JPH04310071A - Picture reader - Google Patents

Abstract

PURPOSE:To attain binarization at an optimum exposure through the use of an optimum slice level independently whether the picture read density is set automatically or manually. CONSTITUTION:This reader is provided with a halogen lamp 7 to expose an original, a line sensor 13 to read the exposed original picture, a binarization circuit 31 binarizing a picture signal from the line sensor 13, a detection signal generating circuit 34 to detect a frame of an original picture and a CPU 25 controlling the light quantity of the halogen lamp 7. A slice level of the binarizing circuit 31 is decided based on a photometry value obtained by the preliminary scanning before the main scanning for read of the original picture and a value representing the light quantity in the main scanning.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading apparatus, and more particularly to an image processing apparatus for reading and processing an image having a frame around the periphery of the image.

0002

BACKGROUND OF THE INVENTION Conventionally, in a microfilm having a negative image, a negative image is recorded in a frame f of a microfilm F, and the periphery of each frame f becomes transparent, as shown in FIG. 19(A). There is. When this microfilm F is printed out using a negative/positive reversal information recording device, the transparent portion around the frame f is developed. As a result, as shown in FIG. 19(B), a solid black frame B is printed on the transfer material P around the image area G, which not only spoils the aesthetic appearance of the printed image but also increases toner consumption. There was a problem.

In order to solve the above problem, the area of the image frame f of the microfilm F is detected, and the area to be recorded on the transfer material P is controlled based on the area determined based on the image area. Accordingly, an information recording apparatus has been proposed that can obtain a print without a frame as shown in FIG. 19(C).

[0004] In this device, image reading information during image scanning is binarized using a predetermined threshold value and stored in the RAM. This information is, for example, as shown in Figure 20, where L in the figure is
x and Ly correspond to the horizontal and vertical lengths of the transfer paper P, respectively. Then, the CPU sequentially determines whether one column of data in the RAM contains all 0s or 1s, starting from the bit string of l1, and if the first column is all 0s, this judgment is made using l1, l2. , l3, . . . are repeated in sequence. Then, when the first column ln containing 1 is found (l6 in the illustrated example),
The CPU sets the value LF of the image recording area to n. Furthermore, the above operation is similarly performed at the left end and both upper and lower ends of the image area G to determine the image recording areas lx1, ly1, and ly2.

[0005]

[Problems to be Solved by the Invention] However, in the conventional example described above, it is difficult to increase the speed because the entire area image sampling data of the original image is stored in the RAM, and then the data is processed and the frame is erased. and
In addition, a large RAM capacity has the disadvantage of increasing costs.

[0006]Therefore, there has been a demand for real-time processing in which image signals read by image scanning are subjected to image processing using pipeline processing, and at the same time automatic border erasure processing is performed sequentially using partial area image sampling data of the original image. .

However, when performing this automatic frame deletion, 2
Setting the value slice level to a fixed value is relatively good if automatic density adjustment (AE) is effective, but depending on the user's preference, the density may be finely adjusted using AE, or AE may be canceled and manual density adjustment performed. If you do so, there is a high possibility that malfunctions will occur.

[0008] Furthermore, when performing this automatic frame erasing, the AE
If they are performed at the same time, for example, when determining the slice level for automatic frame erasing, the slice level can be easily determined by measuring the light intensity at the time of exposure, but AE always uses a constant light intensity in any case. Since photometry must be performed with AE, separate pre-scans (twice) must be performed with AE and automatic frame erasure.

[0009]

[Means for Solving the Problems] The present invention has been made in view of the above points, and includes a light source for exposing an original, a reading means for reading an image of the exposed original, and an image signal from the reading means. a frame detection means for detecting a frame portion of the original image; and a light amount control means for controlling the light amount of the light source. The present invention provides an image reading device that determines a quantization parameter of the quantization means based on a photometric value obtained in the original scan and a value indicating the amount of light in the main scan. The present invention provides an image reading device that controls the amount of light in the main scan and determines the quantization parameter of the quantization means based on the photometric value obtained in the pre-scan.

0010

Embodiment FIG. 3 is an external view of the present invention applied to a digital reader printer for microfilm, and FIG. 4 is a schematic diagram of its mechanism.

