JPH03184469A - Black reference level correcting device - Google Patents

Abstract

PURPOSE:To correct a black reference level without fail by holding picture element data in the non-incident area of an image sensor separately for the black reference level of an odd-number-th picture element and the black reference level of an even-number-th picture element and selectively outputting the black reference levels corresponding to the odd-number-th and even-number-th picture elements when reading picture data in an effective picture element area. CONSTITUTION:A CCD 101 executes photoelectric conversion to original picture information. Picture data AV(102-a), to which photoelectric conversion is executed by the CCD, are inputted to first and second black reference level holding circuits 103 104 and by utilizing the dummy bit of the CCD, the level of the odd-number-th picture element is sampled and held. For a first black reference level signal 103-a, the level of the even-number-th picture element is sampled and held and a second black reference level signal 104-a is generated and synthesized by a selector 107 so as to obtain a black reference level signal 107-a in the case of A/D conversion. Then, odd-number-th and even-number-th picture element sample signals 105-b and 105-c and a select signal 105-d of a selector 107 are outputted from a timing generator 105.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a black level correction device for an image sensor in a document reading section of an image scanner device, a facsimile device, or the like.

[Prior art] A conventionally known image sensor (for example, CCD)
In , image data in a non-incident light area of a CCD was sampled, and the average level thereof was used as a black reference level when quantizing image data in an effective image area.

[Problems to be Solved by the Invention] However, in the above conventional example, when the black reference level is generated by the pixel separation output method often used in high-speed driving CCOs, the odd-numbered black pixel level of the pixel separation output and the even-numbered Since there is a level difference between the black pixel level of the odd-numbered pixel and the even-numbered pixel, if you calculate the average to determine the black reference level for quantization, the black reference level of the odd-numbered pixel and the even-numbered pixel will not match. Is there tlA on the output image? There is a drawback that R occurs.

The above drawbacks will be explained in more detail with reference to the drawings.

The internal configuration of a CCD using the pixel separation output method is as shown in FIG. That is, in this pixel separation method, odd-numbered pixels and even-numbered pixels are processed in separate systems, respectively, so that the charges received by the photodiode 101-1 and photoelectrically converted can be outputted at high speed. Charges photoelectrically converted from odd-numbered pixels and even-numbered pixels are stored in storage electrode 0 (101-5).

e (101-2). Charges corresponding to image information of each pixel for one line are stored in each storage electrode.

These charges are shifted to the shift gate 0 (1
01-6) and e (101-3), and analog shift registers o (101-7) and e (101
-4). The charges sent to these analog shift registers 0 and e are sequentially transferred by transfer signals φO and φE, and are transferred to amplifiers o (101-8) and e (
101-9) and output as analog data Vo of odd-numbered red pixels and analog data Ve of even-numbered pixels. These Vo and Ve are combined by a combining circuit 102 to form image data AV for one line.

In this way, odd-numbered pixel data and even-numbered pixel data are output from different processing units, so data with different levels of Ve and Vo are output alternately, as shown in Figure 3. In Figure 3, which shows the 6 figures that will be stored, the CCD
Ve and Vo in the non-incident light bit (dummy bit) of
This shows the level difference between the two levels. From the image levels Ve and vO of these dummy bits, the average value is calculated as follows:
This corresponds to the two-dot chain line (3-a) in the figure.

Therefore, if this level is used as a black reference level when quantizing image information, a difference will occur in the black level between odd-numbered pixels and even-numbered pixels, and this difference will result in shading in the output image. There was a problem with it appearing.

Therefore, in view of the above-mentioned points, an object of the present invention is to provide a black level correction device configured so that no shading occurs in an output image.

[Means for Solving the Problems] The present invention provides a black reference when reading image information using an image sensor and quantizing image data by maintaining a pixel level in a non-incident light area of the image sensor. In the correction device for adjusting the levels of odd and even pixels in the non-incident light area, first and second pixel levels are adjusted.
first and second black reference level holding means that hold the first black reference level as a black reference level when quantizing odd-numbered image data in the effective image area of the image sensor; When quantizing even-numbered pixel data in the effective image area, the second
control means for setting the black reference level of the black reference level to the black reference level.

[Function] In the present invention, pixel data in the non-incident light area of the image sensor is held separately in the black reference level of odd-numbered pixels and the black reference level of even-numbered pixels, and When reading image data, by selectively outputting the black reference level corresponding to odd-numbered pixels and even-numbered pixels, the black reference level can be reliably corrected when using a pixel-separated output type image sensor. be able to.

[Example] Hereinafter, the present invention will be described in detail with reference to the drawings.

Figure 'fJ1 is a block diagram showing one embodiment of the present invention. Here, 101 is a CCD that photoelectrically converts original image information.
, 102 is an amplifier that amplifies the charge-converted image data, 103 is a first black reference level holding circuit that holds odd-numbered pixel data of dummy bits of the CCD, 104
107 is a second black reference level holding circuit that holds even-numbered pixel data of dummy bits of the CCD; 107 is a first black reference level signal 103-a output from the first black reference level holding circuit 103; 2 black reference level holding circuit 10
108 is an A/D converter; 108 is an A/D converter;
105 is a timing generator, and 108 is a circuit unit that generates a white reference level during A/D conversion.

