for Solving the Problems]
In the image data creation method of the present invention, when creating a plurality of image data groups each representing a plurality of image portions sharing the same area to create image data representing one whole image, the same area in the whole image As the image data of each position, for each position, a weighted addition process is performed in which the weight of the image data located closer to the end of the image portion is reduced with respect to a plurality of image data at that position belonging to each image data group. An image data creation method for obtaining one image data and creating image data representing an entire image using the image data, wherein the image data obtained by the weighted addition processing is reduced by the weighted addition processing. The present invention is characterized in that image data representing the entire image is created by compensating for the noise component.
[0012]
When the image data creation device of the present invention creates image data representing one whole image by combining a plurality of image data groups each representing a plurality of image portions sharing the same area, the same area in the whole image As the image data of each position, for each position, a weighted addition process is performed in which the weight of the image data located closer to the end of the image portion is reduced with respect to a plurality of image data at that position belonging to each image data group. An image data creation device for obtaining one image data and creating image data representing an entire image using the image data, wherein the image data obtained by the weighted addition process is reduced by the weighted addition process. The present invention is characterized in that image data representing the entire image is created by compensating for the noise component.
[0013]
Another image data creation method according to the present invention includes a plurality of line sensors having a large number of light receiving portions arranged in a line, and the light receiving portions located at the end portions of these line sensors are located at the same position of the image carrier. The image carrier is moved while moving the image carrier relatively in the sub-scanning direction intersecting the main scanning direction with respect to the line detecting means arranged in the main scanning direction so as to detect the light emitted from each other in an overlapping manner. An image data creation method for creating image data representing image information carried by an image carrier by detecting light emitted from a body by a line detection means, wherein the light is detected by the light receiving portions of the plurality of line sensors. As the image data at each position, the line data is obtained for the image data obtained from the light receiving units of the plurality of line sensors that detect the light emitted from the position for each position. The image data obtained from the light receiving unit located at the edge of the sensor is weighted and added to reduce the weight to obtain one image data, and the image data obtained by the weighted addition processing is obtained by the weighted addition processing. Image data representing the entire image information is created by using image data obtained by compensating for the reduced noise component.
[0014]
Another image data creation apparatus according to the present invention includes a plurality of line sensors having a large number of light receiving portions arranged in a line, and the light receiving portions located at the end portions of these line sensors are located at the same position of the image carrier. Line detection means arranged in the main scanning direction so as to detect light emitted from each other and the image detection body is moved relative to the line detection means in the sub-scanning direction intersecting the main scanning direction. Based on the image data obtained by detecting the light emitted from the image carrier with the line detector while moving the image carrier relative to the scanning unit and the line detector in the sub-scanning direction, An image data creation device comprising image data creation means for creating image data representing image information carried by a carrier, wherein the image data creation means comprises the plurality of line sensors. At each position where the light is detected by the light receiving unit, the image data obtained from each light receiving unit of the plurality of line sensors that have detected the light emitted from the position is positioned at the end of the line sensor. A weighted addition processing means for obtaining one image data by performing a weighted addition process for reducing the weight of the image data obtained from the light receiving unit, and reducing the weighted addition process for the image data obtained by the weighted addition process. Noise compensation means for compensating for the noise component, and adopting image data supplemented with the noise component as image data at each position where light is detected by the light receiving units of the plurality of line sensors, Image data representing the image information is created.
[0015]
The compensation means not only the case where noise components reduced by the weighted addition process are completely compensated, but also the case where a part of the noise components reduced by the weighted addition process is compensated.
[0016]
The light emitted from the image carrier is light reflected from the image carrier, light transmitted through the image carrier, light generated from the image carrier upon irradiation with excitation light, etc. It means light carrying image information.
[0017]
As the image carrier, for example, a paper or film having image information formed on the surface, or a radiation image conversion panel formed by stacking a stimulable phosphor layer on a substrate can be used. When a radiation image conversion panel is used as an image carrier, stimulated emission light generated from the stimulable phosphor layer by irradiation of excitation light, that is, a radiation image recorded as a latent image on this stimulable phosphor layer ( Light indicating (image information) is light emitted from the image carrier.
