JPS62136966A - Picture signal processing method - Google Patents

Abstract

PURPOSE:To attain automatic setting of binary-coded threshold value efficiently and properly even when background density changes by smoothing a picture data, deciding a film density corresponding to a minimum value of background and setting an optimum threshold value based on the obtained background density. CONSTITUTION:When a digital picture data is inputted to a low pass filter 606, the data is smoothed and outputted as a picture data. The picture data is a smoothed data of dust, flaw and background noise component, a minimum value L is found out by a minimum hold circuit, where the value is held, a multiplexer 610 is switched by a selection signal from a CPU 182 of a control section 180 and sent to the CPU 602 of the next stage. A density corresponding to the minimum value is stored in advance in the CPU 602, the corresponding density is read by the input of the minimum value and an optimum threshold level to the density is outputted from the read density and the result is applied to a comparator 604.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image signal processing method for setting an optimal threshold value for performing binarization processing of image data.

(Prior Art) Conventionally, image reading devices include copying machines and facsimile machines, and recently there are microfilm reading devices and the like developed as image processing systems.

Myro Film 1 First, prior to explaining the present invention, a conventionally proposed microfilm reading device for an image processing system will be explained.

In recent years, as the amount of information has increased, the amount of manuscripts recording various information has increased dramatically. There is a strong need for a system that can store this large amount of manuscripts in a compact space and that can be searched easily. In order to achieve this requirement, it is necessary to once compress the information recorded on the original and record and store it in some means. Microfilm or optical discs are used as recording means. Microfilm records two-dimensional visible information and has advantages such as long-term preservation, legal evidence, and multiple copies of the same information.

In addition, optical discs can not only record large amounts of information, but also allow information to be added and updated in real time, and because they are recorded as digital information, the recorded information can be directly connected to a computer and used for communication as is. It has the following advantages.

Therefore, in order to take advantage of the advantages of both, the original manuscripts, such as paper manuscripts, are photographed one after another in batches to create microfilm, and then the information recorded on the microfilm is used for future retrieval. It has been proposed to record retrieval information on an optical disc to enable the search. In this way, if the original information is recorded on an optical disk together with the search information, anyone can read out the recorded information and print it out or create a microfilm when necessary, and it can also be used directly with a facsimile machine. It can also be transmitted over long distances, allowing for advanced utilization of information.

FIGS. 2 to 4 are block diagrams for explaining an example of a microfilm reading device of an image processing system for writing information from an original document to an optical disk, which has been conventionally proposed for this purpose. .

First, the overall outline of this device will be explained with reference to FIG.

This microfilm reading device first records original image information such as text, drawings, and digital information recorded in computer memory as microfilm information, and then uses a microfilm scanner to read only the desired information from the microfilm information. Configured to file to optical disk.

In FIG. 2, 1, for example, original image information of a paper document 10, etc. is photographed on microfilm by a microfilm photographing device 20 (hereinafter simply referred to as a camera), and 1. a developing process 3o is performed as usual to create a microfilm 40. . This microfilm 40 includes 16mm roll microfilm 42.35mm roll microfilm 44. There is a microfiche film 48 and an aperture film 48.

Next, the image information stored on this microfilm 4o is read by a microfilm scanner loo and converted into digital image information PS.
and 1. Management information for making correspondence with the imaging order of this image is sent to the next optical disc recording device fi20G, where this image information PS is stored on the optical disc 50 together with the management information. When reading the information Ps, microfilm 4
0 into a dedicated kit 102, set it at a predetermined position, scan the loaded microfilm 40 with the image sensor 104, and scan the screen lO8 as required.
It is also possible to perform digital conversion while viewing a single image by projecting an image onto the screen. Furthermore, when recording the image information PS on the optical disc 50 in the optical disc recording device 200,
The scanner 202 reads the input image information PS, the CR7 display device 204 displays the image, and while looking at the displayed image, the keyboard 20B inputs index information indicating the image content, type, etc., and the information is stored on the optical disc 50. Writing is in progress. Note that this index information may be entered using the keyboard 206 while looking at the image projected on the screen 108 of the microfilm scanner 100. Such input work is normally performed by one operator, but the processing capacity is Approximately 3000 pages/8 hours. Therefore, in order to speed up the input work, a personal computer system 60 may be installed to perform the work. In this case, a plurality of computers 62 and 64, for example, are installed, and one personal computer 62 is connected to the microfilm scanner 100. Enter index information using the keyboard 82a while looking at the screen 10fl on the other computer 84.A dedicated microfilm reader 100a is separately installed on the other personal computer 84, and while looking at the screen 108a, enter index information using the keyboard 22a to read each floppy disk. 68
These index information are written to the floppy disks 6B and 68 in the optical disk recording device 2.
By loading the optical disk into the floppy disk drive 70 of No. 00 and reading data, index information can be written therein in correspondence with the image information stored on the optical disk 50.

