JPS62500960A - Method and apparatus for identifying and recognizing objects - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] The present invention is based on the method 7! described in the general concept of the financial I-required range! I am leaving.

In various fields, objects that have been marked with identification marks (fl:M) are is needed. Second 11 Ewo f1 Transfer container is marked with individual identification markings. A symbol (・sign power!) is installed, and from this identification marking symbol, especially the owner and the start point of operation υ; 4 can be read. Also, for example, the beverage industry It is possible to trace the transnephritis 9 of an empty container.

Furthermore, in the industrial field, recognition of workpieces and control of opponents are required. child If such control is performed manually, sufficient reliability cannot be obtained≠Z　t , G’! t.

It will happen. In my experience with such monotonous work, each worker must be attentive. It is known that the power actually declines. In addition, this type of work requires manual labor. It costs a lot of money.

Conventional technology Therefore, there are many methods and methods for recognizing objects based on mechanical recognition of symbols. The device is already known.

Among these, optical scanning methods involve changing the distance between the reading device and the object to be recognized. It has the advantage of being able to

On the other hand, the position of the code relative to the optical sensor does not correspond exactly to the predetermined position. If the code is not available or is used in a rough factory environment, the code may A problem arises in the work place, which is subject to partial changes in schedule.

For this reason, until now, readers have been unable to read the postal code numbers on letters and luggage tags. Character codes are hardly used in industrial applications, except for computer equipment.

Known correlation methods for pattern recognition are are overly sensitive to errors, especially differences in rotational position as well as symbology. the Therefore, known character detection methods require large technical outlays.

For the above reasons, barcodes are mainly used in various industrial processes. As a result, the position of the sign is detected by the minute sequence of bars. There is. In addition, this barcode has a certain degree of tolerance against interference. . In addition, crude symbols in the form of holes and ridges are used; The number is extremely difficult for humans to interpret.

The use of orientation symbols is known per se in the image processing field. However, public In the method of knowledge (DB 3205189A1), it is evaluated because of its special form. A special orientation symbol is used which allows the form to be recognized as such.

However, this method is hampered by errors in the orientation symbol, e.g. mechanical shock. , the orientation symbol may not be recognized. Therefore, the cognitive process may be hindered. be harmed. This is because the orientation symbol recognition is done at the beginning of the whole process. be. In the method according to the invention no special form of orientation mark is required. example For example, a small square, square or disc-shaped point or other geometric shape. be able to. Also, if the code or orientation symbol is partially missing, partially occluded, or defective, In some cases, the functionality of the method according to the invention is not affected within large tolerances.

Description of the invention - The method according to the invention as defined in the characterizing part of the main claim includes the following: Readable codes can be recognized with a small error rate, and deviations in the position of the code can be recognized. and any rotational position can all be detected with relatively little technical outlay, thereby This has the advantage that the recognition process is not affected by the shadow V. Furthermore, the sign symbol Also, even if the orientation mark is partially damaged, it can be read reliably. This gives you the advantage of

In particular, permanently attach character symbols that are easily readable to the naked eye to workpieces, containers, etc. In that case, the nature or characteristics of the segmented disclosure should be included in the claims. However, this symbol can also be read mechanically by the method according to the invention. can be done.

In the field of image processing, the concept of "mask" is generally used to describe related parts of an image. The image cropping means used to screen out the minute and irrelevant parts. trout The disk can have various forms, in particular disc-shaped or rectangular. . If using a video camera, pre-determine the length of the line within multiple predetermined lines. It is expressed by the pulses that appear in each predetermined portion. These pulses H2 The signal is then supplied to a matching circuit along with the converted video signal. In that case, the output of the matching circuit On the power side, only the signals corresponding to the parts of the recognized symbol that are within the mask are generated. .

When implementing the corresponding method using computer technology, it is necessary to use pre-stored cards. Assign addresses to individual pixels (picture elements or bike cells) in the camera image, and The image content is checked for the presence of symbols only at addresses stored as Check. In that case, in order to obtain the desired effect, it is generally necessary to Each address or just a few values and the associated algorithm are stored. It doesn't matter whether it is or not.

For example, in the case of a rectangular mask, remember the corner points, and when using the mask, Checks only points with coordinates located between points, and also in case of a disk-shaped mask. The center point and radius are stored.