[0018]
【The invention's effect】
The image data creation method and apparatus according to the present invention, when creating a plurality of image data groups each representing a plurality of image portions sharing the same region to create image data representing one whole image, As the image data at each position in the same region, for each position, for the plurality of image data at that position belonging to each image data group, weighted addition in which the weight of the image data located closer to the end of the image portion is smaller When the image processing is performed to obtain one image data, and the image data representing the entire image is created using the image data, the noise component reduced by the weighted addition processing with respect to the image data obtained by the weighted addition processing. Since the image data representing the entire image is created by compensating for the noise component, the noise component that appears in the same area in the entire image Since the difference from the noise component appearing in the area outside the same area in the entire image can be reduced, and the difference in the texture of the image in each area can be reduced, the quality of the synthesized entire image can be reduced. The decrease can be suppressed. Here, for example, in the case where the same region is elongated, the occurrence of streaks appearing along the same region in the entire image can be prevented by the effect of reducing the difference in image texture in each region.
[0019]
Another image data creation method and apparatus according to the present invention includes a plurality of line sensors having a large number of light receiving portions arranged in a line, and the light receiving portions located at the end portions of these line sensors are the same on the image carrier. This image is detected while moving the image carrier relatively in the sub-scanning direction intersecting the main scanning direction with respect to the line detecting means arranged in the main scanning direction so as to detect the light emitted from the position overlapping each other. When the light emitted from the carrier is detected by the line detection means and image data representing the image information carried by the image carrier is generated, the light at the positions where the light is detected by the light receiving units of the plurality of line sensors is detected. As image data, for each position, the image data obtained from each light receiving part of a plurality of line sensors that have detected the light emitted from each position is positioned at the end of the line sensor. The image data obtained from the light receiving unit is subjected to a weighted addition process for reducing the weighting to obtain one image data, and the noise component reduced by the weighted addition process is applied to the image data obtained by the weighted addition process. Since the image data obtained by compensating is used to create image data representing the entire image information, the image data acquired from the light receiving unit that detects the light emitted from the same position is mutually overlapped. It is possible to reduce the difference between the noise component of the image area to be represented and the noise component of the image area represented by the image data acquired from a light receiving unit different from the light receiving unit that detects light emitted from the same position. Since the difference in the texture of the image in each region can be reduced, the deterioration of the quality of the entire image can be suppressed. Here, for example, due to the effect of reducing the difference in the texture of the image in each of the above areas, the occurrence of streaks appearing along the image area represented by the image data at the position where light is detected by the light receiving unit in the entire image. Can be prevented.
[0020]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a perspective view showing a schematic configuration of an image data creation device according to an embodiment of the present invention, and FIG. 2 is an enlarged perspective view showing an optical path of light reflected from a document and imaged on a light receiving unit of a line detection unit. is there.
[0021]
In the image data creation device 100 according to the embodiment of the present invention, line sensors 10A and 10B having a large number of light receiving units arranged in a line are positioned at end portions 11A and 11B of the line sensors 10A and 10B, respectively. Line detection in which the light receiving unit is arranged in the main scanning direction (in the direction of arrow X in the figure, hereinafter referred to as the main scanning X direction) so that the light emitted from the same position of the document 30 as the image carrier is mutually detected. A sub-scanning unit 40 that moves the document 30 in the sub-scanning direction (arrow Y direction in the figure, hereinafter referred to as the sub-scanning Y direction) that intersects the main scanning X direction with respect to the line detection unit 20, and the line Based on the image data acquired by detecting the light emitted from the document 30 by the line detection unit 20 while moving the document 30 in the sub-scanning Y direction with respect to the detection unit 20, the document 30 carries the document 30. And an image data creating means 50 for creating image data representing the image information that.
[0022]
The image data creation means 50 is a line sensor 10A, 10B in which light emitted from each position is redundantly detected for each position as image data at each position where the light is detected by the light receiving units of the line sensors 10A, 10B. The weight of the image data obtained from the light receiving units located at the ends of these line sensors is reduced with respect to the image data obtained from each of the light receiving units (hereinafter abbreviated and referred to as overlapping light receiving units). An addition processing unit 51 that performs addition processing to obtain one image data, and a noise compensation unit 52 that compensates for noise components reduced by the weighted addition processing on the image data obtained by the weighted addition processing, The noise component is compensated for as image data at each position where light is detected redundantly by the light receiving portions of the line sensors 10A and 10B. While adopting the image data, document 30 adopts the image data obtained from different light receiving unit creates image data representing the whole of the image information carrying the light receiving portion overlapping said another.
[0023]
Note that the above-mentioned compensation is an image representing the entire image information carried by the image carrier using line detection means comprising one line sensor without a single joint having a large number of light receiving portions arranged in a line. This is to compensate for the goal that the image data that has been subjected to the above weighted addition processing has a noise component equivalent to the noise component that would have occurred in the acquired image data when the data was acquired. However, this compensation is not limited to the case of completely compensating for the noise component reduced by the weighted addition process, as long as the difference in noise component appearing in each region in the entire created image can be reduced. A part of the noise component reduced by the addition process or a part more than the noise component reduced by the weighted addition process may be compensated.
