JPH10283522A - Shading correction method of linear image sensor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method which does not need to adjust an offset level that corresponds to the offset charge value of an image sensor when a light source is turned off and also automates shading correction. SOLUTION: A reference white level of each pixel for one line that corresponds to reference white paper which becomes a reference is read and preliminarily stored in line memory part 51 when a light source is turned on, and next, an offset level for one line that corresponds to an offset charge value of each pixel of an image sensor 1 is read and preliminarily stored when the light source is turned off after the power source is turned on. Here, a correction coefficient value of each pixel is successively calculated based on the reference white level, the offset level and a white level correction target value that is preliminarily set and stored for one line. Next, an image on sheet paper one line each is read when the light source is truned on in a normal operation, the offset level is subtracted from each image level and correction value of each pixel is acquired by multiplying subtraction result value by the correction coefficient value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a shading correction method for an image sensor capable of automating shading correction, and is suitable for application to, for example, a bill discriminating apparatus.

[0002]

2. Description of the Related Art In a conventional image reading apparatus, light from a light source is radiated onto a sheet on which an image to be read is described, and reflected light or transmitted light is input to an image sensor, and the light is transmitted to the image sensor. The above image information is photoelectrically converted into an analog video signal and output. Next, after the analog video signal is A / D converted and converted into a digital value, processing such as shading correction and peak level correction is performed. here,
In the processing such as the shading correction and the peak correction, each coefficient is determined based on the reflected light of the white plate serving as the white reference at the start of the image reading on the paper sheet. Further, in the conventional image reading apparatus, when reading an image from a sheet, the white reference of the sheet is corrected for each sheet immediately after the reading is started. At this time, after detecting the leading end of the sheet, the light source is turned on. Next, the image sensor reads reflected light or transmitted light from a white plate for white reference correction, and performs shading correction based on the white reference.

[0003]

SUMMARY OF THE INVENTION However, LEDs
When the array was turned on, the output level of the image sensor was uneven due to the uneven light amount of the image sensor, the uneven light amount of the light source, the unevenness of the cell hook lens, and the like. On the other hand, when the LED array is turned off, the black level when it is completely dark is output with a specific offset value, so that it is necessary to perform offset adjustment corresponding to the black level. The present invention has been made in view of the above,
An object of the present invention is to provide an image sensor shading correction method that can eliminate the need for adjusting an offset level corresponding to an offset charge value of an image sensor when a light source is turned off, and that can automate shading correction. It is in.

[0004]

According to the present invention, there is provided an image reading apparatus which irradiates a sheet with light from a light source, reads and A / D converts the light using an image sensor, quantizes the read light, and outputs the read value. At the time of lighting, a white level reading step of reading a reference white level of each pixel corresponding to one line corresponding to a reference white sheet as a reference and storing the reference white level in advance, and at the time of turning off the light source after power-on, for each pixel of the image sensor. An offset level reading step of reading and storing in advance an offset level for one line corresponding to the offset charge value; and for each pixel based on the reference white level, the offset level, and a preset white level correction target value. A correction coefficient value calculating step of sequentially calculating the correction coefficient values for each line and storing the correction coefficient value for one line; By subtracting the offset level from the read each image level, and summarized in that and an image level correcting step of obtaining a correction value for each pixel by multiplying the correction coefficient value to the subtraction result value.

[0005]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram illustrating a configuration of the image sensor 1. As shown in FIG.
Is configured such that, for example, 1048 photoelectric conversion elements 13 are arranged in a line. The charges accumulated in the photoelectric conversion element 13 are transferred to the shift register 15 for even-numbered pixels and the shift register 17 for odd-numbered pixels via shift gates 19 and 21 by a shift pulse. The charges transferred to the respective shift registers 15 and 17 are sequentially sent to the output gate 23 by a transfer pulse. Output gate 23
, The outputs of the odd-numbered elements and the outputs of the even-numbered elements are alternately selected and output like P1, P2, P3,..., P10.

FIG. 2 shows a bill discriminating apparatus 31 to which the present invention is applied.
It is a figure showing an example of. As shown in FIG. 2, in the optical system of the banknote discriminating device 31, the irradiation light emitted from the light source 33 is reflected by the paper leaf 37, and the reflected light from the paper leaf 37 is condensed via the cell hook lens array 35. After that, the image sensor 1 is configured to accumulate electric charges in the photoelectric conversion element 13 for each pixel. As the light source 33, an LED array or a fluorescent tube having a function of supplying irradiation light to the paper sheet 37 is generally used. Since the LED array is configured by arranging about 30 LED elements in the main scanning direction, for example, the light amount unevenness is more conspicuous than that of a fluorescent tube, but is an indispensable component for reducing the cost of the apparatus.