For example, using the method of the invention, it is possible to Transporting the object to the recognition device by means of transport known per se; Objects can be recognized with a low error rate despite the inaccuracies of the device.

In one embodiment of the method of the invention, an arrangement is provided for recognizing rotationally symmetrical objects. Before generating the mask of the direction symbol, the first 1 mask and a second mask of the rotational position symbols of the object are generated. 1st masks are required when objects have large positional tolerances. Object recognition device If the rotational position of a rotationally symmetrical object is uncertain when sending it to the is advantageously in the form of a circular ring. The rotational position is approximately given in advance. , a rectangular or disc-shaped mask that encloses the estimated position gi of the rotational position symbol. can occur.

This practical aspect of the method according to the invention is particularly suitable for use in barrels in the beverage industry, especially in the brewing industry. (also referred to as kegs) and containers. in this case , 6 barrels or kegs have a visually readable symbol, especially a pivot or keg It has a code indicating the operation and the number of the barrel or container itself. Bottled or stocked In the fermentation operation, the method of the present invention is applied and data processing equipment is used to produce barrels or Correlate destination and origin addresses when containers enter and exit containers, and You can check the shipping of barrels. Additionally, when the barrels come in, they are It is possible to select barrels that must undergo a certification inspection.

An optical sensor reads the label on the metal cage around the barrel ring or the top pressure of the barrel. It is known how to encode kegs and containers using hole codes that can be used. child The disadvantage of this method is that the barrel must be rotated to increase the bladder as well as the sign The reason is that it cannot be read with the naked eye.

According to another embodiment of the invention (i) K, before generating the mask of orientation symbols, 1 The area GoL center point of the code field with the code is found. This configuration Besides the orientation symbol, there are other optical symbols on the object for detecting the symbol feed A. It can be applied to cases where the system is not provided and cannot be seen with the naked eye. In this case the area The centroid point varies slightly according to the statistical distribution of individual symbols. However, this The variation of is smaller than the play space between the orientation symbol and the mask for the orientation symbol There is an internal pressure within the tolerance range. Therefore, it is possible to subsequently perform orientation symbol detection. .

In another embodiment of the method of the invention, the transport device transfers the object to a code capable of recognizing the code. However, within a predetermined mask, This takes into account the fact that the orientation symbol can be introduced precisely enough to be located at .

According to another embodiment of the method according to the invention, after determining the area centroid, the orientation symbol is Generate a second circular ring-shaped mask, and set the center point k　1llj　52R center of the mask Formed by This ensures that the rotational position of the code field is not ascertained in advance. Even if the orientation symbol is detected, the code field as well as the individual code marks are It becomes possible to detect the position of the number.

In a special configuration of the embodiments described above, a number of orientation symbols have an area of 1114 centroids. The multiple masks of orientation symbols are arranged at different intervals from the center of gravity of the area. are generated at different intervals from each other, in which case at least the first mask has a circular ring shape. and has an area of 7i (concentric with the center).

Depending on the structure of other dependent claims, other methods of the main claim 811 may be used. Advantageous implementations and improvements are possible. Furthermore, the dependent claims include: An apparatus for carrying out the method is described below.

Description of examples FIG. 1 is a block circuit diagram of an apparatus for carrying out the method according to the invention; It is. In the figure, the image of the object to be recognized or the part of the object of interest is electronically The image is captured by camera 1. This camera 1 is equipped with a friction tube or a semiconductor flat sensor. It can be a conventional video camera or also a so-called line camera. camera The output signal of 1 is supplied to camera interface 2. This interface 2 pre-processes the video signal for post-processing. In this case, first the C-video signal is The image is binarized or binarized, and from the many gray values obtained from the camera. , we obtain a binary code that can take only two states depending on the existence of the recognized code. It will be done. The aperture used in binary evolution can be adjusted according to the contrast of the photographed image. be. Techniques of this kind are already known per se and will be described in more detail in connection with the present invention. There's no need to mention it. In addition, the video signal using Monitor 3e is still binary encoded. The reproduction of Gideoi G as well as the display of these two signals is also known.