[0024]
In addition to the line sensors 10A and 10B, the line detection unit 20 forms an image of a linear region S extending in the main scanning X direction on the document 30 on the light receiving units of the line sensors 10A and 10B. Imaging signals 21A and 21B made of a mold lens and the like extending in the main scanning X direction, and light signals propagated through the imaging lenses 21A and 21B are received and detected by the light receiving portions of the line sensors 10A and 10B, respectively. A / D converters 23A and 23B for converting image data composed of digital values, and the imaging lens 21A is an image of a region S1 which is a part of a linear region S extending in the main scanning X direction on the document 30. Is imaged on the light receiving portion of the line sensor 10A, and the imaging lens 21B generates an image of the area S2 which is a part of the linear area S and partially overlaps the area S1. And forms an image on the light receiving portion of the B.
[0025]
The document 30 is illuminated by a line light source 62 including a number of LD light sources arranged in the main scanning X direction and a toric lens for condensing light emitted from these LD light sources in the linear region S. The light reflected by the linear regions S1 and S2 on the original 30 upon receiving the light from the line light source 62 is imaged on the light receiving portions of the line sensors 10A and 10B.
[0026]
Next, the operation in the above embodiment will be described.
[0027]
FIG. 3 is a conceptual diagram showing a state of combining image portions acquired from two line sensors, and FIG. 4 is a conceptual diagram showing a state of combining two image data groups.
[0028]
The original 30 is illuminated by the line light source 62 while moving the original 30 in the sub-scanning Y direction by driving the sub-scanning unit 40. Light that has been irradiated with light from the line light source 62 and reflected by the linear region S on the document 30 is coupled to the light receiving unit of the line sensor 10A and the light receiving unit of the line sensor 10B through the imaging lenses 21A and 21B. Imaged. In the light receiving portions overlapping each other, for example, the light reflected at the position P included in the regions S1 and S2 in the linear region S on the document 30 passes through the imaging lenses 21A and 21B, respectively. The image is formed on the light receiving part Ak that is one of the overlapping light receiving parts located at the end 11A and the light receiving part B4 that is the other of the overlapping light receiving parts located at the end 11B of the line sensor 10B. Light is received (see FIG. 2).
[0029]
The electrical signal received and detected by the light receiving unit of the line sensor 10A is A / D converted by the A / D converter 23A to obtain an image data group A, and the electrical signal received and detected by the light receiving unit of the line sensor 10B. A typical signal is A / D converted by an A / D converter 23B to obtain an image data group B. Each image data of the image data group A and the image data group B represents image information for each position on the document 30.
[0030]
The image data group A and the image data group B output from the line detection unit 20 are input to the weighted addition processing unit 51 of the image data creation unit 50, and the weighted addition processing unit 51 includes the image data group A and the image data group B. The image data obtained from each of the overlapping light receiving units is subjected to a weighted addition process for decreasing the weight of the image data obtained from the light receiving unit located at the end of the line sensor 10A or the line sensor 10B. Thus, one image data for each position is obtained. Thereafter, these image data are input to the noise compensation unit 52, and the noise compensation unit 52 compensates the noise component reduced by the weighted addition processing for each of the image data subjected to the weighted addition processing. Then, the image information carried by the document 30 using the image data supplemented with the noise components obtained from the light receiving units overlapping each other and the image data obtained from a light receiving unit different from the light receiving units overlapping each other. The image data W representing the entire image WW showing the entire image is created.
[0031]
Here, as shown in FIG. 3, the image portion AA represented by the image data group A detected and output by the line sensor 10A and the image portion BB represented by the image data group B detected and output by the line sensor 10B are referred to. Part of it is overlapping. That is, the image region R2 in the entire image WW is a region represented by image data obtained from the light receiving units that overlap each other, and the image data representing the image region R2 is weighted and added by the weighted addition processing unit 51. The noise component is compensated by the noise compensation unit 52 and created.
[0032]
Here, a case where image data W representing the entire image WW is created will be described with reference to FIG.
[0033]
As shown in FIG. 4, each light receiving part of the line sensor 10A is a light receiving part A1, A2, A3,... Ak-3, Ak-2, Ak-1, Ak, and each light receiving part of the line sensor 10B is a light receiving part B1. , B2, B3, B4,..., Be-2, Be-1, and Be, the overlapping light receiving parts are the light receiving part Ak-3 and the light receiving part B1, the light receiving part Ak-2 and the light receiving part B2, and the light receiving part Ak-. 1 and light receiving part B3, and light receiving part Ak and light receiving part B4. Therefore, the image data group A representing the image portion AA is obtained from the light receiving portions A1, A2, A3,... Ak, and the image data group B representing the image portion BB is obtained from the light receiving portions B1, B2, B3,. . Note that image data representing the image region R2 in the entire image WW is received from the light receiving units Ak-3, Ak-2, Ak-1, Ak, and the light receiving units B1, B2, B3, B4, which are the light receiving units that overlap each other. Acquired based on the acquired image data.
