JPH02254864A - Image input device - Google Patents

Abstract

PURPOSE:To improve the deterioration in picture quality causing a difference in each sensor output by measuring the gamma characteristic of each of plural image sensor elements in use prior to the picture input of an original, indexing a correction curve corresponding to each image sensor and making the correspondence to the relation. CONSTITUTION:A scan timing generator 9 sets a scan period of an image sensor 1 and a data subject to digital input conversion by an A/D converter 3 is written in a line memory 4. Then a CPU 6 reads the content of the line memory 4 to sample the data by a decided number for each sensor chip and the average is obtained by each chip to obtain an output level of each chip. Then the CPU selects a curve correcting the sensitivity characteristic depending on the sensitivity characteristic and outputs the number of the curve in a binary number and the result is outputted to latches 7-1 to 7-N. Thus, since the sensitivity of each image sensor is corrected, in the case of reading an intermediate tone original, the read density of each sensor is made equal to improve the deterioration in the picture quality.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a technology for improving the quality of input characteristics of halftone images in image scanners used for desktop publishing and the like.

[Conventional technology]

Various types of image sensors such as tubes and semiconductors are used as image sensors for image input devices, but in both cases, the amount of light received and the sensor output do not have a complete one-to-one correspondence. A non-linear relationship between quantity and sensor output existed. This gamma characteristic may vary slightly between individual parts even if the sensor is of the same type or type. There is a difference. Due to differences in gamma characteristics between multiple image sensors, even if each sensor output is set to output the same output at the reference white level,
When an image of a gray original is input as an intermediate tone, a difference occurs in the output of each sensor, and data corresponding to the same gray cannot be obtained.

In order to correct such gamma characteristics, there is a method of first measuring the gamma characteristics of each of the image sensors constituting the image input device, and then normalizing them using a corresponding correction curve to equalize the outputs of each. According to this method,
The characteristics of each image sensor must be measured and managed individually before the device is assembled, leading to an increase in manufacturing costs.

As an image sensor technology related to this matter, there is, for example, Japanese Unexamined Patent Publication No. 161580/1983.

[Problem to be solved by the invention]

The conventional technology has the above-mentioned problems.

The purpose of the invention is to improve the deterioration in image quality that occurs when inputting a uniform halftone image due to the difference in gamma characteristics of the individual image sensors in an image input device composed of a plurality of image sensor elements. be. In particular, when correcting gamma characteristics, the purpose is to perform automatic correction prior to inputting an original original image without actually reading an original with halftone density.

[Means to solve the problem]

As a method for correcting the gamma characteristics of an image sensor used in an image input device, the present invention provides a circuit and a program for measuring the gamma characteristics of each of a plurality of image sensor elements used prior to inputting an image of an original original. The purpose of this method is to prepare correction curves for each image sensor, and use these to determine and associate correction curves corresponding to each image sensor.

When inputting image data of an actual form, each image sensor output is normalized through a correction curve, so that even if a halftone image is input, uniform data can be input.

To measure gamma characteristics, it is common to read a blank sheet of paper that represents the white level that should be the standard, and a gray sheet of paper that represents each halftone. By reading with , a reference output corresponding to the white level is obtained. In order to provide a change in the amount of light corresponding to an intermediate tone, the fact that the image sensor output corresponds to the amount of received light x scan period (exposure time) is utilized, and instead of providing a change in the amount of received light, a change in the scan period is provided.

For example, a sensor output equivalent to giving 1/2 the amount of light can be obtained by keeping the amount of light the same and reducing the scan period to 1/2. By utilizing this phenomenon, the sensor output on New Year's Day of 1/α can be obtained with a period of 1/α. In this way, the sensor output corresponding to the reference white level and the sensor output corresponding to the intermediate tone can be obtained, and based on these data, a gamma correction curve is calculated for each image sensor element.