Furthermore, the index information created by the personal computers 82 and 84 is configured to be input into the microfilm scanner 100, the microfilm scanner 100 is provided with a search function, and the index information is transferred to the optical disk recording device 200 along with image information and management information. It can also be configured so that it can be transferred to

Microfilm scanner 1 Next, referring to Figure 3, the microfilm scanner 100 is
An example of the configuration will be explained.

This microscanner 100 mainly includes an optical reading section 120 for reading image information on a microfilm, a driving section 140 for the optical reading section 120, and a system for sending a read signal from the optical reading section 120 to an optical device recording device 200. The signal processing section 160 includes a signal processing section 160, and a control section 180 for controlling the driving section 140 and the signal processing section 180.

The optical reading unit 120 includes an illumination system 122 including a light source 1222 and a condenser lens 1224, a microfilm holding device 124 for holding the microfilm 40 between, for example, pressure bonding glasses 1242a and 1242b to prevent image distortion, and a projection system. Lens 12θ2. Imaging lens 1
284 , 126B, an image projection system 12B comprising a half mirror 1288 and a screen IOB, e.g. feed and take-up reels 128a and 128b for feeding the microfilm 40 into the illumination optical path, and a microfilm 40
A mark sensor 130 for optically detecting the blip mark attached to the frame or the density difference between frames, and an automatic exposure control sensor for detecting the legality information of the microfilm 40 in order to read images under optimal conditions. and a sensor unit 134 for scanning optical image information beam-split and projected by a half mirror 1288 with an image sensor 104 and converting it into an electrical signal. In this case, if an image sensor is used that can read an image enlarged to the original size before being reduced to a micro image as it is, the imaging lens 1
284 and 128B can be omitted.

The drive unit 140 includes a drive control unit 142 for driving the supply and take-up reels 128a and 128b in response to signals from the mark sensor 130, and a mechanical connection to the image sensor 104 in order to advance the microfilm 40 frame by frame. and a drive circuit 150 for controlling a motor 148 for driving the coupled screw-and-nut mechanism 144. This rotation of the motor 148 allows the image sensor 104 to scan the optical path surface.

The signal processing section 160 includes a drive reading circuit 182, a linear density switching circuit 164, and an R3422 data line 18B. The drive reading circuit 1B2 moves the image sensor 104 to an appropriate focus based on the film density information measured by the sensor 132 so as to perform image reading under optimal conditions, and also performs photoelectric conversion based on the image sensor 104's detection. The image information thus obtained is sent to the linear density switching circuit 164. The linear density switching circuit 164 has, for example, 166 trees/mm or 8 trees/mm.
The image information is transmitted at a suitable arbitrary linear density, and is sent to the optical disk recording device 200 via the data line 18B.

The control unit 180 controls these drive unit 140 and signal processing circuit 1.
80, and is a circuit for controlling the central processing unit (CP
For example, an R5232 data line 184 for exchanging information C5 such as management information 1 image information, etc. between the CPU 182 and the optical disk device 200, and a line for transmitting commands to the CPU 182 via the data line 184. A computer 186, and a keyboard 1
The drive controller 144 and the drive circuit 150 are controlled by the command 88 via the interface 190, and the CP
Configured to provide other commands to U 182.

Desuph・Me Next, the optical disc recording device will be explained.

FIG. 4 is a block diagram showing an example of the configuration of the optical disc recording apparatus 200.

In this device 200, a CPU 210, a ROM
212 , RAM 214 , CPU 210 , keyboard 218 , and interface 220 , which are connected to a common path line 222 . The interface 220 is the floppy disk device 70 or the host side C.
Connected to P U 224.

Furthermore, this CPU 224 includes a graphics processor 228 for editing, adding, deleting, and enlarging/reducing images through a path line 228. A scanner 202 and a microfilm scanner 100 are connected.

These include a write/read controller 240 and a drive unit 280 that control writing and reading from and to the floppy disk 50 via an interface 230 and a pass line 232.

The write/read controller 240 is the disk data controller 24
2 controls writing of image information and reading of filed image information. Writing is performed by generating a laser beam LB from a laser driver 246 via a modulator 244 in accordance with the image information, management information, and index information read out by the microfilm scanner 100 and scanner 202.

, and reading from the optical disk 50 is performed using the read head 24.
This cell 25 is performed by the cell 250 joined to
The information optically picked up by cell processor 2
52 and demodulated by a demodulator 254. In this case, the position of the head 248 is adjusted to the optical disc 50 by the focusing mechanism 25B.
The structure is such that accurate information can be read from the cell 250 by focusing on the pit of the cell 250.

On the other hand, the drive unit 280 is divided into a sector control system and a crossfeed control system, with the sector control system being managed by the sector control unit 262 and the crossfeed control system being managed by the crossfeed control unit 272. The sector control unit 282 controls the spindle motor 288 via the drive unit 264, the actual control position is detected by the sector wheel 268 and the sector pulse counting unit 270, and this detection data is fed back to the sector control unit 282 to control the spindle motor 288. control to the commanded position.