The camera interface 2 has an address, data and control bus 4 and a It is connected via a video bus 5 to the further components of the device shown in FIG. Figure 1 Referring to , specification 6 includes a real-time gray value analyzer with a mask generator. ing. With this configuration group 6, a specific gray value identification code is obtained from the scan by camera 1. is derived, and the binary threshold and the automatic illumination device q of camera 1 are controlled accordingly. Ru. The mask generator included in configuration group 6 generates a mask and uses this mask to image areas that can be set or adjusted for evaluation purposes. this Various circuits and methods are also known in principle and will not be described in more detail in connection with the present invention. I don't think it's necessary to mention it.

The next configuration group 7 includes image processing devices having various functions. For example, image processing area measurement, center of gravity determination, edge detection, and This can cause problems.

The processor component includes a microprocessor and a program for controlling the method steps. Equipped with RAM memory.

If captured images are made available over multiple image periods, various image processing The physical steps can be carried out in a preferred manner. This is before taking an image with camera 1. Either determine the length corresponding to the object to be recognized, or if it is impossible or impossible. If it is meaningless, store the captured image in the image memo 'J91C. This can be achieved by

FIG. 2 shows a keg moving on a roller path 12 through the field of view 13 of camera 1. is container) 11'? FIG. Figure 3 shows this device. FIG. As shown in FIG. 1, the camera 1 includes an image processing computer 1. Connected to 4. A ring-shaped lamp 15 illuminates the field of view 13 of the camera 1. There is. A symbol 16 is provided on the upper side of the keg 11. 16 is read using the camera 1 and the image processing computer 14.

The light play 18 serves to briefly stop the keg in the three-point positioning device. The three points The positioning device consists of a movable abutment cylinder 91 and two fixed position roll bearings. It is composed of 92°93. Note that the three-point positioning using the optical array 18 is performed within the device. A brief stoppage of the falling keg initiates the recognition method procedure. fixedly provided Gray value scale (scale) 17 or gray value surface of an object that is originally colored The part is processed by an image processing computer to determine the illuminance standard for the evaluation of No. 1 16. get.

FIG. 4 is an enlarged view of the upper side of the keg and the gray value scale 17. roller The path 12 and the positioning are f1t91,92.93, and the center point of the keg 21 is at the target position. Position the keg accurately so that it only deviates a few millimeters from the original position.

If necessary, for accurate center point measurements prior to evaluation of the configuration or orientation symbol. A high quality steel fixture image can be used for the first location recognition. in this case, Use a black rubber valve 20 as a coordination symbol or use a cover lid. If so, a similar mark or dot in the center on the cover shall be used as the orientation symbol. It can be used for initial position recognition. For this purpose, an image processing computer Inside the keg, a disk-shaped mask is placed in a rotational relationship with respect to the target position of the center point 21 of the keg. generated.

It is larger than the outer diameter of the equipment 22, and smaller than the diameter of the equipment 22.

In this way, the rubber valve or the above mark is placed inside the mask to indicate whether the equipment or Appears as a dark area against the bright surface of the retained cap. inside the dark area The center point calculation determines the exact location of the center point of the keg.

To recognize the keg, a code field 23 is provided on its upper side; This code field 23 is filled with the code detailed in connection with FIG. . Two orientation symbols 24, 25 are provided juxtaposed to the code field 23.

A further orientation symbol 26 is provided between the center point of the keg and the code field 23. It is. These symbols can be made bright with a dark background, or vice versa. You can also make it dark with a dark background.

Since the keg is not fixed in terms of its rotational position, the rotational position must first be determined. It will be done. For this purpose, a circular ring-shaped mask 27 is generated. Placed within mask 27 The image portion to be placed is a series of correlated I images that have contrast against the background. i! It is used as the rotational position difference orientation symbol 26 when the cable has a ii cord. The name of this place, The number of pixels corresponds to the width of the rotation position orientation symbol. Rotational position (for the box recognition itself, For example, a position estimation method sound could be used as described below.

In this case, the angular displacement 1 end of each pixel starts from the first pixel of the whole circle and the last 11 Clothes ending in 11 elements are stored in 1fC and are 36<.

Next, the pixels of the circle are sequentially scanned. If the above pixel rows are arranged, then f + - and 1ij One center pixel of the j-thread still has two centers ii! ii element by 1 rotation from the above pixel row ff, ) k To identify, evaluate by reading the relevant angular position from the above table. worth it. However, the accuracy of the recognition is limited by the existing tolerances. receive.

