HU196008B - Method and device for digitizing information content of printed or written texts, figures, graphs on an information carrier favourably to the reading machines - Google Patents

Abstract

In the present invention, the information is coupled to a reader having an analogue-to-digital converter, an associated signal processing and control computer, and an information store and / or information display connected to the processed data path of the computer, preferably in a speech synthesizer. It is new in the apparatus that the box (7) of the scanning device (1) is divided into three compartments, in which the cathode ray tube (15) and its control and control means (16) are disposed in the first space portion (8) towards the inside of the box (7). auxiliary drive units (12, 13), preferably a TV monitor, a second portion of the connector (9, walls coated with a light absorbing coating (11) and a second space portion (9) with a screen (16) of the cathode ray tube (15)) opposite and a third part (9, 10) of a separating partition (23) of the separating partition (23) of the cathode ray tube (15) with the optical system elements (17) mapped to the information carrier (6), preferably a lens / lens dispenser and directed to the information carrier (6). an optical-electric signal transducer (18) is installed, while a third part of the wall covered with a light-absorbing coating (11) (10, with a partition wall (23) is wound up) and the third space part (10) is closed by the information carrier (6) from the point of view of the ambient light sources, and the main projector output (14) of the computer (3) is the cathode ray tube (15, auxiliary control unit (15)). 12) is connected to the control input (23) of the controller. (Figure 1) 196008 -1-

Description

Nowadays there is an increasing number of tasks where the various visually readable information must be entered directly from the information carrier into the computer horse, or recorded in the form eg. magnetic tape so that it can be fed into the computer.

Similarly, there is a need for the blind to read written texts aloud to the blind.

To solve the tasks outlined above, various drawing and text digitizers and reading machines are used.

Such a reader discloses U.S. Patent No. 173,648 to ύ HU, wherein the illuminated moving information carrier is scanned with one n-member photodiode array per line, the analogue signal of the photodiode is converted into a digital signal by a segmenting step each, and then processed by the A / D converter.

In this solution, the resolution is determined by the density of the number of ye of the photodiode. The resolution of this solution is limited {cf. resolution of a TV screen) and reading speed is very slow, and the solution is expensive, so it can only scan graphs to a computer.

The solution described in DE 2,847,857 DE solves the problem of increasing the scanning speed by scanning a photocell matrix, scanning the matrix, and processing the signals after the A / D conversion.

Optical character recognition (OCR) technology, which has been around for nearly 20 years, has been revolutionized by the advent of microprocessors and personal computers. In scanners made from a personal computer and an image digitizer, an image digitizer decomposes an image into a dot matrix B-section map, so that the digitizer unit scans the surface of the page to be scanned one by one and classifies each raster dot as dark or light. The output of the digitizer unit displays a series of digital signals corresponding to the dark world signal. The digital signal sequence transforms the raster image into the memory of the computer and the actual recognition process is the interpretation of the raster image. OCR devices have developed a handy procedure for this: the mask fit method and the contour analysis methodBzer.

Also, DE-2 926 673 discloses a reading device of suitable resolution and speed for readers, where the stationary information medium is scanned with a specially designed camcorder and the video signal is fed to a largerBBB processor.

This latter solution, although technically satisfactory to the assignee, results in an extremely complex and expensive piece of equipment in terms of resolution as read speed, and is thus used only as a component of professional equipment.

The object of the present invention is to provide a high-resolution, high-speed reading method and apparatus that can be used to produce a wide range of low-cost, professional, and widely used text-to-speech devices for the general public.

Thus, the present invention is, firstly, a method for digitizing the information content of printed, Ir, texts, diagrams, and graphs on information carriers, preferably for reading machines. In this method, the medium is illuminated by transforming the light reflected from the medium into optical signals by means of an optical / electric transducer, and then comparing this electrical signal to a discrete reference value, converting it to a discrete value and computing and processing it.

SUMMARY OF THE INVENTION It is an object of the invention to generate a point of light to illuminate the information carrier by means of drive and control units on a cathode ray tube screen, preferably on a TV monitor screen, and run across the n is used to map it to the information carrier so that it runs across an area of η xm light spot on its surface.

