JPS58200666A - Picture input device - Google Patents

Abstract

PURPOSE:To eliminate the influence of the sensitivity dispersion of a photoelectric converter by carrying out white level pursuit in vertical direction to the scanning direction of the photoelectric converter, and by providing a correction circuit and storage circuit to perform the level pursuit in the vertical direction. CONSTITUTION:A storage circuit 2 is provided with an address corresponding to the picture element unit of a photoelectric converter to perform level pursuit in direction vertical to scanning direction, and a value corrected by a correction circuit 4 is written to it. At first, the value of the circuit 2 is initialized to the threshold of white level, and, if the output of the photoelectric converter in a storage circuit 1 exceeds the value, the pursuit of the white level is started. At this moment, a normalization circuit 3 performs division to obtain a normalized output and supplies the normalized output to the circuit 4, which performs multiplication for obtaining a corrected value to write it in the circuit 2. Such process is carried out for every picture element of the photoelectric converter, and, as a result, the newest level of white level of each picture element is stored in the circuit 2. Namely, the reference white level independently performs the pursuit of black level for every picture element of the photoelectric converter.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image input device, and more particularly to an image input device equipped with a correction circuit suitable for image input using a -dimensional photoelectric converter.

Conventional image input devices scan a blank form in advance and use it as a reference to correct for uneven illuminance distribution due to the light source illuminating the form, uneven sensor sensitivity, and a decrease in the amount of light from the imaging lens. The above-mentioned non-uniformity was corrected by accumulating the data and comparing this accumulated data with the data of the reading area.

As a means for comparison, there are known methods such as dividing the two pieces of data or using a read-only memory to speed up the process. (Japanese Unexamined Patent Publication No. 55-112685) However, these human-powered devices have the following drawbacks. That is, one point of ijg is to scan a blank form in advance and accumulate reference data. In general, the reflectance of a blank form differs depending on the location due to the unevenness of the paper. Therefore, the data of blank forms scanned and stored as reference data is affected by the reflectance, and using this as a reference poses a problem in terms of usability. Furthermore, the second point is that the light source often fluctuates over time (such as ripples in the power supply in a short period of time), which is also a point of reference.

An object of the present invention is to eliminate the effects of variations in photoelectric converter output due to variations in sensitivity between elements and pixels of a photoelectric converter, changes in document illuminance due to the influence of lenses and light sources, An object of the present invention is to provide an input device that faithfully reproduces figures and the like.

In order to achieve this object, the present invention is characterized by the following configuration.

(1) In order to absorb the output fluctuations of the nine photoelectric converters mentioned above, the present invention tracks the white level on the form.

(2) The direction of tracking in (1) is perpendicular to the scanning direction of the photoelectric converter.

(3) It has a level correction circuit and a level tracking storage circuit for vertical level tracking.

FIG. 1 is a block diagram of an image input device using the principle of the present invention.

In Fig. 1, reference numeral 11 denotes a subject on which an image to be inputted is written 111'.
The reading section PQ on the top is illuminated. 11 is moved in the Y direction in FIG. 1 by a paper feed mechanism, which is also not shown. A -dimensional photoelectric converter 13 is provided in the direction (X direction) perpendicular to the movement direction of the main l! Image correction processing circuit 14 which is a part
There is no difference 9 from the photoelectric conversion unit in humans.

Hereinafter, the present invention will be explained in detail based on examples.

i@2 diagram! Ilm (elephant correction processing circuit 1) in Figure @1
FIG. 4 is a schematic diagram of one embodiment of FIG. In the present invention, a random access memory (RAM) is used as the storage circuit, but it is clear that it can be realized using other storage circuits.

1@2 In the figure, l is the first memory circuit that latches the data for each pixel after A/D conversion of the electrical signal from the one-dimensional photoelectric converter, and 2 is the scanning direction and vertical direction. a second storage circuit for storing the tracking level of the first and second
4 is a tracking level correction circuit that writes the result of the normalization circuit into the storage circuit 2, and 5 is an output register that latches the normalized data. -j, n indicates that the data width is n bits.

In order to perform level tracking in the direction perpendicular to the scanning direction, 2 is a memory circuit with an address corresponding to the Lu11 element position of the photoelectric converter, and the value corrected in 4 is stored. Initially, a value of 2 is initialized to the white level threshold and starts tracking the white level when the 10 photoelectric converter output exceeds this value. At this time, division by 3 is performed to obtain a normalized output, and the normalized output is given to the correction circuit 4, where the correction value obtained by performing multiplication is written into '2.

Such an operation is performed for each pixel of the photoelectric converter, and as a result, I! The latest level of white level for each Ll accumulation is stored.