In FIGS. 3 and 4, 1 is a scanner for projecting or reading the microfilm onto a screen and performing image processing, and 2 is for processing the image information read by the scanner 1 and then printing it out on plain paper. 3 is an operating unit for making various settings and displays; 4 is a roll carrier having a mechanism for sending and rewinding the photographed image of the roll film to the projection position using a knob on the operating unit 3; , 5 is a projection zoom lens with a zoom mechanism, 6 is a screen for projecting the microfilm image, 7 is a halogen lamp for irradiating the microfilm, and 8 is a condenser for condensing the diffused light of the halogen lamp. An optical lens, 9 a microfilm sandwiched between pressure plate glass (not shown) in a roll carrier, 10 a reflection mirror for reflecting projection light, and 11 for selecting screen projection or projection image reading. A sliding mirror, 12 is an axis for rotating the sliding mirror 11, 13 is a line sensor for reading a projected image, and 14 and 15 are sensors for detecting the position of the line sensor.

The movement of image reading in the above configuration will be explained.

First, the sliding mirror 11 is normally located at a position as shown by the broken line. At this time, the halogen lamp 7, the condensing lens 8
The light irradiated onto the microfilm 9 is reflected by the reflection mirror 10.
, is projected onto the screen 6 through the sliding mirror 11 (
(dashed optical path).

When the copy button provided on the operation unit 3 is pressed, the sliding mirror 11 first moves around the shaft 12 to the position shown by the solid line, and the microfilm projection light follows the optical path shown by the solid line. At this time, the line sensor 13 separates from the home position sensor 14 and starts moving in the A direction. (Hereafter, this direction will be referred to as the pre-scan). When the line sensor 13 moves to the end position of the image light, it is applied to the start position sensor 15 and begins to move in the B direction. (Hereafter, this direction will be referred to as the main scan). When the line sensor 13 touches the home position sensor 14, the movement of the line sensor 13 stops and the sliding mirror 11 returns to the broken line position.

FIG. 1 is a schematic block diagram of an embodiment of the present invention, and FIG. 2 is a detailed block diagram of an image processing circuit 26. As shown in FIG. In the following explanation, the inversion of a signal is represented by * (for example, the inversion of signal A is represented as signal A*)
.

In FIG. 1, 21 is the line sensor 1
3 is an amplifier for amplifying the image signal output, and 22 is an A for converting the output signal of the amplifier into a digital 8-bit signal.
/D converter; 23 is a shading correction circuit for shading correction of the digitalized image signal;
24 is a sampling circuit for obtaining a photometric value from the shading-corrected image signal, and 27 is a lamp drive circuit for driving a halogen lamp.

Reference numeral 25 denotes a light amount control and operation unit 3 for the control halogen lamp 7 that takes in sampling data in response to an interrupt signal (INT*) from the sampling circuit 24.
A CPU that reads or displays the settings of
6 is an image processing circuit that performs various image processing on the shading-corrected image signal.

Next, as detailed blocks of the image processing circuit 26, in FIG. 2, 31 is a binarization circuit for binarizing the image signal VM having 8-bit gradation at the slice level SL, and 32 is a binarization circuit. 33 is a binarization circuit (comparison value ref) for determining the value of the added image signal VMB in blocks; 34 is a binarization circuit 33 that adds bits to the image signal VMB; This is a detection signal generation circuit for generating detection signals MBI and MBIT for detecting a boundary point between an image portion and a frame portion using a converted signal SIG.

35 is a 13-bit counter that counts based on the image block GCLK in synchronization with the main scanning direction synchronization signal, and 36 is an address correction circuit for correcting the delay (k) in the main scanning direction that occurs when forming blocks. A 13-bit adder 37 is an image part falling address latch circuit that latches the address in the main scanning direction at the rising edge of the detection signal MBIT, and 39 is an output address signal NB of 37 by the synchronization signal HSYNT signal generated for each T line.
A T-line latch circuit that performs latching and delays, 40 also outputs 38 output address signals NA by the HSYNT signal.
This is a T-line latch circuit that latches the .

41 is an address timing conversion circuit A which generates timing from the address signals of NB and NA; 42
Similarly, the address timing conversion circuit B generates timing from the address signals PB and PA of output signals 39 and 40.

43 to 47 are gate circuits.

Reference numeral 48 denotes an edge emphasis circuit for emphasizing the edges of the image signal VM; 49 an error diffusion circuit for performing pseudo-halftone processing on the edge-enhanced 8-bit gradation image signal by an error diffusion method; and 50 is an image delay circuit for correcting a delay in the sub-scanning direction that occurs during blocking.

The flow of image signals will be explained based on FIG.

In FIG. 4, in the section where the line sensor 13 moves from the home position sensor 14 to the start position sensor 15, that is, during the pre-scan operation, the signal output from the line sensor 13 is transmitted to the amplifier 21. The image signal is amplified by the A/D converter 22, digitized by the A/D converter 22, subjected to shading correction by the shading correction circuit 23, and sent to the sampling circuit 24.