[Effect]

The feature of the present invention, the function of changing the amount of light in an economical way by changing the scanning cycle of the image sensor, is constructed from an integrated circuit and a simple rough 1-ware program. The scan period uses a counter circuit to generate a period corresponding to the reference period and 1/α period. The counter is usually a digital quantity counter, and the quantity equivalent to 1/a is N/M.
(M, N are integers). If N and M are set to large values, N/M can be brought as close as possible to 1/α. N
By creating a scan period corresponding to 1/α by selecting , M, the image sensor can be set to a light intensity of 1/a, and by A/D converting the image sensor output, the light intensity can be reduced to 1/a. The sensor output value corresponding to /α is determined. By taking several values of α (actually, values of M and H), the software program determines the gamma characteristic suitable for the corresponding sensor. A software program calculates a curve that corrects this gamma characteristic.

Prepared in the memory circuit. By using this in-memory correction curve, the gamma characteristics of each image sensor element are corrected, so that all image sensor elements output the same output even when a halftone image is input.

〔Example〕

An embodiment of the present invention will be described below using screens.

In Fig. 1, 1 is an image sensor (contact type or reduced type can be used) consisting of multiple chips, 2 is an amplifier that amplifies the output of the image sensor, and 3 is an amplifier that converts the output of amplifier 2 into A.
/D conversion A/D converter, 4 is image sensor 1
The output for one scan is A/D converter 3.
A line memory stores the D-converted output, 5 is a line memory address generator that generates an address when writing the output of the A/D converter 3 to the line memory 4, and 6 is a gamma and arithmetic unit for the contents of the line memory 4. Processor (CPU) that selects a correction curve, 7-1~
7-N is a latch where the gamma correction curve number is set from the CPU 6, 8 is a read-only memory (ROM) for gamma-correcting the image sensor output by the output from latches 7-1 to 7-N, and 9 is an image Setting the scan cycle of sensor 1, line memory address generator 5
A scan timing generator outputs a pixel clock to the A/D converter 3, a conversion clock to the A/D converter 3, and outputs an output enable signal to the latches 7-1 to 7-N. This is a binarization circuit that binarizes data.

Next, the operation of this embodiment will be explained with reference to FIGS. 1 to 5.

First, as shown in FIG. 2, the reason why the read density changes from chip to chip due to differences in sensitivity characteristics (gamma characteristics) of each chip of the sensor 1 will be explained. Figure 2(a) is
Shows the sensor output (amplifier output) when reading a white original. Normally, in this state, the gain of the amplifier 2 is changed for each chip to adjust the output levels of all chips to match, so the output of the sensor 1 (amplifier 2 output) becomes flat.

Figure 2 (b) shows the output of sensor 1 (amplifier 2 output) when reading a halftone (gray) original after such adjustment. A step occurs. The occurrence of this level difference means that the read density differs from sensor chip to sensor chip. Incidentally, FIG. 2 is a diagram showing a case where the outputs of a plurality of sensor chips are serially outputted and operated as if they were a single chip sensor.

FIG. 3 is a diagram showing the relationship between scan synchronization and sensor output, and FIG. 3(a) shows scan synchronization T.

FIG. 3(b) shows the sensor output at Tw/2, and FIG. 3(c) shows the sensor output at Tw/3. Image sensor output is (amount of received light) x
(scan synchronization), when the amount of received light is constant,
The sensor output changes by changing the scan period, and if the scan synchronization is set to 1/2.1/3, the sensor output will be as shown in Fig. 3(b).

As shown in (Q), they become 1/2 and 1/3, respectively.

The difference in level that occurs at this time is due to the difference in sensitivity characteristics of each sensor chip. In other words, FIGS. 3(b) and (c)
shows that the sensitivity characteristics of the sensor can be determined by changing the sensor output synchronization when the amount of received light is 1/2.1/3 (when a halftone original is read).