In other words, since the white level, which is the standard for tracking the black level, tracks the black level independently for each pixel of the photoelectric converter, the influence of fluctuations in the output of the photoelectric converter due to variations in sensitivity between pixels of the photoelectric converter It is possible to obtain a normalized output that eliminates the The above will be explained using diagram #I3.

FIG. 3(a) shows the photoelectric converter output of the n-th scan, and FIG. 3(Φ) shows the gtn+i)-th output.

6.7.8 in FIG. 3(a) is the sensitivity variation between elements, and 9.10 in FIG. 3(Φ) is the black level. A case will be described in which the level in FIG. 3(a) is stored as the white level at 2 in FIG. 2 and the (n+1)th scan is tracked. In the (n+1)th scan, if the output of the photoelectric converter is as shown in Figure 3 (b) and is t, then by normalization the reference white level will be the same level in Figure 3 (a) and -) and normalized. The output is as shown in Figure 3 (C), and there is a shadow of sensitivity variation etc. 4
II can be eliminated.

Note that the memory circuit 1 in Figure $2 may not be provided.

FIG. 4 is a detailed diagram of the invention. 13 in Figure 4
is an m-bit one-dimensional photoelectric conversion 6 (sensor), and the l1ij image information converted into an electrical signal is amplified by the logarithmic amplifier 40 and converted into analog-digital conversion! 41 converts the signal into an In-bit digital signal and sets it in the register 42. 40 performs logarithmic amplification in order to simplify the multiplication and division described later, but it is clear that other methods may be used. 406 and 408 are storage circuits for level tracking. 43 is a subtraction circuit for normalization. (Since it is a logarithmic display, the function is division.) One input SA of 43 is obtained from the sensor, and the other input 8B is obtained from the tracking level register 412.
4 stores the result ((1+1) hits. 47 is an adder circuit f, which adds the black level tracking function F (B) to the normalized output obtained in 44 and outputs it to the output register 4B.
(B) is determined by the density of the black level of the image. 401
selects a bit among (n+1) bits according to the slice level preset to 410 in a read-only memory (ROM) and the normalized output of 48. 205 is an output register.

F (B) occurs frequently in the ROM 45 and the ψ raster 46.

The normalized output obtained in step 44 is converted into a white level tracking function F (W) by ROM 49 and temporarily stored in register 400. On the other hand, the tracking level before the first meal is in the register 412,
414t - set in register 403 via 14
12.414 is a register for timing adjustment. )
corrected to a new tracking level by adder circuit 402;
It is stored in a storage circuit 406 or 408 via a register 404.

Here, 411 is 406 or 4 via selector 415 etc.
This is a register that sets an initial value to 08, and is set by a processing device or the like. Reference numeral 415 denotes a selector which switches between the initial value and during operation, and although not shown, may be controlled externally or by including a control bit in the data 411.

Further above &l! Memory circuits 406, 408 and selector 40
9 functions as a shift register. i.e. 40
When 6 is in the write mode, 408 is in the continuous output mode, which is alternately switched.

FIG. 5 is an example of a white level tracking function. Although it is possible to use functions with different shapes other than those shown in FIG. 5, they are omitted because they are not the main purpose of the present invention.

In the jI5 diagram, the Y-axis is the input (the farther from the Y-axis, the level of the mill), and the Y-axis is the output. The output obtained by inputting the normalized output 44 is multiplied by the tracking level of the previous scan stored in 408 or 406 by A (addition since it is a logarithm), and the tracking level is corrected.

As explained above, according to the present invention, it is possible to soften the sensitivity variation between pixels of a photoelectric converter, and even if a device with adjacent bit variation of about 20% is used, this can be substantially reduced by 3 to 30%. It can be conveniently used at a concentration of 6%.

Therefore, although the amount of circuitry required for each pixel of the photoelectric converter increases, it is possible to lower the specifications of the photoelectric converter, which is overall more economical.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the overall configuration of an 1Ii1if image input device using the principle of the present invention, FIG. 2 is a block configuration diagram of an embodiment of the image correction processing circuit in FIG. 1, and FIG. A picture signal waveform diagram showing the effect, Fig. 4 is Fig. 2 t-realizing logic circuit w! A diagram showing an example of 1m formation, and a diagram WJ5, are diagrams showing an example of level tracking interfusion. 4...Correction circuit.

Claims (1)

[Claims] A picture in which a figure on a two-dimensional surface is converted into an electrical signal by a one-dimensional photoelectric converter and the quantized image data is manually input to an information processing device (in the human-powered device, the tracking level in the scanning direction and vertical direction of the photoelectric converter is A storage circuit that stores information for each pixel of the photoelectric converter, and a tracking level that is stored in the storage circuit.
a circuit that normalizes the output of the photoelectric converter; and a correction circuit that corrects the tracking level from the normalized output and the tracking level stored in the storage circuit, and sends the corrected tracking level to the storage circuit. An image input device whose percentage value is that it has a circuit.