FIG. 5 is a diagram showing the sampling method, and the shaded area is one block, which is made up of a rectangle with 64 pixels in the main scanning direction and 3 pixels in the sub-scanning direction.

At this time, the photometry area is set inside the paper size, and the sub-scanning direction 3 is set at the timing of the photometry area.
An interrupt signal (INT*) is output for every line, and the arithmetic processing circuit (not shown) inside the sampling circuit 24 outputs an interrupt signal (INT*) for every line.
The average value of the image signal level within the 4 x 3 area is
25 data buses.

[0027] The output timing of the interrupt signal (INT*) and photometry data register (AED7 to AED0) is shown in Figure 6.
It looks like this. That is, when the INT* falling signal is input to the CPU 25, the CPU 25 enters an interrupt processing routine and reads data from the photometry data registers (AED7 to AED0).

In this way, the photometric data is successively input every three lines and is stored in the internal RAM of the CPU 25. Hereinafter, the operation of the CPU 25 will be explained with reference to the flowchart in FIG.

At this time, the halogen lamp 7 is controlled to have a light amount for photometry, and the light amount data at that time is set as L0.

In this way, image data is sampled locally, and the minimum value MIN of the sampling data is
(Step ■).

Next, MIN'=255×log(MIN)
Logarithmic transformation is performed using the formula /log(255). This is so that subsequent calculation processing can be easily performed using only addition and subtraction expressions.

[0032] Then, in step ■, check the exposure mode, and if the exposure mode is manual, proceed to step ■, AE.
When , the process branches to step ■.

In the manual mode, the exposure data L value for exposure during actual main scanning is calculated from the copy density F value set on the operation unit 3 (step 2).

On the other hand, at the time of AE, the previously determined MIN'
Based on the AE fine adjustment value set on the operation unit 3, the L value is calculated from a preset AE table (step 2).

The AE table is a table showing the relationship between the minimum sampling data value (logarithmization) MIN' and the exposure data (logarithmization) L', and is as shown in FIG. 8, for example. (Here, L'=255×log(L)/log(25
5)).

Using this table, exposure data L is selected so that copies with appropriate density can always be obtained even when using various films with different densities or when lens magnifications are different.

Next, as described above, the calculated MIN' and L'
Calculate the slice level SL based on (step ■
).

Specifically, first, SA' and FIG.
SB' is determined using a table created based on characteristic parameters such as .

[0039] Then, SL'=MIN'+SA'+SB
' to find SL'.

Next, perform antilogarithmic transformation,

[0041]

[Outside 1] Find SL.

Here, FIG. 9 is an exposure amount correction graph obtained empirically from various films, and FIG. 10 is an exposure amount correction graph in which the value is set to 0 when the photometric light amount data is Lo.

During the main scan, exposure is controlled using the previously determined L value, and automatic frame erasure is performed using the previously calculated slice level SL (step 2).

In FIGS. 1 and 4, during the period in which the line sensor 13 moves from the start position sensor 15 to the home position sensor 14, that is, during the main scan operation, the signal output from the line sensor 13 is transmitted to the amplifier 21. After being amplified by the A/D converter 22, digitized by the A/D converter 22, and subjected to shading correction by the shading correction circuit 23, the image signal (VM) is sent to the image processing circuit 26.

In FIG. 2, the image signal VM is compared with the slice level SL calculated by the previous calculation, and converted into a binary signal of 1 or 0 by the binarization circuit 31. A block for one point X on the image has 9×9 sample points as shown in FIG. 14, and each sample point has an interval of 8 pixels in the main scanning direction and 8 pixels in the sub-scanning direction. Therefore, the block size is 64 x 64 pixels, and the sensor 13
Since we are using one with a resolution of 400 dpi, the block will be approximately 4 mm wide. Therefore, dust and dirt enlarged to a width of approximately 2 mm will not cause malfunction.

By the way, the block bit addition circuit 32 calculates the number of 1 sample points (total sum) in a 9×9 block and takes it out as a SUM signal.

This SUM signal is processed by the binarization circuit 33,
SUM compared with the preset number level ref
When ≧ref, it is set to 1, and when SUM<ref, it is set to 0, and it is converted into a binary signal SIG.

Based on this signal SIG, the detection signal generation circuit 3
4, an image falling detection signal MBI and an image rising detection signal MBIT are generated. These two signals are signals that detect the front and rear frame positions in the main scanning direction.

The 13-bit counter 35 is a synchronous counter that is cleared by the HSYNC* signal as shown in FIG.
Counts up at the rising edge of CLK.

