JPH06105145A - Picture input device - Google Patents

Abstract

PURPOSE:To reduce the number of correction coefficients required for shading correction of a face light source by providing a stationary facial light source (rod fluorescent lamps) for a transmission type film original on the picture input device having a linear CCD line sensor 4 and reflecting light use linear moving light source lamps on a transparent glass plate. CONSTITUTION:A correction coefficient for each picture element of a sensor 4 is stored to a nonvolatile memory 8 for each device. A quotient being division of a read output of a reference board 1 by an output of the memory 8 (output of 7-1) is given as a correction coefficient of modulation shading of a lens 11 and linear luminous quantity shading of lamps 2-1, 2-2 and the coefficient is stored in a memory 9-1 as a correction coefficient at reading of a reflection type original. A quotient being a division of an output obtained through scanning of face light sources 2-3-2-6 before original mount by an output of the memory 8 (output of 7-1) is given as a shading correction coefficient for the lens 11 and a transparent facial light source and the coefficient is stored in a white level memory 9-2.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image input device capable of inputting an image of reflected light of an original image such as a paper surface of a document by using a one-dimensional image sensor, and more particularly, to an image input device of an X-ray film as well as an image of reflected light The present invention relates to an image input device provided with a two-dimensional shading correction means capable of using a surface light source for an original image and inputting an image by the transmitted light thereof and correcting uneven irradiation of the surface light source.

[0002]

2. Description of the Related Art Generally, as an image input device (imaging device) using a one-dimensional image sensor such as a CCD line scanner, a reflection type image input device for receiving reflected light on a paper surface, an X-ray film and a photographic film are used. A transmission type image input device that receives the transmitted light of a film is known. In many reflection type devices, a one-dimensional light source for illumination moves together with the image input unit (light receiving unit), and in transmissive devices, the light source for illumination moves within the moving range of the image input unit (light receiving unit). Many of them are fixed in two dimensions.

In this type of image input device (image pickup device),
Even when an object (subject) having the same (uniform) brightness is read, the output signal f (x,
y) changes depending on the position (x, y) of the object. This is called a shading phenomenon. Shading is caused by variations in sensitivity of each element (that is, pixel) of a one-dimensional scanner (CCD, etc.), modulation shading caused by aberration of an optical system (lens system), or uneven light quantity of an illumination system such as a lamp. There is light amount shading that occurs. Here, x is the arrangement direction of the elements (pixels) of the one-dimensional scanner, and y is the moving direction of the one-dimensional scanner perpendicular to it.

Conventionally, as a technique for correcting such shading, for example, No. 106 of "Image Processing" by Hirosaki Ozaki et al.
Page (hereinafter referred to as "Reference 1") and May 3, 1988
Published by Shokodo on the 1st, "Industrial image processing" by Masato Ejiri
There is a method described on page 24 of the above. In order to perform shading correction, first, the entire surface of a subject having uniform brightness is scanned and the image output value at that time, that is, the measured value fo
(X, y) is measured on the entire surface and this value is stored in the memory in advance. Then, for an image signal output f (x, y) read from a general subject thereafter, fc (x, y) = K × f (x, y) / fo (x, y) (K is originally When an image output of fc (x, y) is obtained by performing a correction process of a desired brightness), the influence of shading can be removed.

In the conventional shading correction technique,
A white reference plate is an object of uniform brightness that is read prior to image input from the target subject. By illuminating the white reference plate with a lamp and reading the reflected light, storing that value in memory, and using this stored value as a coefficient for correction, it is possible to obtain an image of the target object that is generally obtained. Make a correction. The data stored in the memory at this time is called a shading correction coefficient array. When inputting an image by transmitted light like film input, the surface light source is read prior to image input and the value is stored in memory.
It is conceivable to perform correction on a generally obtained image.