[0034]
Here, when creating image data from W (1, j) to W (m, j) in the row (j) of the image data W representing the entire image WW, the image data W (1, j) is used. Up to the image data W (k−4, j), the image data A (1, j) to image data A (k−4, j) in the image data group A are adopted, and from the image data W (k + 1, j). Up to the image data W (m, j), the image data B (5, j) to the image data B (e, j) in the image data group B are employed.
[0035]
For image data representing the image region R2 in the entire image WW obtained from the light receiving units that overlap each other, the image data W (k-3, j) is the image data A (k-3, j) of the image data group A. ) And image data B (1, j) of image data group B, and image data W (k-2, j) is generated from image data A (k-2, j) and image data B (2, j). The image data W (k-1, j) is created from the image data A (k-1, j) and the image data B (3, j), and the image data W (k, j) is created from the image data A ( k, j) and image data B (4, j).
[0036]
Next, a weighted addition process performed on image data obtained from overlapping light receiving units will be described. 5 and 6 are diagrams showing the relationship between the position of image data and the coefficient of weighted addition processing applied to the image data, and FIG. 7 shows the difference in standard deviation depending on the location of the noise component in the image after weighted addition processing. FIG. In addition, the whole image represented by the image data before the weighted addition processing is performed and the noise component is compensated is represented by the whole image VV, and the image data representing the whole image VV is represented by the image data V.
[0037]
<Weighted addition processing>
Image data V (k-3, j) of j rows created by weighted addition processing of image data A (k-3, j) of image data group A and image data B (1, j) of image data group B ) Is obtained by the following equation with weighting coefficients α1 to α4.
[0038]
V (k−3, j) = α1 × A (k−3, j) + (1−α1) × B (1, j)
V (k−2, j) = α2 × A (k−2, j) + (1−α2) × B (2, j)
V (k-1, j) = [alpha] 3 * A (k-1, j) + (1- [alpha] 3) * B (3, j)
V (k, j) = α4 × A (k, j) + (1−α4) × B (4, j)
Here, since the weighting of the weighted addition process is smaller as the image data obtained from the light receiving portion located at the end of the line sensor among the light receiving portions overlapping each other, the weighting representing the image region R2 as shown in FIG. When the value of the column (k-3) of the image data V subjected to the addition process is obtained, that is, for the column (K-3) of the image data group A and the column (1) of the image data group B. When the weighted addition process is performed, the weighting on the image data group B is small, and 0.2 is given to the value of the weighting coefficient (1-α1) in the above equation (here, the column (K of the image data group A) The weight α1 of −3) is 0.8). When the weighted addition processing is performed on the column (k) of the image data V, that is, the column (K) of the image data group A and the column (4) of the image data group B, the image data group A The weight is small, and 0.2 is given to the value of the weighting coefficient α4 in the above formula (here, the weight (1-α4) of the column (4) of the image data group B is 0.8). That is, as shown in FIG. 6, the weighting coefficient α for the image data group A and the value of (1-α) for the image data group B are the images obtained from the light receiving units located at the ends of the respective line sensors. It is determined according to the distance from the light receiving portion located at the end of the line sensor so that the data becomes smaller. In this way, the weighting coefficient α1 = 0.8, (1-α1) = 0.2 of the column (k-3), the weighting coefficient α2 = 0.6, (1-α2) of the column (k-2). = 0.4, column (k-1) weighting factor α3 = 0.4, (1-α3) = 0.6, column (k) weighting factor α4 = 0.2, (1-α4) = 0 .8 is defined.
[0039]
In this way, the image data obtained from each of the overlapping light receiving units is subjected to weighted addition processing to obtain one image data for each position, thereby creating image data V representing the image region R2.