[0050] For 13 bits, line sensor 13 is 400
The number of bits assumes main scanning with a dpi A3 size width, so it varies depending on the resolution of the line sensor and the image reading width. Blocking causes delays in both the main scanning direction and the sub-scanning direction, but this can be dealt with by correcting the address using the 13-bit adder 36 in the main scanning direction. The correction value is K, but for convenience the following timing is K=
It is set to 0. The address value of the main scanning direction counter corrected in this way is set as HAD.

FIGS. 12 and 13 are timing charts showing the process of masking a frame on an image and forming a black border at the boundary between the image and the frame. (Hereinafter referred to as HSYNT
C (1 line)).

The top signal in FIG. 12 is HSYNC*,
Pulses that fall at intervals of one pulse (two cycles of GCLK) are output on one line. The image area is this HSYN
It exists between the pulses of the C* signal, but due to the relationship with the printer, one normal line has a clock timing that is longer than the image reading width, and there are also non-image areas before and after the image area. This figure 12 shows the timing when an image appears from a state where there is no image in the sub-scanning direction, and when there is an image,
The SIG signal becomes 1.

However, since the text portion of the image portion has a density close to that of the frame portion, the SIG signal may become 0. At this time, address signal NB indicates the first rising point of the SIG signal, and address signal NA indicates the last falling point of the SIG signal in one line.

This is because the MBIT signal becomes 1 at the first rise of the SIG signal and becomes 0 at the next fall of the HSYNC* signal, so the NB address is latched at the rise of the MBIT signal. Since the MBI signal rises several times during the image section, it is latched several times, but within one line,
Since the last rising address NA is latched, the next H
When latched at the rising edge of the SYNC* signal, the NB and NA addresses are latched. Therefore, the newly created timings NL and NLM become as shown in the figure, and the timings PL and PLM created again using the PB and PA addresses latched by HSYNC* also become as shown in the figure.

Then, as shown in FIG. 2, the outputs of the EXOR gates 43 and 44 are

[0056]

[Outside 2] And further, the OR output is

[0057]

[3] Edge enhancement, error diffusion, and AND output of the image signal VD and NLM through a delay circuit.

[0058]

[Example 4] The waveform at each point is as shown in FIG. 12. Image signal VD
is indicated by diagonal lines.

As a result, in the line where the image first appears, the entire image area becomes a black line, and from the next line onwards (H
If SYNT is not HSYNC but every n lines, the next n
(after the line), the edge of the image area in the main scanning direction becomes a black band with the width of the value set in D (see the VDO signal in FIG. 12).

Further, when the SIG changes as shown in FIG. 13, the width of the black screen changes like the VDO signal. As a result of this change in the width of the black screen, the image shown in FIG. 15(A) is converted into the image shown in FIG. 15(B).

The edge emphasizing circuit 48 uses Laplacia's 3×
3 convolution mask is used. The mask coefficient is shown in Figure 1.
6.

Furthermore, as is well known, the error diffusion method (ED method) compares a pixel of interest with a certain threshold value, and calculates the resulting error (difference value between the density of the pixel of interest and the threshold value) to the density of the next plurality of pixels. This is one of the typical pseudo-halftone processing methods.

The delay circuit 50 uses the output V of the error diffusion circuit 49.
It is output as a serial image signal VD delayed in the sub-scanning direction by the number of lines set in DI.

In this way, in an information recording apparatus that records an image by performing image processing on an image signal read by image scanning in a pipeline process, when automatic frame erasing processing is performed in real time, the document image By determining the slice level of the binarized signal used to detect the upper frame based on the photometric data from the previous scan and the exposure amount data from the main scan, the exposure mode is AE or manual. Frame erasure can be performed accurately and with high precision.

[0065] Furthermore, in an information recording device that records an image by performing image processing on an image signal read by image scanning in a pipeline process, when automatic frame erasing processing is performed in real time, it is necessary to By determining the slice level of the binarized signal used to detect the frame based on the photometric data that is also used in AE acquired during the previous scan, one pre-scan is performed regardless of whether the exposure mode is AE or manual. You can quickly erase the frame by using only the button.

(Other Embodiments) In the embodiments described above, the minimum value MIN of the image sampling data is used as the photometric value and the SA
' was calculated and used, but for example, the sensor output level PK1, which is the left peak on the histogram as shown in FIG. 17, may be used as the photometric value, or the maximum value MAX of the image sampling data may be used as the photometric value. The sensor output level PK2, which is the peak on the right side of the histogram, may be used as the photometric value.

Furthermore, these feature points, MIN, MAX,
By performing arithmetic processing on PK1 and PK2 in combination, SA
For example, SA'=(MAX'+MIN'
)×K (K is a constant, 0<k<1) to calculate SA'.