[0006]

An image input device having a one-dimensional image sensor is used not only to read the reflected light on the original paper surface (original image), but also to use such an image input device as in an X-ray film. It is convenient to make it possible to read the transmitted light from the film and to have both the reflection type and the transmission type. Here, the illumination of the image input device that uses the one-dimensional image sensor and reads the reflected light on the paper is 1
A dimensional light source may be used, and the shading correction coefficient array is a one-dimensional array. Next, the illumination of the image input device for reading the transmitted light requires a two-dimensional surface light source, and the shading correction coefficient array needs a two-dimensional array. Considering the input of an image with a pixel number of 400 DPI (dots per inch, about 16 dots per millimeter) from an A3 original, the number of the one-dimensional shading correction coefficient arrays is 500.
Although it is about 0, the number of two-dimensional shading correction coefficient arrays necessary for the correction of the surface light source is 5000 × 6400 = 3.
You will need 2 mega units. The number of bits required for shading correction is about 8 bits (1 byte) per pixel,
The data amount of the two-dimensional shading correction coefficient array is 32.
Mbytes (8 times the size of a binary image of the same size) are also required, which is difficult to realize.

Therefore, the first object of the present invention is to solve the above-mentioned problems of the prior art and to obtain a good image quality for both the reflected light on the paper and the input light for the film by using the one-dimensional image sensor. An object is to provide a secure image input device.

A second object of the present invention is to reduce the number of two-dimensional shading correction coefficients necessary for shading correction as much as possible when reading a two-dimensional image with transmitted light using a one-dimensional image sensor and a surface light source. It is to provide an image input device capable of

[0009]

In order to achieve the above first object, the present invention provides an image input device having an image input section for inputting an image by reflected light of an original image using a one-dimensional image sensor, A surface light source for transmitted light is provided so that an image of the transmitted light of the original image can be input by the image input unit, and the image input unit further performs at least two-dimensional shading correction of the surface light source. Means are provided.

In order to achieve the above second object,
Basically, the present invention provides shading correction, which is a correction of sensitivity unevenness of a sensor element that requires correction in pixel units, and corrections other than that, which does not require correction in pixel units and allows a large sampling point interval. The processing is divided into two groups, that is, the correction for the modulation shading of the lens system and the correction for the uneven light amount of the illumination system. Specifically, the shading correction means stores in advance a sensor one-dimensional shading correction coefficient array for correcting sensitivity unevenness between pixels of the one-dimensional image sensor, and, after this storage, the surface of the image sensor by the image sensor. By scanning a light source and correcting the scanning output with the sensor one-dimensional shading correction coefficient array, at least a two-dimensional shading correction coefficient array representing two-dimensional irradiation unevenness of the surface light source (actually, an optical lens system). And a two-dimensional shading correction coefficient array for storing the one-dimensional shading correction coefficient array for the modulation shading) and storing the two-dimensional shading correction coefficient array separately from the one-dimensional shading coefficient array.

Here, the interval of sampling points in the sensor primary shading correction coefficient array is one pixel, and the other shading correction coefficient array is a two-dimensional shading correction coefficient array (optical lens system) by the surface light source. 1-dimensional shading correction coefficient array of 1) is sufficiently wider than one pixel according to the rate of change of the shading amount of the surface light source (may include the rate of change of shading amount by the optical lens system). Interval (eg 1
6 pixels).

The two-dimensional shading correction coefficient array thus obtained is used together with the sensor one-dimensional shading correction coefficient array to read all the shading corrections of the read image when the film is read in a transmissive type using the surface light source. Used for.

[0013]

The operation based on the above configuration will be described.

According to the present invention, in a reflection type image input device using a one-dimensional image sensor, a transmission type image input device using a one-dimensional image sensor is provided by disposing a surface light source for transmitted light. Can also be used as. In this case, by providing the shading correction means for performing the two-dimensional shading correction of the surface light source, it is possible to obtain a good image quality even when an image of transmitted light is input.

Further, when performing shading correction, it is necessary to perform shading correction for each pixel because there is a discontinuity in sensitivity of each pixel of the image sensor, but in a one-dimensional image sensor, the sensitivity unevenness is corrected. For the shading correction coefficient array, one line is enough, and the memory only needs to have a storage capacity for one line. On the other hand, since the two-dimensional irradiation unevenness (change in light amount) due to the surface light source has continuity, the sampling point interval for shading correction may be sufficiently wider than the pixel interval, but since it is a two-dimensional light source, shading is performed. The correction coefficient array requires two dimensions, and the memory also requires a storage capacity for two dimensions of the entire surface. With respect to the modulation shading by the optical system (lens system), the sampling point interval may be sufficiently wider than the pixel interval because of continuity, and the shading correction coefficient array is required for one line.