[0040]
Note that the image data V subjected to the weighted addition process is represented by the noise reduction effect in which the noise component included in each image data is canceled by the calculation between the image data when the upper weighted addition process is performed. The noise component that appears in the image area R2 in the entire image VV is smaller than other areas. That is, as shown in FIG. 7, only the standard deviation σ1 of the noise component of the image in the image region R1 formed by only the image data from the image data group A in the entire image V, and only the image data from the image data group B Is approximately equal to the standard deviation σ3 of the image noise component in the image region R3 formed in the above, but the standard deviation σ2 of the image noise component in the image region R2 created from the image data group A and the image data group B is , Smaller than σ1 and σ3. For this reason, due to the difference between the noise component in the image region R2 and the noise component in the other region, the image region R2 becomes a region having a different texture from the other regions. Further, since the noise reduction effect increases as the weighting coefficient when the weighted addition process is performed approaches 0.5, streaks appear in the column direction (sub-scanning Y direction) at the center of the image region R2.
[0041]
Next, a case where the noise component reduced by the weighted addition process is compensated will be described. FIG. 8 is a diagram showing a standard deviation of noise components for each column in image data after weighted addition processing, FIG. 9 is a conceptual diagram showing how noise improved image data is obtained, and FIG. 10 is an image subjected to weighted addition processing. FIG. 11 is a diagram showing a correspondence relationship between the value Do of the data V and the value Nr of the noise-enhanced image data N, FIG. 11 is a diagram showing the normalized improvement function, and FIG. 12 is the image data W supplemented with the noise component. It is a figure which shows a mode.
[0042]
<Noise compensation>
Noise components included in the image data of the image data group A or the image data group B before the weighted addition processing are classified as follows.
[0043]
Nx: X-ray quantum noise
Ni: Document structure noise
Nh: photon noise
Ne: Electric noise of CCD element
Here, the X-ray quantum noise Nx and the document structure noise Ni are noise components of the reading object itself, and do not change even if the above-described weighted addition processing is performed. On the other hand, the photon noise Nh and the electrical noise Ne of the CCD element are noise components generated independently in the CCD element, and are reduced by weighted addition processing.
[0044]
The above-mentioned noise component compensation is performed by predicting in advance the photon noise Nh and the electrical noise Ne of the CCD element that are reduced by the weighted addition process, and this reduced noise component (herein also referred to as noise improving component). It can be implemented by adding to the image data that has been subjected to the weighted addition processing.
[0045]
Here, there are various methods for obtaining the noise improving component from the photon noise Nh and the electrical noise Ne of the CCD element, but as a direct and simple method, a so-called solid manuscript having a constant density is prepared regardless of the place. Then, a two-dimensional pattern representing the noise improving component can be obtained based on the image data for each position obtained by performing weighted addition processing on the image data obtained by actually reading the solid original.
[0046]
Here, the noise improving component is a value conceptually shown as follows. That is, as shown in FIG. 8, in image data V (i, j) obtained by reading a so-called solid original having a constant density regardless of a place and performing weighted addition processing, the image data V (i, 1) If the standard deviation of the noise component included in each image data representing the linear area of the column (i) represented by the image data V (i, n) is σi, the image area R1 composed only of the image data A , And the standard deviation of the noise component included in the image data representing the linear region of each column (i) in the image region R3 composed only of the image data B is approximately constant and σh≈σi (h = 1 to k−). 4, i = k + 1 to m). However, the standard deviation of the noise component in the i = k−3 to k column subjected to the weighted addition process changes for each column, and the relationship between the position of each column and the standard deviation of the noise component is quadratic. It can be approximated by the curve g. A difference δ between the straight line σ = σn and the quadratic curve g in FIG. 8 indicates a noise improving component.
[0047]
Actually, this noise improving component is represented by a random noise component having a different value for each position on the image VV, and the two-dimensional image data V () obtained by reading a solid document. The noise component reduced in the image region R2 on which the weighted addition processing in i, j) is performed becomes the noise improving component. Here, as shown in FIG. 9, the solid document is illuminated so that the maximum amount of light that can be detected by the line detection unit 20 is received and the maximum value of the image data is obtained, and this brightly illuminated solid document is read. The image data V (i, j) obtained by performing weighted addition processing on the obtained image data and the above-described brightly illuminated solid original are read by a line detection unit including a seamless line sensor, and weighted addition processing is not performed. Reference image data V ′ (i, j) equivalent to the obtained image data is prepared, and image region R2 is obtained by subtracting image data V (i, j) from reference image data V ′ (i, j). The noise-enhanced image data N (i, j) representing the noise-enhancing component at each position is obtained. Since the noise component included in the reference image data V ′ is constant regardless of the area, the reference image data V ′ is the image data V in which the image area R2 in the image data V is the image area R1, R3. It can be obtained by interpolating with image data.