The present invention also relates to apparatus for digitizing the information content of printed, irote texts, graphs, and graphs on an information carrier, preferably readers, having a reading device coupled to the information carrier, an analog-to-digital processor and a computer for processing signal connected thereto. an information store and / or information display, preferably a speech synthesizer, connected to its data output. The essence of the invention is that the box of your scanning device is arranged in three compartments, the first compartment having a cathode ray tube and a control and even auxiliary unit thereof, preferably a TV monitor, with a screen directed towards the inside of the box. The walls of the second space portion of the connector are provided with a light-absorbing coating.

In a partition facing the cathode ray tube and facing the cathode ray tube, the second space portion is directed towards the information carrier by an optical system azeride (s), preferably a lens / lens system and an information carrier directed towards the information carrier. is installed. The third space portion of the wall having a light-absorbing coating facing the partition is open and configured to be connected to the information carrier. Λ the third space, closed with the information carrier, forms a closed space in terms of ambient light sources. The computer's point-of-light output is wired to the control input of the cathode ray tube control unit.

The invention will be explained in more detail by reference to the drawings, where

Fig. 1 is a schematic diagram of a possible application of the invention,

Figure 2 shows an outline of the image memorisation program

As shown in Fig. 1, the present invention is a Bzorlnti-based reading machine of the reading device 1 having an analog-to-digital converter 2 connected to its electrical signal output, e.g. for reading black-and-white printed texts, a simple jol level sensor, a controller connected to its output, and a signal processing computer 3 consists of 4 speech synthesizers and 5 speakers connected to its processed data output.

The reading device 1 is contained in a box 7, which can be divided into three distinctly spaced sections. In the first space portion 8, a cathode ray tube 15 is disposed so that its umbrella 16 is directed toward the inside of the box, i.e., the second space portion 9. In the first space 8 there is also located a helper driver 13 of the cathode ray tube 15 and a control auxiliary unit 12 controlling the light point generated on the cathode ray tube 15, whose control signal input 24 is connected to the light point control output 14 of the computer. The wall of the second space part 9 is provided with a light-absorbing coating 11 on the inside. The second space 9 is separated from the third space 10 by the screen 16 and is separated by a partition 23 into which an optical system 17, e.g. a lens / lens system, on the other hand, an optical-to-optical converter 18 is incorporated. The optical relay portion of the optical-to-electrical transducer 18 is directed toward the third space portion 10, while its electrical signal output forms, directly or indirectly through an amplifier, the electrical signal image of the leo-display device. The third compartment 10 and the baffle 23 are also provided with a heat-absorbing coating 11, but the opaque side of the baffle 10 is missing b. for books, sheets of paper, graphs etb. gumiharmonikáva! so that the third space 10, when closed with the information carrier 6, forms a closed field with respect to the ambient light sources, so that no interfering light can reach the surface to be read. Kal It is preferable to use a TV monitor instead of 15 palletBuilding units and 12 control units and 13 drive units.

The operation of the apparatus according to the present invention having regard also to the process of the invention is carried out as follows:

Jén On the screen 16 of the cathode ray tube 15, the driver auxiliary unit 13 and the control auxiliary unit 12 generate a light point 19 of a size such that the screen 16 can form a raster of n rows and m columns. This single light point 19 is then run through the control auxiliary unit 12 with the control signals from the light output control output 14 of the computer 3 across the screen m 16 n. The light spot 19 passing through the screen 16 is continuously mapped onto the information carrier 6 by means of the optical system 17, so that a single spot of light 21 is imaged through its surface, which also has a raster point. The area m χ n can be subdivided into the area x b.

These sub-areas eg. the size of a letter. This sub-area resolution is performed on the computer 3 during processing. Z Z1 luminescence running on 6 media on the brightness of the illuminated point, eg. it transmits light of different intensities to white or black depending on the optical-electric converter 18, which converts the luminance signal into an electrical signal.

Thus, in this way, the running light point 19 generated on the screen 1G is copied to the information carrier 6 in the direction 20 from which the optical signal corresponding to the scanned point is transmitted to the optical converter 18 according to the direction 22 and then output 2 analog-to-digital converter, in this case, since only the black-and-white difference has to be detected in a comparator, then from here to the processing computer 3, where we divide the whole axb into sub-pictures and then with reference pictures eg. compares it with images taken from letters to determine the letter read. The electrical signal corresponding to the letter read is stored in the memory 3 through the processed data output of the computer 3, or as a speech synthesizer from a reading machine shown in the figure, and then assembles into read text at the speaker 5.