Image data AV (102
-a) is the first black reference level holding circuit 103, the second
The black reference level holding circuit 104 samples and holds the odd-numbered pixel level using a dummy bit of the CCD, and outputs the first black reference level signal 103-a and the even-numbered pixel level. Hold the sample and hold the second
The selector 10 generates a black reference level signal 104-a of
Black reference level signal 107 when synthesized in step 7 and converted to /D
- get a. Sample signals 105-b and 105-c of odd-numbered and even-numbered pixels and a select signal 105-d of the selector 107 are output from the timing generator 105.

FIG. 2 shows the relationship between one line of CCD data and each control signal output from the timing generator 105. The illustrated shift pulse S)+(105-a
) from one falling edge to the next falling edge (one cycle) is 1
corresponds to the line. The first 32 bits in one line of the CCD are dummy bits, and this portion is used to generate a black reference level. Thereafter, the image of the original is read using 5000 bits as an effective image area. The last 4 bits are also dummy bits. Here, the output of ^V is shown when a white uniform image is read.

In FIG. 2, the ODD signal (105-b) has the level Vo[1,3°5, . . .
...31], and is a signal for sampling the analog switch 103- of the first black reference level holding circuit 103.
1 is 0N10FFL/. The sampled data is held as a first black reference level in a subsequent hold circuit and output as a signal 103-a.

Similarly, the EVEN signal 105-c is also at the even-numbered pixel level Ve[2, 4, 6, . . . 32
], and the analog switch 104-1 of the second black reference level holding circuit 104 is set to 0N10F.
I'm doing F. The sampled data is held as the second black reference level in the subsequent hold circuit, and the signal 10
It is output as 4-a.

Next, generation of the black reference level will be explained in more detail with reference to FIG. Figure 3 shows the dummy bit section 1 in Figure 2.
This is an enlarged view of the portion from the 6th pixel to the 6th pixel.

Here, the odd pixel level Vo and the even pixel level Ve are different. The black reference level of the odd-numbered pixel sampled and held by the EVEN signal and the ODD signal and the level of the even-numbered black pixel are indicated by the dashed line (103-a) and the dotted line (103-a).
It will look like 04-a). Regarding the black reference level of odd-numbered pixels and the black reference level of even-numbered pixels generated in this way, in the effective image area, as shown in FIG.
When reading an odd-numbered effective pixel, the black reference level of the odd-numbered pixel is set by the select signal SEL (105-d).
When reading even-numbered effective pixels, the black reference level of the even-numbered pixel is selected as the black reference level of the A/D converter.

FIG. 4 shows how the black reference level for odd-numbered pixels and the black reference level for even-numbered pixels are selected. FIG. 4 is an enlarged view of the effective image area from the 33rd pixel to the 38th pixel. At the 33rd, 35th, and 37th pixels, the SEL signal becomes “H” and is output from the first black reference level holding circuit 103. Then, the black reference level signal 103-a of the odd-numbered pixel is selected.

In addition, the SEL signal is “L” in the 34th, 36th, and 38th pixels.
'', and the black reference level signal 104-a of the even-numbered pixel output from the second black reference level holding circuit 104 is selected.

That is, in one embodiment of the present invention, the level of the even-numbered pixel of the dummy bit and the level of the odd-numbered pixel are separately held as black reference levels, and when reading the effective image pixel, the level of the odd-numbered pixel is By selectively outputting the corresponding black reference level for each even-numbered pixel,
It is possible to correct the black reference level during ^/D conversion.

Note that although a 5000-bit CCD is used in the above embodiment, it goes without saying that the number of bits does not matter. Further, the present invention can be similarly applied to various image sensors such as a touch sensor instead of a CCD.

[Effects of the Invention] As explained above, according to the present invention, pixel data in the non-incident light area of the image sensor is divided into the black reference level of odd-numbered pixels and the black reference level of even-numbered pixels. When holding and reading image data in the effective pixel area,
By selectively outputting black reference levels corresponding to odd-numbered pixels and even-numbered pixels, it is possible to reliably correct the black reference level when using a pixel separation output type image sensor.

[Brief explanation of drawings]

Figure 1 is a block diagram showing the configuration of an embodiment of the present invention, Figure 2 is a diagram showing the relationship between output data for one line of the image sensor and control signals, and Figure 3 is a diagram showing the non-incident light of the image sensor. An explanatory diagram when generating a black reference level using a region, Fig. 4 is a diagram showing the process of performing correction by switching the black reference level signal between odd-numbered pixels and even-numbered pixels, and Fig. 5 It is an internal configuration diagram. ...CCD. ...Amplifier, ...First black reference level holding circuit, ...Second black reference level holding circuit, ...Timing generator, ...＾0 converter, ...Selector.

Claims (1)

[Claims] 1) When reading image information using an image sensor,
In a correction device that maintains a pixel level in a non-incident light region of the image sensor and uses it as a black reference level when quantizing image data, odd-numbered and even-numbered pixel levels in the non-incident light region are first and first and second black reference level holding means that hold the first black reference level as a second black reference level; 1. A black reference level correction device, comprising: a control means that uses the second black reference level as a black reference when quantizing even-numbered pixel data in the effective image area.