Next, we will discuss the operation of the image input device shown in Figure 1 in the fourth section.
This will be explained with reference to the figures. First, the scan timing generator 9 sets the scan period of the image sensor 1 to T
. , (T0□<'rw), and A/D converter 3
The data converted into digital input is transferred to line memory 4.
write to. Before the above operation, as explained in FIG. 2, the output level of the sensor (output level of amplifier 2) is adjusted to the same level by adjusting the gain of amplifier 2 during scan period Tv. shall be taken as a thing. Thereafter, the CPU 6 reads the contents of the line memory 4, samples a predetermined number of data for each sensor chip, and calculates the average of the sampled data for each chip, thereby determining the output level of each chip. Next, the scan timing generator 9 sets the scan period of the image sensor 1 to "rGZ, TQ3-T(IN (Ta
)l-<T (12<Ta,<Tw), perform the same processing as above, find the sensor output level for each scan timing, and find the sensitivity characteristics for each sensor chip. Note that the scan synchronization setting is performed using a control signal from the CPU 6.

Thereafter, the CPU selects a curve for correcting this sensitivity characteristic from among the curves shown in FIG.
N corresponds to the sensor chip number. ). Latch 7-1~
The output of 7-N is wired OR/connected to the upper address of the ROM 8 containing the sensitivity correction curve, and its output is connected to the chip select signal 7-N from the scan timing generator 9 shown in FIG. 2(c). 1a, 7-2a
...7-Na are enabled in sequence.

Further, the output of the A/D converter 3 is connected to the lower address of the ROM 8, and data before sensitivity correction is input thereto. Through the above steps, the sensitivity is corrected for each sensor chip.
When reading halftone originals, it is possible to eliminate the difference in the sensor output level as shown in FIG. 2(b), and the reading density for each chip can be the same 2j. However, in another embodiment, correction data is obtained at multiple points using the above method.

A correction curve is created by the interpolation algorithm under program control by the CPU 6, and this correction curve is stored in a memory such as RAM, and if this RAM is used instead of ROM 8, the correction curve can be prepared in the ROM in advance. It is possible to correct the sensor output even without it.

Further, in the above embodiment, after the correction curve is switched for each sensor chip and the sensor output level is corrected to be the same level, the binarization circuit 10 converts the output of all the sensor chips into two values at the same slice level 10a. However, instead of correcting the sensor output level conversely, by creating a correction curve for the slice level loa in the same manner as in the above embodiment, and switching the correction curve for the slice level 10a for each chip, Effects similar to those of the above embodiment can be obtained.

Furthermore, by incorporating two sensors in FIG. 1 instead of the image sensor 1, the above correction can be made within the image sensor, and it is possible to make the sensor output level the same when reading a halftone original.

〔Effect of the invention〕

According to the present invention, since the sensitivity characteristics of each image sensor can be corrected, the reading density of each sensor becomes the same when reading a halftone original, which has the effect of improving image quality deterioration. In addition, when correcting the sensor sensitivity characteristics, the actual halftone J
Since there is no need to read the M draft, the operability is greatly improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing sensor output when reading a halftone original, FIG. 3 is a diagram showing the relationship between scan synchronization and sensor output, and FIG. 4 is a diagram showing the relationship between scan synchronization and sensor output. FIG. 5 is a diagram showing a correction curve of sensor sensitivity characteristics, and a diagram showing the relationship between sensor output and slice level.

Claims (1)

[Claims] 1. In an image input device having an image sensor composed of a plurality of semiconductor chips, means for changing the scan period of the image sensor, and a sensor output corresponding to the change in the period under a constant amount of received light. an image input device comprising: means for reading the gamma correction curve for each chip and calculating a gamma correction curve for correcting the sensor output; and means for correcting the sensor output for each chip based on the correction curve. . 2. The image input device according to claim 1, wherein the sensor output is obtained at a plurality of points by changing the scan period, and a correction curve is calculated using an interpolation algorithm. 3. In an image input device having an image sensor composed of a plurality of semiconductor chips, means for changing the scan cycle of the image sensor, and reading sensor output for each chip with respect to the cycle change while the amount of received light is constant. Means for calculating a gamma correction curve for correcting the binarized slice level of the sensor output, and means for correcting the binarized slice level of the sensor output for each chip based on the correction curve. Featured image input device.