Once the rotational position and center point of the keg has been determined at least approximately; The image processing computer 14 calculates the estimated positions of the orientation symbols 24 and 25.

However, in order to find an accurate position, two masks 30 are used around the estimated position. ．． 31 is generated.

The diameter of these masks can compensate for inaccuracies in traditional methods. chosen large enough to accommodate

However, these masks 30 and 31, for example, must not be so large that the code field 23 falls within the mask and causes an erroneous measurement. stomach. In a practical example, l? i: Center point of direction symbol 24.25 and sign field 23 A radius of mask 30 or 31 corresponding to one-half of the gap between is valid. The score is 190-.

Subsequently, area 1 (center of orientation iji'' No. 24 and 25) is processed using a known 1III image processing method. Follow the law. Orientation symbols 24 and 25 are predefined for code field 23. y) has c in the position given by t6. but (determined) The subsequent method step [T1] will be explained with full reference to FIG. sign - No.) field 23, for example, select from 7 threads/enemy portion or esegment. For example, four characters are represented by selecting, for example, 33,34°35.3 It has 6. These sentences j'' are written by l, fc, and fi with the human eye. Not only can it be recognized by JIV, but it can also be reliably recognized by an image processing computer. . In contrast to the example of No. 41 C1, in FIG. It is arranged and the segment of the letter is adapted to the pressure. rectangular mask 37 is the fourth VC and is a rectangular mask instead of a circular mask. (vinegar· It shows that it can be used. Segment that exists! call and non-existence Segment position j1. i IIC sequential 5i', using another mask for the cow, It is possible to determine which segments exist. Like this I/Cl-te4 ・The information is then written in a known manner, for example, in ASCII code format. is converted into a misleading signal.

In connection with Figures 2 to 5, the method steps "n5 daughter V"+- are as a program. It is stored in the image processing unit 14. This program is used Depending on the method of calculation and taking into account other factors, experts in the field may be able to You can create it with Therefore, below, iD<)'J (required Describe only the gram stages or steps.

*Time the optical array signal to register the presence of the next keg at the station. do. As an application example, the code field may contain 7 segments of D or 9 digits. shall be entered.

Image for gray value analysis in constant scanning mask (elephant imaging and mask formation).

The mask contains a fixedly set gray value base surface at the edge of the image field. Ru. Generation of gray value sum and illuminance change I! The measured gray to compensate for + - Adaptation of the binary threshold for the sum of values.

*Binary image @i[! by real-time storage of information] Li image record.

*Detection of rotational position using rotational position orientation symbol (26) (rough recognition).

*Estimated center point M1' of fine orientation note ++ (24, 25).

M2' word 1 calculation.

*Two squares formed as disks with center points M1', M2' and radius r Measurement of the center of 6V on the inner surface. where r is the center point of the fine orientation symbol and the sign field The claimed area corresponding to approximately one-half distance between The points correspond to the actual center points Ml, M2 of the fine orientation symbols (24, 25).

*Using the locus tf'A transformation, the 7-segment segment is calculated from the position information of MlF5 and Ml. ment center point M S t n: Output.

*Quick cementation using a circular mask with diameter and center point MS scanning. Here, D corresponds to the half-width of one segment (11 tatami mats).

All II inside each circular mask! Li accumulation occurs. (The pixel address of the disk mask is Determine the center point address of each segment using a persistently stored mask expression If you give it out, you will get a lot of benefits).

*All pixels S1. Store jf. Here 1-number 1ci=1. ..., 1o ) index, and index of j-segment notification number (j = 1..., 7).

*Emission of 10fl, 1 flag pie) FB, each with 7 upper bits. Live here, 81, if j is larger than Sg, i-J: /'! bit j of item 1 is set to “1” equally, If ’+j is less than or equal to Sg, then bit j of byte 1 is rOJ is equal to K, and S is about a quarter of the sum of pixels in one cement area. l+J corresponding to the limit value for the pixel sum S at Ru.

Bit 8 of byte 1 is always rOJ.

