JPH05227427A - Shading correction system - Google Patents

Abstract

PURPOSE:To realize highly precise and minute shading correction and to reduce the trouble of a fine adjustment process for the amplifier of a sensor. CONSTITUTION:A multilevel quantization picture signal for plural sample density values is obtained, and an actual characteristic is estimated from plural pieces of measurement point information. Thus, a shading correction control part 7, an average value calculation part by individual elements 8, a measurement function estimation part 10, a measurement function information memory 11, a standard function information memory 21 and a sample density information memory 9 are newly provided. Thus, the dynamic range of a density area to be noticed can be taken large since a photoelectric conversion characteristic can be defined by a non-linear function.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a shading correction system used in a document reading device such as a scanner,
In particular, it relates to a system for correcting a digital multi-value quantized image signal after A / D conversion.

[0002]

2. Description of the Related Art In inputting an image, the level of the peripheral portion may be lowered or the gray level may not be uniform over the entire screen due to the characteristics of the image pickup device or the influence of lens aberration. This is the shading phenomenon.
Conventionally, a shading for removing this shading phenomenon
As the Ding correction method, there is a process of correcting the photoelectric conversion output itself with reference to a reference white original document, and a process of correcting the threshold value as well (for example, "Electronic Information and Communication Handbook").
Second Volume, March 30, 1988, published by OHMU Co., Ltd., pp.26
05). By the way, a scanner or the like for reading a form by using a photoelectric conversion function includes a sensor unit having a function of converting light reflected from the form into an image signal and a function of a signal amplification amplifier for amplifying the image signal, and a digital image signal processing unit. The multi-value quantized image signal is converted into an A / D converter, and the multi-value quantized value is stored in the image memory.

In such a scanner, the characteristics of the light receiving element of the sensor section usually vary due to changes with time. To absorb the difference of the variation,
The correction value for the original characteristic with respect to the actual characteristic is obtained, and the shading correction section corrects the multi-value quantized image signal after A / D conversion based on the value. Figure 3
It is a conceptual diagram of conventional shading correction. In Figure 3,
The characteristics of one element of the sensor unit are shown, where the horizontal axis X represents the density value and the vertical axis Y represents the multi-value quantized value. The standard function R (X) is the original characteristic, and the measurement function S (X) is the actual characteristic. Of the two inclined straight lines in FIG. 3, the lower part shows the characteristics of the measurement function S (X), and the upper part shows the characteristics of the standard function R (X). Therefore, the value of Yout with respect to the multi-valued quantized value Yin is the shading correction value. Here, the measurement function S (X) is a measurement point (X which is defined by the multi-value quantized value Ys measured at a certain sample concentration value Xs.
s, Ys). In the conventional method, measured values (Xs, Ys) at one sample concentration
And a straight line passing through the point (Xi, 0) where R (X) = 0
Was expressed.

[0004]

In the above-mentioned conventional method, S (Xi) = 0 is not always satisfied due to the fluctuation of the offset level. In addition, the change in the multilevel quantization level cannot be approximated by one straight line in all the density value ranges. Therefore, there is a problem that the estimation accuracy of the measurement function S (X) is not always sufficient. In addition, S (X
In order to guarantee i) = 0, it can be solved by adjusting the offset level of the signal amplification amplifier, but for that purpose, a delicate adjustment process of the sensor amplifier is required. This adjustment has a problem that it is extremely troublesome. The object of the present invention is to solve such conventional problems,
A shading correction system in which the photoelectric conversion characteristic, that is, the estimation accuracy of the measurement function S (X) can be improved to finely perform the shading correction, and the signal amplification amplifier can be adjusted without trouble. To provide.

[0005]

In order to achieve the above object, the shading correction system of the present invention includes a sensor section (2) for converting reflected light from a form to be read into an image signal and amplifying the image signal. , An A / D conversion unit (3) for converting an image signal into a digital multi-valued quantized image signal, a correction value of the multi-valued quantized image signal is stored, and the stored value is stored using the multi-valued quantized image signal as an address. In a shading correction system including a shading correction memory (4) for reading and an image memory (5) for storing corrected multi-valued quantized values, each unit for determining the shading correction value And a shading correction control unit (7) that prompts the operator to insert a form for shading correction, and an average value of the multi-value quantized image signal for each sensor unit element and sample density value. Calculate and calculate An element-by-element average value calculating section (8) for storing values, and a sample density information memory (9) for storing in advance information on a plurality of sample densities for measuring a multilevel quantized image signal for shading correction. , The average value of the multi-valued quantized image signal at each sample density value in a certain element from the element-by-element average value calculation section, and each sample density information from the sample density information memory, and based on these, the apparatus Measurement function estimation unit (10) for estimating photoelectric conversion characteristics, measurement function information memory (11) for storing measurement function definition information from the measurement function estimation unit, and pre-stored standard photoelectric conversion function function definition information However, it is characterized in that it is provided with a standard function information memory (12) which outputs a multi-valued quantized value serving as a shading correction value, using the content of the measurement function information memory as an address.