The crossfeed control section 272 controls the linear motor 276 via the drive section 274, and its control position is set to Moire 1t.
427B and the grating device 280, and the detected data is fed back to the crossfeed control section 272 to control the crossfeed command position.

If such conventionally proposed microfilm reading devices are used, it is possible to read not only existing microfilmed information but also non-microfilmed information as well as information that will be obtained in the future that is suitable for filing on optical discs. The information can be stored on microfilm, and the stored information can be searched and used online.

By the way, in the above-mentioned microfilm reading device, as shown in FIG. The -dimensional image sensor 104 scans and reads the image data in the main scanning direction (X direction) in the direction of the sensor array and in the sub-scanning direction (Y direction) which is the mechanical driving direction of the sensor. Through this reading, video data (image signals) are obtained from the sensor in time series for each pixel, and this video data is binarized and output for post-processing.

(Problems to be Solved by the Invention) When binarizing this image signal, it is necessary to determine a threshold level for the binarization. When applying the conventional threshold level setting method to a microfilm reader, average photometry is performed on the projected image to detect the film density separately from scanning to obtain this image signal, and the film density obtained by this average photometry is The threshold level is determined from the concentration value.

However, in the case of average photometry, the density distribution is averaged over a rather wide range, such as 1/4 of the entire projection area, so if there is a wide gap in the photometry range, for example, , the measured density value shifts toward the density value of this missing portion, and it can be said that it does not necessarily reflect the film density accurately.

Therefore, when the conventional average photometry method is applied, there is a problem in that the threshold level for binarization cannot be accurately determined.

There is also a method of manually adjusting and setting the threshold level, but the background level varies significantly depending on the film, making it difficult to adjust the threshold level appropriately.

Therefore, an object of the present invention is to provide an image processing method for automatically and accurately binarizing video data obtained by reading an image with an image sensor.

Means for Solving Problem C) To achieve this objective, the image signal processing method of the first invention of this application includes the following steps: At the same time, the output of the image data obtained by scanning the projected image of the negative film to be read with an image sensor is smoothed, the minimum value is determined from the output distribution obtained by smoothing, and the backing of the image data is performed from this minimum value. The method is characterized in that a ground density is determined and an optimal threshold value for binarization is set according to this background density.

In carrying out the present invention, the optimum value for binarization is stored in advance in a table memory in association with the background density value, and the optimum threshold value is read out in accordance with this background density value. is suitable.

Furthermore, according to the image signal reading method of the second invention of this application, when setting the optimal threshold for performing the binarization process of image data, the projected image of the positive film to be read is scanned with an image sensor. Find the maximum value of the output from the output distribution of the image data obtained by It is characterized by setting a threshold value.

In carrying out the present invention, an optimum value for binarization is stored in advance in a table memory in association with a background density value, and an optimum threshold value is read out according to this background density value. is preferable.

(Function) As described above, according to the present invention, since the minimum value or the maximum value of the output of the read image data from the image sensor is used as the background density, this background density can be measured accurately and quickly. An appropriate binarization threshold can be set from this background density.

(Example) Hereinafter, with reference to the drawings, an example of the image signal processing method for measuring the background density of image data according to the present invention will be described in the case where it is applied to the microfilm scanner shown in FIG. 3. Furthermore, although the following embodiments will mainly be described with reference to negative films as an example, it is clear that the invention can also be applied to positive films.

FIG. 1 is a block diagram for explaining the present invention, and this diagram shows the sensor section 134 of FIG. Drive reading unit 182, drive circuit summer 50. Parts corresponding to the motor 148 and the control unit 180 are shown.

In FIG. 1, a projection image 300 on a microfilm is electrically scanned in the main scanning direction X and mechanically in the sub-scanning direction Y by the -dimensional image sensor 104.

Sequential time-series video data for each pixel read from the sensor 104 is sent to the shading correction circuit 32 via the amplifier 310.
0, and here the two-dimensional shading of the video data is corrected in order to improve the quality of the reproduced image.The shading-corrected video data is binarized by the binarization circuit 340 and then passed through the interface circuit 380. The output signal is outputted to other necessary processing circuits such as a linear density switching circuit (indicated by 164 in FIG. 3) at the subsequent stage. Note that this two-dimensional shading correction can be performed by the method disclosed in a previous application filed by the present applicant and is not the gist of the present invention, so its explanation will be omitted here.

In the present invention, the processing method for binarization includes performing various image signal processing for binarization in the binarization circuit 340 on image data obtained by shading correction in the shading correction circuit 320. As will become clear from the following description, there are various methods, and the image signal processing for binarization will be explained in the case of a negative film.

2 Flume * 1 = Concentration measurement First, one method of binarization is the method shown in Figure 6. In this method, film density is detected from this shading-corrected image data, a certain threshold level is set from this film density, and one image data is compared with this threshold level to automatically binarize it. It's a method.

This case will be explained.

In order to avoid the influence of the film background when setting the threshold for binarization, it is necessary to know the film density.