*Obtained flag bytes and 10 reference bytes for numbers ZO, ..., z9 Comparison with

If the comparison result is positive, derive the relevant ASCII symbol from the reference list and enter the symbol. In the case order ZQ, ..., Z9, the output clothes [11 are included.

*Output of ASCII symbols via serial interface In Figure 6, two letter symbols 41 and 42 are shown, and the segment of letter symbol 42 is 43 is scanned with a mask 44. The mask 44 has a disc shape and has a segment. The image executed by the program mentioned above covers 430 pieces. Generation of elementary sum and checking whether pixel sum exceeds a predetermined limit value Yes, mask 440th position 1u for segment 43 or segment 43 self Inaccuracies regarding body position (wide or narrow) are compensated for.

Additionally, as shown in Figure 6, to improve the readability of the symbols with the naked eye, In order to It can be supplemented to give a correct symbol. Area 4 indicated by diagonal lines in Figure 6 5 is not recognized by the image processing computer 4 except for the evaluation.

In FIG. 7, all ten numerical symbols are shown for complete understanding.

If it is advantageous for reasons of manufacturing technology or for other reasons, the code field may be changed to It can be formed into a nine circular ring shape as shown in the figure. In this case, one symbol 5 1, orientation symbols 52, 53 can be provided in association.

In the embodiment of the method of the invention described with reference to FIGS. The center point of can be located within the field of view of the camera with relatively high accuracy.

If these points cannot be obtained by applying the method of the present invention, other methods of the present invention may be used. All embodiments of can be applied. For details regarding this embodiment, please refer to FIG. Describe. In this case, the object to be recognized is not shown in FIG. 9 within the field of view of the camera. In addition to the symbol, other symbols can be introduced so that they do not appear in the camera liljm. This is a prerequisite.

The code field 55 has, for example, two character symbols 56 and 57. moreover Two coordinate symbols 59 and 61 are provided. Object is within the field of view of the camera After that, the image processing computer should take into account all the symbols shown in The approximate area center of gravity 62 is then determined. A circular ring mask centered on this point 63 is generated. The average radius of this mask is the 7ft center of the above surface M and the orientation symbol 5 9 corresponds to the distance from the center. Therefore, the position of the orientation mark 59 and the The approximate rotational position of the code field 55 [- can also be determined. other discoids Determine the exact position of the orientation marks 59 and 61 by means of a mask, e.g. 58.60. , from these positions, the exactness of the code field 55 and the character symbols 56 and 5T position can be derived.

However, if the rotational position is approximately known, even after the area fish center , the mask for the orientation symbol can be determined as described in connection with FIG. Ru.

FIG. 10 shows an alternative arrangement of the orientation symbols in the λ4 diagram. code field 23 is centered on the axis of rotation 21 of the workpiece or transport means, as in FIG. It is provided on the virtual ring 50. Orientation marks 52 and 53 are on the same radius 54 It is used as a rotational position/orientation symbol, and also as a code feed. This is used to determine the exact position of the lead z3.

Next, in connection with FIGS. 11 and 12, another example of how to carry out the method according to the invention will be described. The device will be explained below.

In this example, recognizing the symbol 69 stamped on the bottom of the bottle It is based on iib. For this purpose, bottle 64ti, camera 1 on roller path 65 passed through the field of view. A lighting device 66 illuminates the bottle from below.

This illumination device 66 operates in a flashing manner, and when the bottle 64 is in continuous motion, This contributes to the generation of tilt@ from which blur caused by motion is removed. Purpose of evaluation The corresponding image signals are then stored as described in connection with FIG. roller path 6 The rollers 67 and 6B of No. 5 transmit light. The identification or recognition process This can be done in the manner described in connection with FIG.

For objects that are subject to rough handling, the VC is large enough that the code is retained permanently. Special consideration must be given to the buttons to ensure high contrast. For example, print .

There are often cases where the given label is not appropriate.

An example of such a permanent code is shown in FIGS. 16 and 14-1. letter Segments of the symbol are represented in the form of ridges on the surface of the object. This kind of prominence These marks can be easily formed on objects that are injection molded from synthetic materials. Wear. The light sources 71 are arranged such that the raised symbol 72 produces a shadow 73. Therefore, there exists an illuminated surface that is not occluded by the bordering segments for the light source. There is enough contrast between the existing segment and the surface being shaded. live.