[0048]
Note that the noise-enhanced image data N (i, j) represents a noise-enhancing component when the maximum amount of light that can be detected by the line detector 20 is received. Here, the photon noise Nh is a line detection. The amount of light detected by the line detection unit 20, that is, the amount of light detected by the line detection unit 20 changes non-linearly according to the value of the image data acquired by the line detection unit 20. Based on the read image data (for example, by changing the density of the solid document or the intensity of the illumination light), the value D indicated by the image data V (i, j) subjected to the weighted addition process, and this image A correspondence relationship with the value Nr indicated by the noise-enhanced image data N (i, j) corresponding to the data V (i, j) is measured and determined in advance. As shown in FIG. 10, this correspondence relationship is expressed by functions Nd ′ and Nr = Nd ′ (D) in which the value Nr increases as the value D increases.
[0049]
Since the image data that exists in each of the image data group A and the image data group B to be subjected to the weighted addition processing is image data obtained from the light receiving units that overlap each other, the values almost coincide with each other. The value D of the image data V obtained as a result of the weighted addition process is also substantially equal to the above value although the noise component is reduced.
[0050]
Next, as shown in FIG. 11, the correspondence relationship Nr = Nd ′ (D) between the value D of the image data V subjected to the weighted addition process and the noise improvement value Nr is set to the maximum expected value D. Normalization is performed so that the noise improvement value corresponds to the value Dmax. As a result, the standardization / enhancement function Nd, Nd (D) = (Nd ′ (D) / Nd ′ (Dmax)) is obtained, and this function reduces the value D on the basis of Nd (Dmax) = 1. Therefore, the value Nr is also reduced.
[0051]
Here, the value of the image data W (i, j) representing the entire image WW supplemented with the noise component is defined as DD (i, j), and weighted addition processing corresponding to the image data (i, j) is performed. Assuming that the value of the image data V (i, j) is D (i, j), the noise improving component Nr (i, j) at the position (i, j) is Nr (i, j) = Nd (D ( i, j)) × N (i, j). As shown in FIG. 12, the noise improvement component Nr (i, j) is added to the value D (i, j) of the image data V that has been subjected to the weighted addition process, thereby compensating the noise component. A value DD (i, j) of W is obtained. That is, the value DD (i, j) is obtained by the following formula: DD (i, j) = D (i, j) + Nd (D (i, j)) × N (i, j).
[0052]
Thus, the image carried by the document 30 using the image data supplemented with the noise components obtained from the overlapping light receiving units and the image data obtained from a light receiving unit different from the overlapping light receiving units. Image data representing the entire information is created. As a result, the image area R2 that has been subjected to the weighted addition process in the image WW represented by the image data W does not appear to have a different texture from the other areas (for example, no streaking occurs). The deterioration of quality can be suppressed.
[0053]
In addition, the method of creating the image data representing the entire image may be a method of creating image data representing one entire image based on a plurality of image data groups each representing a plurality of image portions sharing the same area. For example, when small images projected by a plurality of projectors are combined to create one large image, or a plurality of three-dimensional image data representing a small region are combined to form a single image This method can be applied when creating large three-dimensional image data.
[0054]
As described above, according to the present invention, it is possible to suppress deterioration in the quality of an image created by combining a plurality of image portions sharing the same area.
[0055]
The document is not limited to being moved in the sub-scanning Y direction with respect to the line detection unit, and the document may be moved relative to the line detection unit in the sub-scanning Y direction.
[0056]
The line sensor can be a CCD element, a CMOS element, or the like. In the above embodiment, the light receiving units that overlap each other are the four light receiving units in each line sensor. However, this number may be one or more, for example, 100 light receiving units in each line sensor are mutually connected. It may be an overlapping light receiving unit.
[0057]
In the above-described embodiment, an example of the line detection unit configured by arranging two line sensors is shown. However, the line detection unit includes a plurality of line sensors, and a light receiving unit positioned at each end of each of the line sensors. As long as the light emitted from the same position of the image carrier is arranged side by side so as to detect each other, the line detection means composed of three or more line sensors may be used as the line sensor. It goes without saying that the effect of can be obtained.
[0058]
In addition, when the original is a radiation image conversion panel, an excitation light irradiation unit that emits a linear excitation light and generates a stimulated emission light from a linear region of the radiation image conversion panel instead of a line light source. The radiation image recorded on this radiation image conversion panel can be read. In this case, the light generated from the radiation image conversion panel is stimulated emission light. More specifically, a side view showing the excitation light irradiation unit, the line detection unit, and the radiation image conversion panel in FIG. 13, and a region where the stimulated emission light is generated from the radiation image conversion panel in FIG. As shown in the top view showing the positional relationship with the line sensor that receives light, by providing an excitation light irradiation unit 60T instead of a line light source and a line detection unit 20T composed of five line sensors, the original 30 The radiation image recorded on the radiation image conversion panel 30T can be read from the radiation image conversion panel 30T instead of the above.