As already explained, when performing average photometry of a projected image, if the image has a very wide range of areas, the average density value will shift all at once to the density of the area that is the center of the image.
As a result, most of the information such as one character is skipped and no longer appears, and the original image information is no longer accurately reflected. This is due to the fact that average photometry is performed over a wide area such as 5 cm square, for example.

On the other hand, since the image sensor 104 reads information in each very narrow area of 178 mm square of the projected image and sends out the image data, it is possible to measure the film density from the output of the obtained image data by measuring the average photometry. It is possible to measure the concentration much more faithfully than in the case of .

Therefore, in the present invention, the density is measured using the output of the image data obtained from the image sensor 104 and which has undergone some degree of correction such as shading correction. Therefore, in the following explanation, the magnitude of the output of image data may be simply referred to as density.

By the way, this corrected image data, for example,
As is clear from the output distribution curve shown in Figure (A) with time on the horizontal axis and output level on the vertical axis, there are information areas such as characters A1 and A2 through which light passes;
It consists of background B. However, even with this corrected image data, if there are dust or scratches, there may be parts where light does not pass through and the output (corresponding to density) drops sharply, as shown by C1 and C2. Also, the background noise is not completely removed, making it extremely noisy.

Therefore, it is necessary to know the density level of the background in order to set the threshold level to avoid the influence of noise caused by such dust, scratches, etc.

Therefore, dust, scratches, background noise components, etc. contained in such image data are smoothed by a low-pass filter (LPF), and the minimum value (minimum level) of the output is calculated from the output distribution of the smoothed image data. ), the corresponding density can be determined from this minimum value and a threshold level suitable for this density can be set. Therefore, the threshold value for binarization that is faithful to the original image information can be set. You can set the level. In the case of positive film, the corresponding density value can be determined by measuring the maximum output value (peak level).

FIG. 6 is a block diagram showing an example of an image processing apparatus for explaining a method of measuring film density as described above to determine a threshold level and binarizing image data using this threshold level. In FIG. 6, 600 is a film density measurement section (image data output measurement section);
2 is a central processing unit (CP) for setting the binarization threshold.
U), 604 is a comparator for binarization. still,
This CPU 602 is the CPU 18 of the control unit 180.
This film density measuring section 600, which may also be used as 2, is a low-pass filter 60B and a multiplexer (MPX).
It is equipped with 810.

Next, the operation will be explained. Now, when a line worth of digital image data having an output (0 degree) distribution as shown in FIG. 7(A) is input to the low-pass filter 808, it is smoothed and the horizontal axis shows time and vertical axis as shown in FIG. 7(B). It is output as image data having an output distribution with the output level plotted on the axis.

This image data has dust, scratches, and background noise components smoothed out, and a minimum hold circuit searches for and holds the minimum value, and the multiplexer 81
0 is switched by a selection signal from CPU 1B2 of control unit 180 in FIG. 3 and sent to CPU 802 at the next stage.

The minimum value of this output can be determined by simply finding one point for one line in one scanned screen, but for example, if this minimum value is based on dust or scratches, it may not necessarily be an appropriate minimum value. Therefore, find one point for each of several lines, for example four lines, on one screen,
The average value of these may be determined as the minimum value.The control unit 1 shown in Fig. 3 determines which line's minimum value is to be held.
It can be controlled by a control signal from the CPU 182 of 80.

Further, these average minimum values can be configured to be calculated by the CPU 802. In this case, CPU80
2, for example, a memory that captures and stores the lower limit value determined for each line, an average minimum value calculation means that reads each minimum value from this memory and calculates the average value, and a device suitable for performing this procedure. Other necessary means can be provided.

The minimum value thus obtained is sent to the CPU 602. This CPU 802 stores the density value corresponding to this minimum value in advance, and can read out the corresponding density value by inputting the minimum value. The configuration is such that the threshold level can be output and supplied to the comparator 604. It is also possible to configure the threshold value to be output directly from the minimum value.

8th rI! J is the relationship (curve I) between the film background density (horizontal axis) obtained in this way and the minimum value (vertical axis) of the output of the image data that has passed through the low-pass filter B06, and the relationship of the binarization described later. FIG. 6 is a characteristic diagram showing the relationship (curve ■) with the threshold value.

FIG. 8 shows the experimental results when the output after A/D conversion at the snout portion of the test film was set to FFH (H indicates a hexadecimal number). The vertical axis on the left side of the figure shows the minimum value of the image data output after passing through the filter 60B, and the background density of 1.0 on the horizontal axis is the density at which the contrast is optimal. As the value increases, the concentration increases, and conversely, as the value decreases, the concentration decreases.

As is clear from the illustrated curve I, there is a one-to-one relationship between the minimum value and the film density. Therefore, the relationship between the minimum value and the film density is stored in advance in a table memory 818, such as a table RAM or a table ROM. If stored in a suitable memory according to needs, when the minimum value measured as described above is input to the CPU 802, the lower limit value of the background density value corresponding to this minimum value can be read out and known. I can do it.