Another example of achieving high contrast is to place the code on the wall of the object. An example of this is realized as a notch. Figure 15 and Figure 16 are In order to identify the bottle box 75 that is passed through the field of view of the camera 77 at an inclined position on the road 76, The second device is shown. The camera is provided in the form of a notch with the symbol TB gold. The irradiated surface or self-luminous surface 79 is imaged using the valve.

The code distortion caused by the tilted posture of the camera 7T is detected by the image processing computer 1. 4 (Fig. 1) can be easily compensated for by programming accordingly. Ru.

Figures 17 to 19 show the various possibilities of code arrangement in an all-metal keg. shows. In the arrangement shown in Figure 317, the code is provided in the center of the lower surface. The center of the circular surface 81? , (+ is formed by fL, but the same sequence is formed by two different In the examples shown in FIG. 18 and FIG. de) 82 wraps the filling port HV and is in contact with the upper surface of the keg insertion part 84.

In a preferred embodiment, the fill port is also welded to the long V, so that the case is between the loading port and the rondo.

There exists a closed intervening space of 10+83. preferably flat and horizontally formed On the upper surface of the rond, an index mark and an orientation mark are formed by punching.

In order to obtain the desired contrast, the space between the charging port and the rondo is filled with synthetic foam. The sheet is filled with a material, preferably black polyurethane, so that it can be punched out. The symbols formed stand out in dark relief against the light-colored metal surface of the rondo. Go up.

Instead of the rondo, the keg was also placed at the top of an imaginary circle determined around the axis of rolling. Rectangular plates (medals) or pieces of sheet material can be used. In that case, The longitudinal side of the plate or sheet material is aligned with an imaginary ring centered on the keg 1 axis of motion. and extends tangentially.

The upper surface of the keg loading opening is two-dimensionally curved, allowing for adaptation of the board or box. Five possible methods of fixation can be adopted.

1. Gold plate (medal), two-dimensionally produced by pressing, for example, before or after encoding. It deforms and adapts to the upper surface of the keg. In this case, the resulting plate piece (medal) The distortion in ζ on the perspective view of the image is calculated by the necessary process 6 li □ + 1; shall be compensated by.

2. The adaptation of the board is achieved by means of a thick layer of adhesive or synthetic material, ,,The upper surface of the plate remains in the dry state. Place one side of the board in a position where the top surface is covered with water. It is advantageous to supplement the light with a 51/C adhesive or a synthetic material*1 to increase the value fPt.

6. Is the sheet molding material in the shape of a sheet material box with the bottom open? t, 4 (4:, molding The side wall has its free lower edge number ′ζ, adapted to the keg entry surface and welded. It looks like it can be formed. The two sides of the box carrying code information are horizontal The position should be fZi. The rudder is a rondo as shown in Figures 18 and 19. Against the other hand, it would be possible to create an LV type.

For subsequent installation @VC#i, a fork-shaped extension is provided on the sheet molding material. , it is preferable that the arrow extension part is located around the charging port and has one end so that it can be welded to the charging port. It is advantageous.

Another advantageous application of the present process is the use of known polyurethane foam coated beer kegs. One example relates to the encoding of objects lost in synthetic forest materials, such as kegs.

The segment configuration of the code symbol and orientation symbol allows the symbol to be made into a rod of several millimeters. Formation of geometric island areas within the polyurethane foam mantle as a material or cylindrical body The surface can be buried with a keg mantle, thereby being surrounded by synthetic forest material. Together they form a flat surface.

In this way, the integration of objects and codes, which has been difficult in the past, can be easily achieved. simultaneous For example, the segments may be polygonized under program control using manually operated automation equipment. Can be embedded within a urethane foam envelope.

In order to embed the segments inside the plate, the plate has already been made into a white injection molded material. For those who want to set up all the numerical segments in their original state and create all the required breaking points. The law is known. In this case, undesired segmentation may occur in a program-controlled extruder. ment to form a predetermined number. The coordination symbol according to the present invention is also pi1-like. It can be provided on the plate piece depending on the method. For the foam covering of the keg, cover the board with foam. Flat with an orientation symbol with segments buried in piles and not removed and the top side of the container Forms a flat upper surface. The area where the segment was removed is covered with polyurethane foam. do.