[0059]
Here, the excitation light irradiation unit 60T is perpendicular to the surface of the radiation image conversion panel 30T from each LD in the excitation light source 61T, and the excitation light source 61T composed of a number of LDs arranged in the main scanning X direction. And a toric lens 62T extending in the main scanning X direction for condensing the excitation light Le emitted in the direction on a linear region S extending in the main scanning X direction on the radiation image conversion panel 30T.
[0060]
Further, the line detection unit 20T includes five line sensors 7A, 7B, 7C, 7D, and 7E that are composed of CCD elements having a large number of light receiving units arranged linearly in the main scanning X direction. The light receiving units located in the unit are arranged in the main scanning X direction so that light emitted from the same position on the radiation image conversion panel 30T, which is an image carrier, overlaps each other. Here, the line sensors 7A, 7C, and 7E are arranged on the side of the linear region S on the radiation image conversion panel 30T in the sub-scanning-Y direction (arrow -Y direction in the figure), and the line sensors 7B and 7D are The linear region S on the radiation image conversion panel 30T is disposed on the sub-scanning + Y direction (arrow + Y direction in the figure) side. That is, the line sensors 7A, 7C, and 7E and the line sensors 7B and 7D are disposed on opposite sides of the optical path of the linear excitation light that is irradiated from the excitation light irradiation unit 60T to the radiation image conversion panel 30T. Arranged in a staggered pattern.
[0061]
In addition to the line sensors 7A, 7B, 7C, 7D, and 7E, the line detection unit 20T uses the line sensors 7A, 7B, and 7C to display an image of a linear region S extending in the main scanning X direction on the radiation image conversion panel 30T. An electrical image that is received and photoelectrically converted by an imaging optical system 21T composed of five imaging lenses, which will be described later, and the line sensors 7A, 7B, 7C, 7D, and 7E. There are provided five A / D converters 23T for converting each image signal into image data composed of digital values. The imaging optical system 21T includes imaging lenses 21A, 21B, 21C, 21D, and 21E formed by arranging a large number of gradient index lenses in the main scanning X direction, and is linear on the radiation image conversion panel 30T. The area portions Sa, Sb, Sc, Sd, and Se are imaged on the line sensors 7A, 7B, 7C, 7D, and 7E by the imaging lenses 21A, 21B, 21C, 21D, and 21E, respectively. As a result, the stimulated emission light Ke generated from the linear region portions Sa, Sb, Sc, Sd, Se on the radiation image conversion panel 30T by the irradiation of the excitation light Le from the excitation light irradiation unit 60T is converted into each line sensor 7A. , 7B, 7C, 7D, and 7E are imaged and received.
[0062]
Here, the stimulated emission light Ke generated from the overlapping region Sab of the linear region portion Sa and the linear region portion Sb is detected by the light receiving unit located at the end of each of the line sensors 7A and 7B. Stimulated emission light Ke generated from the overlapping region Sbc of the region portion Sb and the linear region portion Sc is detected by the light receiving portions located at the ends of the line sensors 7B and 7C, and the linear region portion Sc The stimulated emission light Ke generated from the overlapping region Scd with the linear region portion Sd is detected by the light receiving portions located at the end portions of the line sensors 7C and 7D, and the linear region portion Sd and the linear region portion are detected. The stimulated emission light Ke generated from the overlapping region Sde is detected by the light receiving portion located at the end of each of the line sensors 7D and 7E, and image data representing each position on the radiation image conversion panel 30T is obtained.
[0063]
For the image data obtained from the overlapping light receiving portions located at the end portions of the line sensors 7A, 7B, 7C, 7D, and 7E, the light receiving portion located at the further end of the line sensor is added by the addition processing portion as described above. The image data obtained from the unit is subjected to a weighted addition process for reducing the weight to obtain one image data, and the noise compensation unit reduces the image data obtained by the weighted addition process by the weighted addition process. The image data supplemented with the noise component is adopted as the image data at each position where the noise component is compensated and the light is detected by the light receiving unit of each line sensor. The image data representing the entire image information carried by the radiation image conversion panel 30T using the image data obtained from the unit is It is made.
[Brief description of the drawings]
FIG. 1 is a perspective view showing a schematic configuration of an image data creation device according to an embodiment of the present invention.
FIG. 2 is an enlarged perspective view showing an optical path of light reflected from a document and imaged on a light receiving unit of a line detection unit.
FIG. 3 is a diagram illustrating a state in which image portions acquired from two line sensors are combined.