In the present invention, paying attention to such characteristics of various shading correction coefficients, a sensor one-dimensional shading correction coefficient array for correcting sensitivity unevenness of a one-dimensional image sensor and irradiation unevenness correction of a surface light source are used. Since the two-dimensional (illumination) shading correction coefficient array is separately stored in the memory, the number of sensor one-dimensional shading correction coefficient arrays and the memory capacity for storing the same are
It is enough for the number of pixels in one line, and it is two-dimensional (illumination)
The number of shading correction coefficient arrays and the memory capacity for storing them need only be the number of sampling points that is extremely smaller than the number of pixels of a two-dimensional screen, and the number of these shading correction coefficient arrays can be reduced. The storage capacity can be reduced. The modulation shading correction coefficient array for the modulation shading of the optical system (lens system) can be combined with either the sensor one-dimensional shading correction coefficient array or the two-dimensional (illumination) shading correction coefficient array. (Illumination) It can be obtained together with the shading correction coefficient array.

Now, the shading correction operation of the present invention will be described in more detail. The shading correction coefficient can be represented by the product of the following three coefficients.

(1) Coefficient of sensitivity variation between bits of an image sensor (between each element, that is, between each pixel) (2) Coefficient of modulation shading of an optical system (lens system) (3) Irradiation unevenness of an illumination system (lamp) Considering the required number of the above three coefficients, as described above, in (1), since there is no continuity between bits (one pixel), the number of coefficients is as many as the number of pixels. Regarding (2) and (3), since there is continuity between bits and the rate of change thereof is gentle, it is not necessary to sample all pixels, and a large sampling point interval can be set. The number of coefficients can also be reduced.

When using a one-dimensional image sensor,
The coefficients (1) and (2) become a one-dimensional coefficient array that does not change in the moving direction (y) of the image sensor but changes only in the line direction (x).
(Sensitivity variation of sensor element) and fm (x) (modulation shading coefficient) are set.

When using a surface light source, the coefficient of (3) is
It becomes a two-dimensional coefficient array that changes in the x direction and the y direction, and this is designated as fl (x, y).

If the output of the surface light source that has not been corrected is fo (x, y), this output is a shading correction coefficient array including all three types of shading, and is a product of three types of shading correction coefficient arrays. equal. Expressed as an expression, fo (x, y) = fs (x) × fm (x) × fl
It becomes (x, y).

In the above equation, fs (x) is a value peculiar to the image sensor and hardly changes with time.
It is possible to register in a non-volatile memory or the like (preliminarily obtained using a light source with a known light amount distribution, a light source in which the light amount shading is corrected, etc.), so that fs (x) of the surface light source can be read. Correction can be applied, and if this output is fi (x, y), then fi (x, y) = fo (x, y) / fs (x) = fm (x) × fl (x, y) As shown, this value is equal to the product of fm (x) and fl (x, y).

The number of coefficients of fm (x) and fl (x, y) can be made sufficiently smaller than the number of pixels as described above, and its product fi (x,
The same applies to the number of coefficients in y).

As described above, the correction coefficient fs (x) for each element (pixel) of the image sensor is previously registered in the non-volatile memory or the like, which is measured in advance for each component, and this correction coefficient is read when reading the surface light source. The number of two-dimensional shading correction coefficients that must be stored in the memory can be reduced by applying the correction according to the value of. If the sampling point interval is 16 pixels, the number of two-dimensional shading correction coefficient arrays is reduced to 1/256, the data amount is 1 Mbit, and only one memory element is required, so that shading correction of a surface light source is facilitated. It becomes feasible. In the above, fl
The same thing can be said even if (x, y) is replaced with a correction coefficient fl (x) for the light amount shading of the one-dimensional light source. 1
A one-dimensional light source is usually used as a light source of an image input device that inputs reflected light on a paper surface using a three-dimensional image sensor. To perform shading correction, a white reference plate having a constant lightness is read and its value is read. It has been performed in the related art to store a one-dimensional shading correction coefficient array in a memory. In this case, if the correction by fs (x) is performed at the time of reading the white reference plate, as in the case of the surface light source. The number of one-dimensional shading correction coefficients that must be stored in the memory can be reduced.

Further, according to the above shading correction method, both shading corrections of the light source for reflected light and the light source for transmitted light can be performed by the same circuit.