One aspect of the generation and installation of code symbols and orientation symbols according to the present invention is , for example, by stamping and embedding a symbol in a particular shaped plate and then printing the plate with polyurethane. Concerning foam embedding within a rethane mantle.

The above will be explained with reference to FIG. 23. In this case, a flat plate made of the relevant material metal or composite that can be made by punching or molding from a sheet of metal A plate piece (medal) 94 made of material is used.

The plate piece has a rectangular cutout 95 in the form of a rectangular cutout 95 having a rotation position orientation n pin on one side. or oval form of goldstone.

The minute 11t, direction symbol 96.97 and number segment 98 are, for example, punched or These symbols and bottle C segments are formed by stamping and forming. It appears in the plate piece 94 as a hole. During the foam coating of containers, the plates are covered with foamed composites. The material forms a flat and smooth surface, and the chain holes 96, 97.9 provided in the plate piece 94 8 is filled with foamed synthetic forest material, 5 is filled with foam, i! I! be played a role. for this purpose Before the foaming process, the plate is placed in the defined position of the foam mold. death However, in the metal processing process of container manufacturing, forming I By welding to metal surfaces, an improved locking connection with the container can be achieved.

In order to generate as high a contrast as possible, the plate 94 is generally sheathed. When the object is corrugated with synthetic material, as is customary in kegs, the coating If the color is dark brown or black (indicated by diagonal lines in Figure 26), it may be white, for example. Constructed from synthetic forest materials or polished high-quality steel.

The special feature of image processing containers is that they have a large number of identification symbols, such as those on alcohol containers. This may occur when a conventional identification code reader is used to evaluate the Ru. Geometrical form and correlation of orientation and code symbols and number and number of code symbols and geometrical form are assumed to be the same. .

Depending on the glaze pressure of the image being generated, the following cases occur when all symbols appear bright on a dark background. (e.g. white symbols made of synthetic material in eight containers covered with black foam material).

Case B All symbols appear dark on a light background (for example, if the rondo is foam covered with black PU) A light-colored metal rondo t-pattern is applied to the base and the symbol is stamped on top of it).

Case C Rotational position and orientation symbols appear in light colors and other symbols appear in dark brown, each on an opposite colored background. (Example shown in FIG. 23).

Therefore, according to a further embodiment of the invention, the original Before code recognition, a disk-shaped mask is created that covers the container surface with an area as large as possible. It is proposed that this occurs. White I in the binary image within the mask! ! Ii elemental total The sum or summation of gray values is determined and the result is a metal keg or Deriving information on whether the keg is covered with foam material.

In the case of a metal keg, subsequent image processing will result in all cross-section image areas being dark areas and The binary image is inverted so that all dark image parts appear as bright parts.

As a result, even in the case of dark brown symbols on a light background, the algorithm described above can be applied as is. \Can be applied.

In the case of a foam-covered keg (in black), rotational position recognition counteracts the binary f-idea. It is executed without any change.

After determining the rotational position according to the above-mentioned hand jlx, the image processing computer The sum of white pixels in the binary image □□□ calculates the sum of grays 114.

Using the values thus obtained, cases A and C described above can be identified.

This is because there are large differences depending on the background brightness.

In case C, binary i[! j image is inverted. In the case of a person, the binary image is Not transferred. Therefore, the process described above will be performed.

The configurations described in connection with the cage and bottle introduction examples with respect to encoding also apply to other transport methods. Also suitable for stage 6 and workpieces. For example, in automobile ceremonial equipment, it is difficult to see later. Attach the identification symbol shown in the shape 1 of the metal plate piece with the riveted holes to the area where to the finishing section that includes the section.

In this case, even after applying lacquer, the stud holes forming the symbol will be completely covered. For example, the code remains readable even in backlight. In the well-known riveter hole encoding The difference between the encoding according to the present invention and the encoding according to the present invention is that readability by the naked eye is guaranteed. be.