Thus, it can be seen that a very fine resolution (the resolution of the screen is determined by the size of the pixel 19 of the cathode ray tube 16 and the quality of the optical system 17 used) can be achieved with the aid of our method and apparatus. THE

The operation of the -3196008 (control, processing, speech synthesis juice) requires the use of equipment corresponding to the operation and memory capacity of an average personal computer, so that a complete reading machine according to the invention can be produced at a cost less than previously known.

The operation of the device according to the invention and the computer program controlling it are briefly described below:

The screen resolution of your machine is the same as that of your computer. For example, for Commodore 64, we can distinguish 320 pixels horizontally and 200 pixels vertically.

A letter requires a minimum of 8 lines in vertical resolution. It follows that the optics of the screen projectors must be designed so that this condition is met by opting for optics. For Commodore 64, eg. In the case of a 1.6 mm newspaper letter, this means a line spacing of 1.6 / 8 = 0.2 mm.

Λ the size of the image you are looking at gives 200 * * 0.2 = 40 mm, horizontally approx. We see a section of 67 mm. (Horizontal and vertical resolutions differ slightly.)

To read the entire page, you have to move the viewed surface in some way. The easiest way to move this is by placing and rotating several optics, e.g. on a given C64 machine, three vertically fishy images should be viewed horizontally on the Α / 4-ββ format. A preferred embodiment may be e.g. placing three optics horizontally and moving the information carrier itself in increments of 40 mm.

The computer program that runs the reader operates from a main program and the following subroutines:

Figure 1. Memo program

2. blank line search program

3. empty column finder

4th character identification program

5. speech synthesizer program

The image memory program puts the image image into memory.

The stroke finder searches for stroke-free raster lines between strings. Finds strings with these n lines. After the program runs, the line positions know the program, which is stored in the variable v (n). Where n is the sequence number.

The empty column finder program provides the position of characters within a line. The characters are drawn between these lines. The horizontal raster positions of the empty lines are stored in the variable E (n, u) where n is the rasterized position u is the in-line position.

The character finder program compares the contents of the dot matrix defined by the lines defined above with the values of the predefined font table. Given a margin of error! matches the character and stores its ASCII value according to n lines. Such a well-known program is the Recognita program developed by considerROPISR 16 ~ The talking program is a known program that produces sounds that match the ASCII characters you enter. Such a program eg. Comtnodor C64 is a program called "satnu speaches" that speaks English as you type.

How the main program works:

Reading is done in multiple segments. When you read a new tab, you begin initialization by setting the stigma index.

After adjusting the optics segment, the memorization program scans the text read in raster and stores the scanned values in memory.

The image stored in memory is analyzed by the empty line spacing program and then stored by the raster positions of the blank lines (spacing).

As before, the nz empty column finder finds the vertical fish boundaries of the characters and stores their values in the specified variable E (n, u).

The characters that are searched for in this way are identified by the character finder in the ponmatrix and stored in the corresponding ASCII values according to the string.

You will then see if you have all the segments, if not you will do the above before you finish the bold segment.

After reading the whole page, or after recognizing the texts you have already read, the speaking program utters the text.

Λ After running the program, to read a new page, initialize progremoL and wait for the new page.

Λ a description of how subprograms work.

Figure 1 Memorization Program:

The program has a vertical b point. stores as a perlon.

The program starts with the memory pointer zero. For example, in the case of a C64 computer, the Yo = inlXin / 320 * b ÖBB dependence at the upper yo value of the raster line. The horizontal value of X is given by the relation X - Quality (320).

The program uses a Z variable to determine the vertical bit position. In variable 'i', the tenth position of the bit 1 is always probed within the eight vertical points defined by Y0. I represents the offset of the raster sequence number. The real raster position is X, (Yo + I). The program calculates their values as a function of M, setting i = = 0.

U / These calculate after Y = Yo + I

In the next step, the grid point of X, Y coordinate is lit.