By the way, in FIG. 8, the vertical axis on the right side shows the threshold value for binarization in arbitrary units. FIG. 8 shows the relationship between the background density and the optimal threshold for binarization. As can be understood from this figure, when the film density is changed, for example, from about 0.5 to 1.5, the binarization threshold for obtaining an optimal image changes over a wide range. Therefore, if you try to manually set the binarization threshold using this relationship between the background density and the threshold, in areas where the threshold adjustment range is narrow, if the adjustment is even slightly off, the entire background will appear in the reproduced image. In this way, the optimum threshold value for a single film must be set very precisely, and adjustment is extremely difficult.

Therefore, in order to be able to automatically set the binarization threshold level, the threshold level TH is set by shifting the level by a value that can remove this density (output) minimum value/<Tsugground B. If the relationship between the background and the optimal threshold value is set in advance in a one-to-one relationship, it is possible to appropriately binarize the information portions A and A2 in the image and the background B. Figure 7 (B)).

The setting of this threshold level is based on the CPU [10
This is done by writing a table in advance in the table memory 616 of No. 2 in association with the minimum density value. For example, if the output of one image data is 20H, the corresponding threshold level is 5
Set it to AH. Note that the shift amount in this case can be set as any suitable amount as required.

In this way, the image data output, that is, the background density, is measured for each film to be read, the optimal threshold level for binarization is automatically set from that value, and this threshold value is stored in the table of CP 0802. It can be output from the memory 616 to the comparator 11104.

With this configuration, image data can be automatically binarized accurately and quickly.

Further, although the above embodiment has been explained with reference to a negative film, in the case of a positive film, the maximum value of the image data output is held, and the optimum threshold value corresponding to this maximum value is stored in the table memory 6.
By setting it to 16, the binarization threshold can be automatically set in the same manner as described above.

Although the circuit shown in FIG. 8 is configured to process digital signals, it can also be configured to process analog signals. In this case, no table memory is required.

Contrast of -5, 1) ■ ■ The above-mentioned binarization processing may not be able to perform binarization accurately when the image frequency becomes high.

As is generally known, when the frequency components of an image become high, the contrast becomes low because the output difference between white and black becomes smaller compared to the low frequency components. For example, as shown in the image data waveform shown in Figure 9 with time on the horizontal axis and output level on the vertical axis, there is a clear contrast difference between signal A3 and background B in the low frequency region. ,
In the high frequency region, the contrast difference between the signal A4 and the ground B decreases significantly, resulting in so-called image blurring. Therefore, even if the threshold level TH is determined as described above, the image data in the high frequency region There is a possibility that accurate binarization of the data cannot be expected.

As a method for accurately performing binarization by removing the influence of image blur, the present invention can consider three main methods as shown in FIGS. 10(A) to 10(C).

The first method is to perform blur correction on the image data using the anti-repo correction circuit 700, and use the constant binarization threshold level TH obtained as described above or other appropriate threshold level as the threshold level. , a comparator 720 performs the comparison.

In the second method, as shown in FIG. 10(B), the image data to be binarized is sent as it is to one input terminal of the comparator 720, and on the other hand, this image data is sent to the smoothing circuit 722, compressed. Circuit 726 and level shift circuit 730
to form an optimal threshold level TH for binarization,
This threshold level is sent to the other input terminal of the comparator 720, and binarization is performed there.

The third method is to perform the first and second methods at the same time,
The image data to be binarized passes through a blur correction circuit 700, and the threshold level is set by passing this image data through a smoothing circuit 700.
2. After passing through a compression circuit 726 and a level shift circuit 730, the signals are supplied to a comparator 720 for binarization.

Below, the case where these binarization methods are performed by digital processing will be explained. Further, as described above, as an example, image processing will be described regarding image data obtained by two-dimensionally scanning a projected image of a negative film with an image sensor.

First method> To remove the blur in this image, it is sufficient to emphasize the high-frequency components, and this method uses a high-frequency emphasis filter, such as the known unsharp mask (hereinafter simply referred to as mask US). (Reference: Basic optics of light and images (published by the Institute of Electrical Engineers of Japan)).

FIG. 11(A) is a circuit block diagram corresponding to FIG. 1θ(A), which shows in detail the circuit block when the mask US is used as the blur correction circuit 700.

Mask US focuses on the fact that image blur appears at edges between areas of uniform density, and attempts to make the gradient of density steeper. According to this mask US, when considering the image space as an odd matrix in which the central pixel appears,
The pixel data corresponding to the center pixel is emphasized and outputted.
There are 8 pixels, and each line is main-scanned, and data is stored in each line memory in the order in which pixels are read for each line. If one pixel is one byte, this line memory is a 2048-byte line memory.
), consider the case where data corresponding to nine pixels of a 3×3 matrix are read out simultaneously and the data of the center pixel is emphasized. The center pixel data of this matrix is
Assuming that the pixel data a-i are arranged as shown in the figure, the emphasized pixel data E corresponding to the center pixel is given by E=5e-(b+a+h+f) (1).