[0068] In addition, when finding the minimum value MIN,
The specified level (Kcut
), and filtering processing such as ignoring the distribution data below this may be performed.

In addition, in the above embodiment, the application was applied to a digital reader printer for microfilm.
It is easy to apply to paper scanners and digital copying machines, and it can be applied to any other method of scanning documents (whether the line sensor is fixed or movable).

Further, in the above embodiments, recording was performed on paper, but it may be on a medium such as a magneto-optical disk (OMD), or on a memory, and the type thereof is not particularly limited.

As explained above, in an information recording apparatus that records an image by performing image processing on pixel signals read by image scanning using pipeline processing, when performing automatic frame erasure in real time, the document By determining the slice level of the binarized signal used to detect the frame on the image based on the photometry data from the previous scan and the exposure amount data from the main scan, the exposure mode can be set to A.
Automatic frame erasing can be performed accurately and with high precision either manually or manually.

[0072] Furthermore, by determining the slice level of the binarized signal used to detect the frame on the original image based on the photometric data also used in AE acquired during the previous scan, the exposure mode can be set to AE or manual. In either case, automatic border erasure can be performed quickly with one pre-scan.

Furthermore, by performing a pre-scan (photometry) during the forward movement and a main scan (exposure) during the backward movement, real-time automatic frame erasure can be performed with only one round-trip scan.

[0074]

As explained above, according to the present invention, a light source for exposing an original, a reading means for reading an image of the exposed original, a quantization means for quantizing an image signal from the reading means, An image reading device comprising a frame detection means for detecting a frame portion of a document image and a light amount control means for controlling the light amount of the light source, the image reading device comprising: a frame detection means for detecting a frame portion of a document image; and a light amount control means for controlling a light amount of the light source; Since the quantization parameter of the quantization means is determined based on the photometric value and the value indicating the light amount in the main scan, when detecting the frame on the document image when performing automatic frame erasure in real time, By determining the quantization parameter to be used from the photometric data during the pre-scan and the exposure data during the main scan, accurate and precise automatic frame erasure can be performed regardless of whether the exposure mode is AE or manual. , the amount of light in the main scan is controlled based on the photometric value obtained in the pre-scan before the main scan for reading the original image, and the quantization parameter of the quantization means is determined. By determining the quantization parameter used during detection based on the photometric data also used in AE acquired during the previous scan, the exposure mode can be set to AE.
In either manual or manual mode, automatic border erasure can be performed at high speed with a single pre-scan.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the overall configuration.

FIG. 2 is a block diagram of an image processing circuit.

FIG. 3 is an external view of the digital reader printer.

FIG. 4 is a mechanical diagram of a digital reader printer.

FIG. 5 is a diagram showing sampling positions.

FIG. 6 is a timing chart diagram.

FIG. 7 is a flowchart diagram.

FIG. 8 is a diagram showing the relationship between MIN' and L'.

FIG. 9 is a diagram showing the relationship between MIN' and SA'.

FIG. 10 is a diagram showing the relationship between L' and SB'.

FIG. 11 is a timing chart diagram.

FIG. 12 is a timing chart diagram.

FIG. 13 is a timing chart diagram.

FIG. 14 is a diagram showing sampling positions.

FIG. 15 is a diagram showing an example of an output image.

FIG. 16 is a diagram showing filter coefficients.

FIG. 17 is a diagram showing an example of a histogram.

FIG. 18 is a diagram showing an example of a histogram.

FIG. 19 is a diagram showing an example of conventional image output.

FIG. 20 is a diagram showing a memory RAM.

[Explanation of symbols]

13 Line sensor 24 Sampling circuit 26 Image processing circuit 25 CPU

Claims (2)

[Claims] 1. A light source that exposes a document, a reading device that reads the exposed document image, a quantization device that quantizes an image signal from the reading device, and a frame detection device that detects a frame of the document image. and a light amount control means for controlling the light amount of the light source, and the light amount control means controls the amount of light from the light source based on a photometric value obtained in a pre-scan before the main scan for reading the document image and a value indicating the light amount in the main scan. An image reading device characterized by determining a quantization parameter of a quantization means. 2. A light source that exposes a document, a reading device that reads the exposed document image, a quantization device that quantizes an image signal from the reading device, and a frame detection device that detects a frame of the document image. and a light amount control means for controlling the light amount of the light source, and controls the light amount in the main scan based on the photometric value obtained in the pre-scan before the main scan for reading the original image, and also controls the light amount in the main scan. An image reading device characterized by determining a quantization parameter of a quantization means.