[0026]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention, in which 1 is a white reference plate, 2-1 to 2-6 are light sources such as rod-shaped fluorescent lamps, and 3-1 is a document support. A transparent glass plate, 3-2 is a diffusion plate, 4 is a one-dimensional image sensor such as a CCD line sensor, 5 is an amplifying unit for amplifying the output of the CCD sensor 4, and 6 is a digital image signal of the analog image of the amplifying unit A / D converter for converting to signal, 7-
1 and 7-2 are dividers, 8 is a non-volatile memory, 9-1 and 9-2 are white level memories, 10 is a white level memory 9-
A selector for switching the outputs of 1 and 9-2, 11 is a CCD
It is a lens for forming an image on the sensor 4.

The lamps 2-1 and 2-2 are glass plates 3-
1 is a light source for reflected light with respect to the surface of the original placed on 1 and the white reference plate 1 (having at least the same length as the width of the original), and is movable in the y direction (rightward in the figure) together with the CCD sensor. There is. The two lamps are provided to prevent a shadow even on a dented document or the like. The lens 11 forms an image on the CCD sensor with the full width of the original document in view. The lamps 2-3 to 2-6 are glass plates 3
In order to give transmitted light to the original film such as an X-ray film placed on the -1 through the diffusion plate 3-2,
The fixed surface light source for transmitted light is formed by having at least the same length as the width of the original film and arranged at a predetermined interval over at least the length of the original film. Non-volatile memory 8
In, the correction coefficient fs (x) of the variation in sensor sensitivity is stored in advance for each device at the time of shipment or installation of the device.

First, the case where the apparatus of this embodiment is used as a reflection type will be described.

The white reference plate 1 is irradiated with the light emitted from the lamps 2-1 and 2-2, the reflected light is read by the CCD sensor 4, and divided through the signal amplification section 5 and the A / D converter 6. Input to the device 7-1. The output of the A / D converter 6 is affected by three types of shading, and fs
It is represented by (x) × fm (x) × fl (x).

The output fs (x) of the nonvolatile memory 8 is A / D
The sensor sensitivity variation is corrected by dividing the output of the converter 6, and the output of the divider 7-1 is a combination of the light amount shading fl of the lamp 2-1 and the lamp 2-2 and the modulation shading of the optical system. Value fm (x) × fl
It is represented by (x). This output is the white level memory 9-1.
To store. Next, when reading the input original document (placed on the transparent glass plate 3-1) under the diffusion plate 3-2 by the reflected light, the lamp 2-1 and the lamp 2-2 are used as the light source, and the CCD sensor is used. 4 is moved and scanned in the y direction to obtain an image signal output K × fs (x) × fm (x) × fl (x).
In this case, the non-volatile memory 8 and the divider 7-1 correct the variation in the sensor sensitivity to obtain K × fm (x) × fl.
(X), and then the selector 10 causes the white level memory 9-
By selecting the output fm (x) × fl (x) of 1, the divider 7-2 corrects the light amount shading of the lamps 2-1 and 2-2 and the modulation shading of the optical system.

Next, the case where the apparatus of this embodiment is used as a transmission type will be described. In this case, the diffuser plate 3 is illuminated by the lamp 2-3, the lamp 2-4, the lamp 2-5, and the lamp 2-6 to form a surface light source. A transparent original such as a film is placed under the diffuser plate 3-2 (between the diffuser plate 3-2 and the transparent glass plate 3-1) and scanned by the CCD sensor 4 to input an image. In this case, the CCD sensor 4 scans the diffusion plate 3 to read the two-dimensional light amount shading of the surface light source before the transparent original is placed before the image input. At this time, the output of the A / D converter 6 is fs
(X) × fm (x) × fl (x, y), and the output f of the nonvolatile memory 8 is obtained by using the divider 7-1 for this output.
The variation in sensor sensitivity is corrected by s (x). At this time, the output of the divider 7-1 is represented by a combined value fm (x) × fl (x, y) of the two-dimensional light amount shading of the surface light source and the modulation shading of the optical system. This output is stored in the white level memory 9-2. Next, when the transparent original is placed under the diffusion plate 3 and the surface light source is illuminated to input an image, the output fs (x) of the nonvolatile memory 8 and the divider 7-
The sensor sensitivity variation is corrected by 1, and then the selector 10 outputs the output fm (x) of the white level memory 9-2.
By selecting fl (x, y), the divider 7-2
Corrects the two-dimensional light amount shading of the surface light source and the modulation shading of the optical system, and outputs the corrected output (video data) of all shadings.