190 008

Is the program looking at the reflected signal level? If it is black, then Z performs an OR operation on its M-content. So, the bit corresponding to the raezler point is set to 1. If the dot is white, then the program performs an AND operation on Z neg, so the corresponding bit is changed to 0.

Ra The X, Y coordinate raezlerpoint is then extinguished and Z-L is moved one bit to the right. Increment with I-L.

They then look at the computer for Z = 0 to see if the eight steps have been completed. If Z is not zero then it jumps back to U.

11a Z = 0 then checks to see if the memory bar has reached its maximum, if not incremental M and jumps back to calculate Yo and restart the cycle described above.

2. Operation of the empty Wine Coring program.

This program manages the memory organized in bytes according to the real pixel, (x, y) pixel value (bool) (X, Y) = ALMOST (y / b) * 320 + XAND2ü (b-ymodb) described storage thousands.

The program sets X, Y to zero,

i. label, checks to see if X, Y is Zero.

If Zero then increases its x-value by then jumps to I label. If it is not Zero then it will examine Y to reach the maximum and if it does not increment Y it will jump to label I.

As soon as X reaches the end of the line, the spacing is stored and stored in the U (n) variable.

3. Blank column search program.

The program is n = 0; Initializes with x = 0. (Λ starts the search at the first line item.)

Label 1; then sets Y = V (n), examining whether (x, y) is white. If Γοhér, examine whether you have reached the next row. If you have reached it, you will lottery, if not you will increase x further, he earns. 11a Y reaches the last line the program ends.

4. Character identification program

As a result of all the efforts so far, we have been able to locate the characters. The contents of a character space form a matrix. I compare these α matrices to a table with a caseB evaluation and identify the character with the correct match (due to minor printing differences) and then store its ASCII value sequentially.

5. Speech synthesizer program.

Several programs are known for this task.

Λ best known α Commodore program named .samu speches. The program works by sending a Basic typing routine ASCII codes to the appropriate storage location of the machine code program, which is then spoken by pressing the return button.

Claims (3)

1.) A method for digitizing printed information, graphics, graphs on an information carrier, preferably for readers, by illuminating the information carrier and converting the light emitted from the information carrier by means of an optical-electric transducer into electrical signals and then converting converting this discrete signal sequence into a computer, and storing and / or displaying the result, characterized in that it is illuminated on a cathode ray tube by means of its drive and control auxiliaries (13, 12), preferably on a TV monitor screen. generating a point of light, then running this point of light (19) through the n and in columns of the screen (16) of the cathode ray tube (15), (19) an optical system (17), preferably a lencBo / lenc.System, that is mapped onto the information carrier (6) so as to run over the surface of the information carrier (6) in an area of n x m lightbeam (21). Method according to claim 1, characterized in that during processing, the electric signal obtained from the area consisting of nxm light points (19) is divided into partial sequences such that the axb light points correspond to the sub-areas, and then the resulting partial drawing series one by one. comparing with reference sequences, preferably an axb symbol made of printed letters, and generating, by matching, a speech synthesizer (4) corresponding to the image, preferably a letter, of the image found, and comparing and generating the following sub-sequence. ' 3.) Apparatus for digitizing the information content of printed, written texts, diagrams, and graphs on an information carrier, preferably having a reading device connected to the information carrier, for converting an analog-to-digital converter associated with its electrical signal output. -511 and control computer and computer processed data memory has an associated information storage and / or information display, preferably a speech synthesizer, characterized in that the reading device (1, box 7) is divided into three spaces, where the first space ( 8) a cathode ray tube (15) and its auxiliary units (12, 13), preferably a TV monitor, the walls of the second portion (9) of the connector being coated with a light-absorbing coating (11) and the second and third portions (9) of the second portion (9) facing the cathode ray tube (15). 9, 10), the optical system (s) (17) directed to the information carrier (6), preferably directed towards the lens / lens system and the information carrier (6). electrical transducers) (18) are provided, while the third space portion (10) of the wall covered with the light-absorbing coating (11) is open to the side of the opaque tea and is connected to the information carrier (6) and the third space portion (10) closed with the information carrier (6), it forms a closed space with respect to the ambient light sources, and the cathode ray tube (15) of the computer (3) has a cathode ray tube (15) RLO auxiliary unit (12) is connected to the control input (23).