FIG. 4 is a conceptual diagram showing how two image data groups are combined.
FIG. 5 is a diagram showing how weighted addition processing is performed on two image data;
FIG. 6 is a diagram showing the relationship between the position of image data and the coefficient of weighted addition processing applied to the image data;
FIG. 7 is a diagram illustrating a difference in standard deviation depending on the location of a noise component in an image after weighted addition processing;
FIG. 8 is a diagram showing a standard deviation for each column of noise components in an image after weighted addition processing;
FIG. 9 is a conceptual diagram showing how to obtain noise-enhanced image data.
FIG. 10 is a diagram illustrating a correspondence relationship between the value of image data subjected to weighted addition processing and the value of noise-enhanced image data.
FIG. 11 is a diagram showing a correspondence relationship indicated by a standardized improvement function.
FIG. 12 is a diagram illustrating how image data in which noise components are compensated is obtained
FIG. 13 is a side view showing an excitation light irradiation unit, a line detection unit, and a radiation image conversion panel.
FIG. 14 is a diagram showing a positional relationship between a region of stimulated emission light generated from the radiation image conversion panel and a line sensor that receives the stimulated emission light.
[Explanation of symbols]
10A line sensor
10B line sensor
11A end
11B end
20 line detector
30 Manuscript
40 Sub-scanning unit
50 Image creation section
51 Weighted addition processing unit
52 Noise compensation part
62 line light source
100 Image data creation device

Claims (4)

When creating image data representing one entire image by combining a plurality of image data groups representing each of a plurality of image parts sharing the same region, as image data at each position of the same region in the entire image, For each position, a plurality of image data at that position belonging to each of the image data groups is subjected to a weighted addition process in which the weight of the image data located closer to the edge of the image portion is reduced to obtain one image data. An image data creation method for creating image data representing the entire image using the image data,
A method of creating image data, wherein image data representing the entire image is created by compensating for noise components reduced by the weighted addition process for the image data obtained by the weighted addition process. When creating image data representing one entire image by combining a plurality of image data groups representing each of a plurality of image parts sharing the same region, as image data at each position of the same region in the entire image, For each position, a plurality of image data at that position belonging to each of the image data groups is subjected to a weighted addition process in which the weight of the image data located closer to the edge of the image portion is reduced to obtain one image data. An image data creation device that creates image data representing the entire image using the image data,
An image data creating apparatus, wherein image data representing the entire image is created by compensating for noise components reduced by the weighted addition process for the image data obtained by the weighted addition process. A plurality of line sensors having a large number of light receiving parts arranged in a line are arranged so that the light receiving parts located at the ends of the line sensors respectively detect light emitted from the same position of the image carrier. With respect to the line detection means arranged side by side in the main scanning direction, the line detection means emits light emitted from the image carrier while relatively moving the image carrier in the sub-scanning direction intersecting the main scanning direction. An image data creating method for creating image data representing image information carried by the image carrier detected by
As the image data of each position where the light is detected by the light receiving units of the plurality of line sensors, the light was emitted from each light receiving unit of the plurality of line sensors that detected the light emitted from the position for each position. The image data obtained from the light receiving unit located at the end of the line sensor is subjected to a weighted addition process for decreasing the weight to obtain one image data, and the image data obtained by the weighted addition process is obtained. An image data creation method, wherein image data representing the entire image information is created by using image data obtained by compensating for noise components reduced by the weighted addition processing. A plurality of line sensors having a large number of light receiving parts arranged in a line are arranged so that the light receiving parts located at the ends of the line sensors respectively detect light emitted from the same position of the image carrier. Line detection means arranged side by side in the main scanning direction, scanning means for moving the image carrier relative to the line detection means in the sub-scanning direction intersecting the main scanning direction, and the line detection means An image carried by the image carrier based on the image data obtained by detecting the light emitted from the image carrier by the line detection means while relatively moving the image carrier in the sub-scanning direction. An image data creation device comprising image data creation means for creating image data representing information,
The image data creation means is obtained from each light receiving unit of the plurality of line sensors that has detected the light emitted from the position for each position where light is detected by the light receiving units of the plurality of line sensors. A weighted addition processing means for obtaining one image data by performing a weighted addition process for reducing the weight of the image data obtained from the light receiving unit located at the end of the line sensor with respect to the image data;
Noise compensation means for compensating for the noise component reduced by the weighted addition process for the image data obtained by the weighted addition process,
An image data creation device that employs image data supplemented with the noise component as image data at each position where light is detected redundantly by the light receiving units of the plurality of line sensors, and creates image data representing the image information. .

Images