According to the present embodiment, the correction means for the element sensitivity variation which needs the correction in the pixel unit, the modulation shading correction which does not require the correction in the pixel unit and the sampling interval can be wide, and the nonuniformity of the illumination system. It is possible to separately provide a correction means for performing the correction. Further, the reflected light shading correction coefficient and the surface light source transmitted light shading correction coefficient can be stored in separate memories and used for the respective corrections.

As described above, according to the present embodiment, the two-dimensional shading correction of the transparent original surface light source can be performed.

The one-dimensional modulation and shading correction coefficient array of the optical system (lens system) can be obtained and stored together with the one-dimensional sensor sensitivity shading correction coefficient array in advance. Only the one-dimensional uneven illumination shading correction coefficient array fl (x) by the lamps 2-1 and 2-2 is stored in the level memory 9-1, and the lamps 2-3 to 2-6 are stored in the white level memory 9-2. 2D uneven illumination shading correction coefficient array fl (x, y)
Only need to store.

[0036]

As described above in detail, according to the present invention, in a reflection type image input device using a one-dimensional image sensor, a surface light source for transmitted light is arranged to provide a transmission type image input device. It is also possible to use both the light source for reflection type and the two-dimensional shading correction means for surface light source for transmission type by providing the one-dimensional shading correction means for the reflection type light source and the two-dimensional input for the reflected light on the original paper surface and the transmitted light on the original film. On the other hand, an effect that a good image quality can be secured is obtained.

Further, a one-dimensional sensor sensitivity shading correction coefficient array, which requires correction for each pixel but needs correction for one line, and an optical system other than that, which has a sampling point interval sufficiently wider than the pixel interval ( Since the lens system) modulation shading correction coefficient array and the illumination system shading correction coefficient array, especially the two-dimensional light amount shading array by the surface light source, are separately obtained and stored, the number of sensor sensitivity shading correction coefficient arrays And other than that, particularly, the number of two-dimensional light amount shading correction coefficient arrays by the surface light source is reduced to reduce the total number of shading correction coefficient arrays and the storage capacity of the memory required for storing the same. The effect that can be obtained is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

[Description of Reference Signs] 1 white reference plate 2-1 to 2-2 movable fluorescent lamp (one-dimensional light source for reflected light) 2-3 to 2-6 fixed fluorescent lamp (surface light source for transmitted light) 3-1 transparent glass plate (Document support plate) 3-2 Diffusion plate 4 CCD sensor (one-dimensional image sensor) 5 Signal amplification unit 6 A / D converter 7-1, 7-2 Divider 8 Non-volatile memory 9-1, 9-2 White Level memory 10 Selector 11 Lens

Front Page Continuation (72) Inventor Masaaki Fujinawa 2880, Kozu, Odawara-shi, Kanagawa Hitachi Ltd. Odawara Plant (72) Inventor Shizuo Yamada 3-5-3 Sendagi, Bunkyo-ku, Tokyo Gran Eclair Sendagi 302 Yamada Medical Sharing Co., Ltd.

Claims (3)

[Claims] 1. An image input device having an image input unit for inputting an image of reflected light of an original image using a one-dimensional image sensor, wherein the image input unit transmits an image of transmitted light of the original image so that the image is transmitted. An image input device, wherein a surface light source for light is disposed, and further, the image input unit is provided with a shading correction means for performing at least two-dimensional shading correction of the surface light source. 2. The shading correction means comprises:
Means for pre-storing a sensor one-dimensional shading correction coefficient array for correcting sensitivity unevenness between pixels of the three-dimensional image sensor, and after storing the one-dimensional image sensor, the one-dimensional image sensor scans the surface light source and scan output thereof By performing correction with the sensor one-dimensional shading correction coefficient array, a two-dimensional shading correction coefficient array representing at least the two-dimensional irradiation unevenness of the surface light source is obtained.
The image input device according to claim 1, further comprising: a unit that stores a three-dimensional shading correction coefficient array separately from the sensor one-dimensional shading correction coefficient array. 3. An interval between sampling points in the sensor one-dimensional shading correction coefficient array is set to one pixel, and an interval between sampling points in the two-dimensional shading correction coefficient array depends on a rate of change in the shading amount of the surface light source. 3. The image input device according to claim 2, wherein the interval is sufficiently wider than one pixel.