Therefore, as a mask US, 3×3 out of 1 image data
Pixel data corresponding to 9 pixels of the matrix are read out in parallel, and these pixel data are multiplied by equation (1) (5e).
, addition (b+d+h+f) and subtraction thereof, an emphasized output E can be obtained.

In the blur correction circuit 700 shown in FIG. 11(A), n×n
(n is an odd number greater than 3) Although it is also possible to configure using a matrix, an example in which the mask US is configured as a 3×3 matrix will be described in detail below.

The blur correction circuit 700 shown in FIG.
0 and a subtraction circuit 712.

The matrix memory section 702 includes a line memory section 703, an address counter 704 for writing and reading pixel data into and from the line memory section 702 in accordance with a control signal from the CPU 182 of the control section 180, and a data counter for nine pixels. It also includes a latch section 70B for parallel output.

Furthermore, this line memory section 703 is connected to the image sensor 1.
Number of pixels of 04 (number of dots in l line, in this example 204
8 pixels) line memories (Ml, M2, and
703a, 703b, and 703c) are arranged in parallel, and the image data read for each line is controlled by a write control unit 703d and cyclically written to one of the line memories for each line. Now, the image data j-2 of the two previous lines is written in the line memory Ml, the image data j-1 of the previous line is written in the line memory M2, and currently the image data j-2 of the previous line is written in the line memory M3. Assume that image data j is being written to and the data is being updated. On the other hand, these line memories M1, M2 . At this point, the image data j-1 and j-2 of the line memory that is not currently a write type are read out in parallel in synchronization with the writing by the read selector 703e. Simultaneously read out from M2 and store the read pixel data in the latch circuit? 03
g, 703h, and then sent to the next stage latch unit 706. At the same time, image data j currently being written is directly sent to the latch unit 708 via a latch circuit 703f via a separate route.

This latch section 708 has two stages of latch circuits 708a, 708b; 708c, 706d; 701 for each line.
3e and 708F are provided, and the configuration is such that output can be taken out from a total of three locations at the same time, front and rear. Now, three sequential pixel data of image data j-2 read from line memory M2 are designated as a, b, and c, and similarly, sequential pixel data of image data j-1 from line memory M1 is designated as d.
, e, f, and the pixel data of image data j currently being written are g, h, i, the latch unit 70B stores these pixel data a-i (9
data) are read simultaneously in time and output in parallel.

Note that in this line memory section 703, the latch circuit 7
03f to 703h are line memories 703a to 703
This is provided as necessary to match the timing of reading in relation to the reading speed of c, and may be omitted.

Next, one pixel data e among the pixel data a to i thus read out is sent to a multiplication circuit 708 that performs multiplication (5×e) of equation (1). On the other hand, pixel data b, d, f
, h to perform the addition operation of equation (1), an addition circuit 710
supply each. The outputs calculated by the multiplier circuit 708 and the adder circuit 710 are supplied to the subtraction circuit 712, where the subtraction of equation (1) is performed, and as a result, (1
) is obtained. Such processing is performed on the image data of the blurred image, and these emphasized pixel data are supplied to the comparator 720 as image data to be binarized.

As can be understood from the image data waveform diagram shown in Fig. 11 (C) where the horizontal axis shows time and the vertical axis shows output level, the image data that has been enhanced through this mask US has signal A4 (Fig. 9) emphasized. The output level of the highlighted signal (white) As increases, and the output level of the background B1, in which the background (black) B (FIG. 9) is emphasized, decreases, so that the contrast difference appears sufficiently.

Therefore, according to this first method, image data with extremely fine patterns that could not be binarized using conventional methods and would be skipped as a white level can be converted to a certain threshold level that is optimal for binarization as described above. With a certain threshold level obtained by TH or known methods, it is possible to binarize accurately and appropriately, thus increasing the resolution.

Second method> This method uses a shading correction circuit (320 in Figure 1).
Instead of supplying the image data with the blurred image sent from ) to the comparator 720 as is, the threshold level is set using a method different from the method described above to perform accurate binarization. .

Therefore, in this method, after smoothing the image data,
This method attempts to obtain a high-quality reproduced image by compressing the image data and then level-silting the compressed signal to form an optimal threshold level for binarizing the image data.

FIG. 12(A) is a block diagram corresponding to FIG. 10 CB), and has a structure in which a known median filter is used as the smoothing circuit 722. This example will be explained. The median filter itself is described in the above-mentioned literature:
Since it is disclosed in "Basic Optics of Light and Images (IEEJ series)", detailed explanation will be omitted.

This median filter 724 is a low-pass filter,
As is already known, this filter is defined as nXn
For example, when configured as a 3X3 matrix filter,
The density values (output levels) of nine pixel data a to t are arranged in descending order and the fifth (center) density value from the smallest one is output. It is possible to output pixel data with less blur at edges and smooth out low-frequency signals. Therefore, by using the median filter 724, even if the background level of image data fluctuates somewhat due to dust, image sensor variations, or other causes, it is possible to obtain a smooth signal with that fluctuation removed.