The present invention is applicable to an image reading apparatus having an image scanner, and as a form of the scanner, for example, a document transport type scanner, a document table moving type scanner, a flat bed type scanner, a manual type scanner and the like are widely used. However, no matter which scanner configuration is used,
It goes without saying that the shading correction method for a linear image sensor according to the present invention is applicable. on the other hand,
In the mechanical system of the bill discriminating device 31, a stepping motor 39 receiving a drive pulse output from the motor control unit 47
Each time is rotated by one step, this rotation is transmitted to the roller 43 via the belt 41, and the paper sheet 37 is conveyed according to the rotation of the roller 43. The rollers 45a, 45b,
Reference numeral 45c denotes a mechanism configured to convey the paper sheet 37 following the rotation of the roller 43 and to assist the conveyance of the paper sheet 37, so that the description and illustration thereof are omitted here.

Next, the configuration of the electronic circuit system of the bill discriminating device 31 will be described. In FIG. 2, an A / D converter 49 performs analog / digital conversion on an output signal (analog) representing the amount of charge photoelectrically converted for each pixel by the image sensor 1 to convert the output signal into an output level (digital). The A / D converter 49 has a reference value VREF for determining how much the input analog signal is to be converted into a digital value.
+, VREF- are input. The line memory unit 51 includes:
This memory is rewritten only in the adjustment mode, and always stores an output level VWi for each pixel of a blank white paper as a reference. In addition, since the line memory unit 51 is backed up by the battery 53, the stored contents are retained even when the apparatus is turned off.

The CPU 55 controls the entire apparatus in accordance with a control program stored in an internal ROM.
Performs various operations related to shading correction. The RAM unit 57 is a volatile memory that stores a black level VDi for each pixel taken from the image sensor 1 when the power is turned on, and a correction coefficient Ki for performing shading correction. The dual-port RAM unit 59 stores a correction output Vi 'obtained as a result of performing shading correction on the output level Vi of each pixel taken in from the image sensor 1 and connects to an interface on the host computer side for high-speed operation. It is a buffer memory for enabling reading and writing from each direction to enable transfer.

The motor control unit 47 outputs a drive pulse for rotating the stepping motor 39 for each step. SW <b> 1 is a switch for supplying power Vcc to the LED array 33 to turn on the LED array 33, and is intermittently controlled by the CPU 55. Next, a shading correction method for the image sensor 1 will be described with reference to FIGS. FIG. 3 shows the output level of each pixel of the image sensor 1. First, the first step is a white level reading mode for reading a white level at the time of adjusting the apparatus. From the opening 61 of the bill discriminating device 31, the paper 37
When the reference blank paper is inserted, the reference blank paper is sequentially conveyed into the apparatus according to the rotation of the roller 43. At this time,
An ON signal for turning on the light source 33 is applied to SW1 from the CPU unit 55, and the light source 33 is turned on. Next, as shown in FIG. 3A, the image sensor 1 outputs an output signal corresponding to the charge amount of the reference white paper for each pixel. A
In the / D converter 49, the output signal output from the image sensor 1 is sequentially subjected to A / D conversion and quantized, and the white level VWi of the reference blank sheet for one line is stored in the line memory 51 via the image data bus 50. Remember. Since the line memory unit 51 is backed up by the battery 53, the memory contents are stored even when the power is turned off.

The second step is a black level reading mode for reading a black level when the power is turned on. CPU unit 5
5 is an SW1 signal for turning off the light source 33.
And the light source 33 is turned off. Next, as shown in FIG. 3A, the image sensor 1 outputs, for each pixel, an output signal corresponding to the charge amount when the light is turned off. A / D
In the converter 49, the output signal output from the image sensor 1 is sequentially subjected to A / D conversion and quantized, and the black level V for one line is stored in the RAM unit 57 via the image data bus 50.
Remember Di. Next, the third step is a correction coefficient calculation mode for calculating the correction coefficient Ki using the apparatus.
The CPU 55 calculates a correction coefficient Ki for each pixel based on the white level VWi stored in the line memory 51, the black level VDi stored in the RAM 57, and a preset white level correction target value V0. Are sequentially calculated as follows, and the calculation result is stored in the RAM unit 57. Ki = V0 / (VWi-VDi) (Equation 1) (0≤i≤192)