[0006]

In the present invention, information on a plurality of sample concentration values is stored, multi-value quantized values corresponding to these sample concentration values are acquired, and the measurement function S (X) is obtained from a plurality of measurement points.
Can be estimated. As described above, since it is possible to obtain multi-valued quantized values for a plurality of sample concentrations, the measurement function S (X) can be estimated with high accuracy, and fine shading correction can be performed. become.

[0007]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a block diagram of a shading correction system showing an embodiment of the present invention. In FIG.
1 is a form to be read, 2 is a sensor for converting the reflected light from the form 1 into an image signal, 3 is an A / D conversion unit for converting the image signal into a digital multi-value quantized image signal, and 4 is a share
Memory for shading correction, 5 an image memory unit, 6 an apparatus control unit, 7 a shading correction control unit, 8 an average value calculation unit for each element, 9 a sample concentration information memory, 10 a measurement function estimation unit, 11 Is a measurement function information memory, and 12 is a standard function information memory. In this embodiment, the multilevel quantized signal is set to 6
Although the description is given with four gradation values, the quantization level can be extended and applied even when the value N is an arbitrary integer.
First, the sensor 2 converts the reflected light a from the form 1 into the image signal b.
After the conversion, the converted image signal b is converted into the A / D conversion unit 3
Send to. The A / D converter 3 converts the analog image signal b into a digital multivalued quantized image signal c and sends the converted output to the shading correction memory 4. The image signal c is sent to the element-by-element average value calculating section 8 only in the shading correction value determination mode, and in the normal reading mode. However, the element-by-element average value calculation unit 8 does not receive this. The shading correction memory 4 stores the corrected signal value for each multi-value quantized image signal, and the multi-value quantized image signal c
Are respectively stored in the areas having the address. The correction value having the address of the multi-valued quantized image signal c sent under the control of the shading correction control unit 7 is used as a shading value.
It is read from the memory 4 for Ding correction. This value is
The multi-valued quantized image signal c after the Ding correction, and the corrected signal c is stored in the image memory 5. The above is the operation in the normal reading mode.

Next, the operation in the mode for determining the shading correction value will be described. First, by inputting the shading correction instruction information d from the device control section 6 to the shading correction control section 7, the shading correction control section 7 outputs the shading correction memory erase signal k. The data in the correction memory 4 is erased by sending the data to the correction memory 4. Shading correction controller 7
Prompts the operator to insert the shading correction form. In the present invention, the shading correction form is required for each sample concentration determined in advance by the system, and the instruction to the operator specifies the sun concentration. In the present embodiment, the number of sample concentrations will be described as four, but the number is not limited to this number, and an arbitrary integer M can be expanded as the number of sample concentrations.

Correction form 1 which is a sample of the first density
Is inserted, the multi-valued quantized image signal c passes through the sensor unit 2 and the A / D conversion unit 3 and the element-based average value calculation unit 8
Sent to. The element-based average value calculation unit 8 acquires the multi-value quantized image signal c from the A / D conversion unit 3 for the inserted shading correction form 1 for several lines, and calculates the element-based average value. And store this. Next,
The Ding correction control unit 7 prompts the operator to insert a form for the next sample concentration. In this way, the above operation is repeated for all sample concentrations. The measurement function estimation unit 10 receives the designated element average value f from the element-based average value calculation unit 8 and the sample concentration information g stored in the sample concentration information memory 9, and calculates the measurement function definition information h by the method described later. And output. Next, the measurement function information memory 11 stores the measurement function definition information h from the measurement function estimation unit 10. The stored data stores the density value at that time by using the multilevel quantized value as an address. In the standard function information memory 12, the information defining the original photoelectric conversion characteristics determined in advance by the system stores the multi-value quantized value at that time using the density information as an address. Therefore, the standard function information memory 12 receives the contents from the measurement function information memory 11 as the reference address i, and the value stored at that address as the correction information j, the part of the corresponding element of the shading correction memory 4. Output to.