In order to input pixel data to this median filter 724, this smoothing circuit 722 is connected to a matrix memory section 702 for reading out pixel data in parallel, which has the same configuration as that described in the embodiment of the first method (FIG. 11). (line memory section 703, latch section 70B, and address counter 704), and since these circuit components operate in the same manner as described above, they will be denoted by the same reference numerals and their repeated explanation will be omitted.

Also in this embodiment, it is assumed that image data j of a certain line is input to the smoothing circuit 722. Similarly to the above, image data j-2 of the two previous lines, image data j- of the previous line
The three sequential pixel data a to i of the image data j of the current line and the current line are simultaneously read out in parallel and supplied to the median filter 724 in parallel as nine pixel data ai. The image data obtained from the median filter 724 is a signal from which background noise has been removed. This smoothed image data is shown in FIG. 12(B), which is a signal waveform diagram in which time is plotted on the horizontal axis and output level is plotted on the vertical axis, and the smoothed background is indicated by B3.

Next, the image data thus smoothed is compressed and outputted by the compression circuit 72B, and the level of the compressed signal is changed to a level above the lower limit level (background level) by the level shift circuit 730. A threshold level at which high-frequency component image data can be appropriately binarized is formed by shifting the image data so that the image data has a high frequency component. If you do this,
A threshold level can be obtained from the image data to be binarized. In this case, the degree of compression and the amount of shift are performed appropriately depending on the output level of the image data, but when the output difference between the minimum peak of the output of the image data before compression and the background level is taken as 100%, The threshold level is set to be 30 to 95%, preferably 80 to 90%, of the minimum peak. Furthermore, in this case, the reason why the compressed signal itself cannot be used as it is as the threshold level is because this signal is at almost the same level as the background and cannot be distinguished.

FIG. 12(C) is a diagram showing the compressed and output signal waveform, with time on the horizontal axis and output level on the vertical axis. In the figure, the part corresponding to signal A3 is A6, and the part corresponding to A4 is A? The background level is shown respectively as follows.

FIG. 12(D) is a diagram showing the signal waveform after the level shift, and the background level L2 is shifted by a certain shift amount from the level L1.

This compression circuit 726 can also obtain a compressed signal by multiplying the image data by an appropriate coefficient, and preferably, in response to the output of one image data, the compressed signal value level corresponding to the table memory 732, for example, a table RAM is stored in advance. is written in advance and configured to output an appropriate compressed signal according to the input signal. Further, the shift amount in the shift circuit 730 is determined in advance by setting an appropriate shift amount in accordance with the output level of the compressed signal in a table memory 734, for example, a table RAM.
It is preferable to write the threshold value TH, read it out according to the input level, and output it as an appropriate threshold level TH. However, these configurations can be modified as appropriate depending on the design.

Note that the table memory used here is the CP of the control unit 180.
The memory of U 182 may be used or may be provided separately.

The threshold level thus set and the image data are sent to the comparator 720, and the image data is binarized. The relationship between the signal waveforms in this case is shown in FIG. 12(E).
In the figure, the image data to be binarized is shown by a solid line ``■'', and the threshold level is shown by a broken line ``■''.

In this way, in this embodiment, the binarization threshold is set from the image data itself to be binarized, so even if the background of one image data fluctuates, the threshold level can be easily set. can be set, and can be binarized accurately.

Third Method> FIG. 13(A) is a block diagram for explaining the third method of the image signal processing method for binarization of the present invention. This example is a method that combines the first and second methods, and thus is a method that combines the methods of FIG.
The configuration includes the circuit configuration shown in ). However, 1
Matrix memory unit 702 for reading out pixel data in parallel. (Line memory section 703, latch section 706
．． Since the address counter 704) is common to the blur correction circuit 700 and the smoothing circuit 722, it is not necessary to provide them separately, and it is sufficient to share one address counter.

Therefore, in this third method, image data to be binarized with a waveform such as that shown in FIG. On the other hand, the threshold level obtained from the image data through the smoothing circuit 722, compression circuit 726, and level shift circuit 730 (see FIG. 11(C)),
(see FIG. 11)) can be supplied to the comparator 720.

Therefore, both signals are inputted to the comparator 720 in FIG. 13 (B
), they are compared and provide a binary signal.

According to this third method, images with fine, high-frequency components can be binarized more accurately and with high resolution.

The image data processing described in the first to third binarization processing methods described above can be performed for each line, but when binarizing with the comparator 720, the average processing for several lines is performed. For example, this average value calculation circuit (
) is provided in the comparator 720, and the CP of the control unit 180
It can be configured to be performed under control from U182.

The binary signal obtained in this way is sent to interface 2.
60 (FIG. 1) is output for further processing.

It is clear that the present invention is not limited to the configuration of the embodiments described above.

For example, although the above-mentioned embodiments have been described with reference to negative films, they can be similarly applied to positive films with just a few appropriate changes.

Furthermore, the configuration of each circuit can be any other suitable configuration depending on the design. Further, in the above-described embodiments, the explanation was centered on digital processing, but if necessary, all or part of the circuit configuration may be replaced with an analog circuit to perform the processing.