The fourth step is a normal operation mode in which shading correction is performed for each read pixel by using the apparatus and output. During normal operation, the opening 6 of the bill discriminating device 31
When a paper sheet on which a character or an image is written is inserted as the paper sheet 37 from 1, the paper sheet 37 is sequentially conveyed into the apparatus according to the rotation of the roller 43. At this time, an ON signal for turning on the light source 33 is applied to SW1 from the CPU unit 55, and the light source 33 is turned on. Next, the image sensor 1 outputs an output signal corresponding to the charge amount at the time of lighting for each pixel. In the A / D converter 49, the image sensor 1
A / D conversion is sequentially performed on the output signals output from the A / D converter to quantize the output signals.
The output level Vi for the line is stored. Next, the CPU unit 55 sequentially obtains a correction output Vi 'for the output level Vi stored in the RAM unit 57 by the following formula. Vi '= Ki (Vi-VDi) (Equation 2) At the same time, the CPU 55 sequentially stores the corrected output Vi' in the dual port RAM 59. Thus, the output level V for each pixel taken in from the image sensor 1
The corrected output Vi ', which is the result of performing shading correction on i, is stored in the dual-port RAM unit 59, and is connected to the interface of the host processor via the dual-port RAM unit 59 to enable high-speed transfer. Can be

The black level VDi used in the present invention is
Is the lower limit of the reference voltage of the A / D converter 49 (V
REF-) must always be set higher. As described above, when the light source is turned on, the reference white level of each pixel for one line corresponding to the reference white paper as a reference is read and stored in advance, and then, when the light source is turned off after the power is turned on, each pixel of the image sensor is turned off. The offset level for one line corresponding to the offset charge value is read and stored in advance. Here, a correction coefficient value for each pixel is sequentially calculated based on the reference white level, the offset level, and a preset white level correction target value and stored for one line. At the time of lighting, the sheet is read line by line, the offset level is subtracted from each image level, and the subtraction result value is multiplied by a correction coefficient value to obtain a correction value for each pixel. In this case, the adjustment of the offset level corresponding to the offset charge value of the image sensor can be made unnecessary, and the shading correction can be automated. Therefore, the light amount unevenness for each pixel of the image sensor,
It is possible to correct variations in the output level of each pixel of the image sensor, which are caused by unevenness of the light amount of the light source, unevenness of the cell hook lens, and the like. Further, the dynamic range of the output signal from the image sensor can be expanded, and the sensitivity variation of the image sensor can be adjusted for each pixel. In addition, since the calculation formula of the correction coefficient value including the division in advance is calculated for each pixel, it is possible to avoid a calculation requiring a large processing time such as division during a normal operation, and to reduce the processing speed of the apparatus. Can be improved.

[0014]

As described above, according to the present invention, when the light source is turned on, the reference white level of each pixel of one line corresponding to the reference white paper as a reference is read and stored in advance, and When the light source is turned off after the power is turned on, the offset level for one line corresponding to the offset charge value for each pixel of the image sensor is read and stored in advance. Here, a correction coefficient value for each pixel is sequentially calculated based on the reference white level, the offset level, and a preset white level correction target value and stored for one line. At the time of lighting, an image on a sheet is read line by line, an offset level is subtracted from each image level, and the subtraction result value is multiplied by a correction coefficient value to obtain a correction value for each pixel. In addition, it is not necessary to adjust the offset level corresponding to the offset charge value of the image sensor when the light source is turned off, and the shading correction can be automated.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of an image sensor 1;

FIG. 2 is a diagram illustrating an example of a bill discriminating apparatus to which the present invention is applied.

FIG. 3 shows an output level of each pixel of the image sensor 1;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image sensor 31 Banknote discriminating device 33 Light source 37 Paper sheet 49 A / D converter 51 Line memory 55 CPU section 57 RAM section 59 Dual port RAM section

Claims (1)

[Claims] 1. An image reading apparatus which irradiates a light source light onto a paper sheet, reads and A / D converts the read light using an image sensor, quantizes the read light, and outputs the read value, wherein a reference white paper as a reference when the light source is turned on. A white level reading step of reading a reference white level of each pixel corresponding to one line and storing the reference white level in advance, and one line corresponding to an offset charge value of each pixel of the image sensor when the light source is turned off after power is turned on. An offset level reading step of reading and storing an offset level for each minute, and sequentially calculating a correction coefficient value for each pixel based on the reference white level, the offset level, and a preset white level correction target value. A correction coefficient value calculating step of storing one line and reading the image on the sheet for each line when the light source is turned on in the normal operation. Subtracted et the offset level, shading correction method for a linear image sensor and having an image level correcting step of obtaining a correction value for each pixel by multiplying the correction coefficient value to the subtraction result value.