FIG. 2 is a characteristic curve diagram showing the operation of the measurement function estimator in FIG. A method of estimating a measurement function in a certain device and a method of calculating measurement function information will be described with reference to FIG. As described above, the measurement function estimating unit 10 sets the four points (X1, Y1) (X2, Y2) (X
3, Y3) (X4, Y4) to approximate the measurement function. Many methods already exist for the approximation, but any of these methods can be implemented in the present invention. Here, a method of obtaining a line segment connecting two adjacent points will be described. That is, as shown in FIG. 2, points (X1, X
2) and the point (X2, Y2) are connected by a line segment, and the point (X2, Y2
2) and the point (X3, Y3) are connected by a line segment, and the point (X3
3, Y3) and the point (X4, Y4) are connected by a line segment. Note that X
When 1 is not 0 and Y4 is not 0, the line segment connecting (X1, Y1) (X2, Y2) is changed to (X3, Y3) (X4, Y4) until X becomes 0. Extend the connecting line segment until Y becomes 0.

In the function estimated by the above method, the expression in the form of the matrix of Y when X is sequentially increased from 0 becomes the measurement function information. The finer the increment width of X, the more accurate the correction can be. As is clear from FIG. 2, the maximum value of X is a value at which Y becomes 0. Further, Y is an integer and has a value of 0 to 63. However, X where S (X) = Yi
There may be more than one, but at that time X when Yi
Average of maximum and minimum values of standard function definition information memory 1
It is necessary to obtain this value X so that it exists as the value of X that becomes the address of 2. As described above, since the multi-value quantized image signal for a plurality of sample density values is acquired by the input from the device control unit 6 and the actual characteristic is estimated from the plurality of measurement point information, the actual characteristic is highly accurately determined. It can be estimated with. As a result, the correction accuracy is high, and it is possible to perform the fine shading correction.

[0012]

As described above, according to the present invention,
Since the estimation accuracy of the measurement function can be improved by increasing the number of sample concentrations, it is possible to determine a more accurate and fine grained shading correction value, and as a result, obtain a high quality image. Is possible. Further, by using the offset density as a sample, it is not necessary to adjust the amplifier of the photoelectric conversion unit during operation, so that the labor of maintenance can be reduced. Furthermore, since the nonlinear photoelectric conversion characteristic is expressed from a plurality of measurement points, it is possible to expand the dynamic range of the concentration region of interest.

[0013]

[Brief description of drawings]

FIG. 1 is a block diagram of a shading correction system showing an embodiment of the present invention.

FIG. 2 is a characteristic curve diagram showing an operation of a measurement function estimation unit in FIG.

FIG. 3 is a characteristic curve diagram showing an operation of conventional shading correction.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Form 2 Sensor part 3 A / D conversion part 4 Shading correction memory 5 Image memory 6 Device control part 7 Shading correction control part 8 Average value calculation part by element 9 Sample concentration information memory 10 Measurement function estimation part 11 Measurement function information memory 12 Standard function information memory a Reflected light b Image signal c Multi-value quantized image signal d Shading correction instruction signal e Sample concentration designation information f Designated element average value g Sample concentration information i Reference address j Correction value information k Shading correction memory erase signal

Claims (1)

[Claims] 1. A sensor unit for converting reflected light from a form to be read into an image signal and amplifying the image signal, an A / D conversion unit for converting the image signal into a digital multi-value quantized image signal, and A shading correction memory for storing a correction value of the value-quantized image signal and reading the stored value using the multi-valued quantized image signal as an address; and an image memory for storing the corrected multi-valued quantized value. In the shading correction system, a shading correction control unit that controls each unit to determine a shading correction value and prompts the operator to insert a shading correction form, and an element of the sensor unit. An average value calculator for each element that calculates the average value of the multi-valued quantized image signal for each different sample density value and stores the calculated value, and a plurality of multi-valued quantized image signals that measure the multi-valued quantized image signal for shading correction. Sun A sample density information memory in which pull density information is stored in advance, the average value of the multi-value quantized image signal at each sample density value in a certain element from the element-specific average value calculation unit, and the sample density A measurement function estimation unit that receives each sample concentration information from the information memory and estimates the photoelectric conversion characteristics of the device based on these, a measurement function information memory that stores the measurement function definition information from the measurement function estimation unit, and A standard function information memory that stores function definition information of standard photoelectric conversion characteristics, and outputs a multi-valued quantized value that is a shading correction value, using the content of the measurement function information memory as an address, is provided. A characteristic shading correction system.