Furthermore, in the circuit configuration of the embodiment described above, the control unit 180
For example, address lines, buses, and other control lines from the CPU 182 are omitted from illustration.

Further, although the embodiment of the present invention has been described as an example in which it is applied to a microfilm reading device, it may be applied to any type of device 44 that includes a device that performs two-dimensional scanning with an image sensor, reads the data, and converts it into a digital signal. It can be applied.

(Effects of the Invention) As is clear from the above description, according to the image signal processing method related to the optimal threshold for binarization processing of the present invention, image data obtained using an image sensor can be smoothed. , the minimum or maximum value of the background is determined from the smoothed signal, the corresponding film density is determined from this minimum or maximum value, and the optimal threshold is set based on the obtained background density. Even if the background density changes, it is possible to automatically set the binarization threshold value more easily and efficiently and accurately than the conventional setting of the binarization threshold value.

[Brief explanation of drawings]

FIG. 1 is a block diagram for explaining the image signal processing method of the present invention, FIG. 2 is a schematic diagram showing a microfilm reading device to which the present invention is applied, and FIG. 3 is a component of FIG. 2. An explanatory diagram of a microfilm scanner; FIG. 4 is an explanatory diagram of an optical disk recording device which is a component of FIG. 2; FIG. 5 is an explanatory diagram for explaining two-dimensional scanning of an image of an image sensor. FIG. 6 is a block diagram mainly showing circuit parts for film density measurement and threshold level setting, for explaining an example of a binarization circuit, and FIG. 7 is a signal waveform diagram for explaining the present invention. FIG. 8 is a diagram showing the relationship between film density and the minimum value and binarization level of the signal passed through the low-pass filter; FIG. 9 is a signal waveform diagram showing an example of input image data used to explain the present invention; Fig. 10 is a block diagram for explaining three methods of binarization according to the present invention, which are configured to accurately binarize even high frequency components of an image, and Fig. 11 is a block diagram of the method shown in Fig. 10. FIG. 12 is an explanatory diagram for explaining the first binarization method; FIG. 12 is an explanatory diagram for explaining the second binarization method in FIG. 10; FIG. 13 is an explanatory diagram for explaining the third binarization method shown in FIG. 100... Microfilm scanner 104...
Image sensor, 134... Sensor unit 148... Motor, 150... Drive circuit IEi2... Drive reading circuit, 180... Control unit 300... Projection image 320... Shading correction circuit 340. ...Binarization circuit, 360...Interface section 600...Film density measurement section 602...Threshold value setting CPU 804, 720...Comparator, 606...Low pass filter 610...Multiplexer 812 ... Minimum hold circuit 614 ... Peak hold circuit 81 [1, 728, 732, 734 ... Table memory 700 ... Blur correction circuit 702 ... Matrix memory section 703 ... Line memory section 703a~ 703c...Line memory 703d...
Write control unit, 703e...read selector 703f
~703h, 70Ela~708e-Latch circuit 70
6... Latch section, 70B... Multiplication circuit 710
... Addition circuit, 712 ... Subtraction circuit 722.
... Smoothing circuit 724 ... Median filter 726 ... Compression circuit 730 ... Level shift circuit. Patent applicant: Fuji Photo Film Co., Ltd. Togono, 46th Moon's picture 2
Sun-moon diagram Figure 1 Time siete 1 shi γ yearly output of regular 4 table 4s flow diagram Ginmon LPF nature nigga 4 word passing diagram Figure 7 t P F , aAt & * h 4% r-f f:h
n1ktr-Kamari 2 4 I and Noru Haruka Mo 1 Incoming power switch r-ta's No. 1 waveform Figure 9 Explanation of binary A IrLr = Jl - 龜 ゛ - When the number of A of 2 is drawn out, P - 1 island 3 ↑; Explanation of the time binary value IB Figure 12

Claims (4)

[Claims] (1) In setting the optimal threshold for performing the binarization process of image data, the output of the image data obtained by scanning the projected image of the negative film to be read with an image sensor is smoothed, and the output of the image data obtained by scanning the projected image of the negative film to be read is smoothed. A minimum value is determined from the output distribution obtained by smoothing, a background density of the image data is determined from the minimum value, and an optimal threshold for binarization is set according to the background density. Image signal processing method. (2) The optimum value for binarization is stored in advance in a table memory in association with a background density value, and the optimum threshold value is read out in accordance with the background density value. The image signal processing method according to scope 1. (3) When setting the optimal threshold for binarizing image data, calculate the maximum output from the output distribution of the image data obtained by scanning the projected image of the positive film to be read with an image sensor. Image signal processing characterized by determining a background density of image data for setting the threshold value from the maximum value, and setting an optimal threshold value for binarization according to the background density. Method. (4) The optimum value for binarization is stored in advance in a table memory in association with a background density value, and the optimum threshold value is read out in accordance with the background density value. The image signal processing method according to scope 3.