20 to 22 show another advantageous design of the code. For example, in Figure 20 , the sign of the character symbol by the 3-point 5-point matrix is shown. This example is an example For example, it can be advantageously applied to encoding conductor plates of electronic equipment. In this case, support The dots are formed as rivet holes in the conductor plate and are visible to the naked eye as well as equipped with an image processing converter. It can be read even under backlight with a high-quality camera. Known rivet hole codes for conductor plates and In comparison, a slightly larger area is required. This encoding allows alpha vents The advantage is that all characters can be displayed.

It is best to insert the conductive plate into the reading device with one edge aligned with the reference circle. It is advantageous. The camera detects the code tack hole under backlight. In that case, the electricity is generated below the conductor plate. Arrange the light surface. The camera has telephoto lenses arranged at regular intervals to To reduce the elemental rarity of the code caused by ovalization due to the perspective effect of the circular hole in the Ru. The conductor plate can also be arranged in the far beam path. The edge of the conductor plate has rivet holes. If not parallel to the raster or pattern, the rotational position of the conductor plate is The centering holes formed by the Ru. To determine the position of the code field in the horizontal and vertical axes, the code Start and end of the code field displaced by one symbol from the field A number of rivet holes or rows of rivet holes are used.

A further application of the method according to the invention includes visually readable symbols and In particular, the symbols provided for automatic g Wdt are common, e.g. the number of blows generated. It is possible to arrange it on one stamping material for. generated in this way An example of the mark is shown in FIG. Such notes can be made using the method according to the invention. A machine-readable mask that scans the individual segments is used to identify the It is generated at the position of symbol 85 or ', and in the area of symbol 86 which is readable to the naked eye. Not generated. Finally, in Figure 22, the symbol 8T is This is an example of admonishing.

Another embodiment of the invention provides additional gray value measurement records in the code field region. Regarding the establishment and evaluation of issues.

Gray value measurement symbols can be used if the print quality of the symbol varies or if the contrast is Determining the optimal threshold for video signal binarization under different conditions and illumination used for.

This is especially true for goldsmiths (whose symbols are etched on the gold Fj4& side). Small against the background, such as when applying a full ink jet to a printed or porous surface. Required for delicate printing methods that produce only small contrasts. death However, if the background etc. changes again, use the method described below for reliable symbol recognition. It is advantageous to apply.

In this case, a rectangular or disc-shaped area, for example, is placed at a fixed level immediately adjacent to the code symbol. area between them in the same way as the other markings, preferably at the same time as the other symbol. Form it so that there are no gaps. Another fixed point in the immediate vicinity of the code symbol , leaving the same size and geometrically same area (hereinafter referred to as "empty symbol"). and set the sign field background to 6th inning.

Gray value measurement symbols and empty symbols should be the same size as the code symbol location or It is advantageous to make it larger.

By the method according to the present invention, that is, the method of reading the code in advance with reference to the binary coded threshold, After detecting the position of the code field, the gray value measurement symbol and the number of empty symbols are It is possible to measure the position in a very concise manner.

Next, a disk-shaped or rectangular mask is set at the center of the two symbols. this square The symbol contour is used whenever the position detection of the hole is inaccurate compared to the symbol. The symbol is selected to be small enough to completely surround it and not overlap the symbol contour.

Next, ii! inside the mask! A sum of 11 gray values is generated in device ρ6.

The sum of the two gray values thus determined (where one is the printed value of the code field) The other represents the gray value of the unprinted part of the code field r. Depending on the value f: input), the adaptation t1 of the binary IAVL is as shown above. This is done so that it is located approximately in the middle of the two gray 1' lines that have been set.

Using the thresholds adapted in this way, we then use vinegar. In this way, high accuracy can generally be obtained. Then, from the method pressure of the present invention, the code Scan symbols.

For example, the symbol 6 is used for certain marking methods such as etching on metals. 5 and 1-1, the trust may change significantly. Therefore in this case requires a contrast measurement for each individual symbol. Generally, fill, : Since the position of the number does not change, there is no need to perform position detection. Furthermore, the background (zu2? Both have the same brightness value.

In this and similar applications (the gray value measurement symbol is gold, each symbol is Mf tangent 1〆t or each? It is best to connect it in contact with the upper %1 or lower side of jli %-. J Tsuru in a meaningful way. In this case, the gray value measurement symbol &′,j can be made relatively small. The gray value 11i11 constant mask and M (may match. This is because the background The gray value of is uniquely defined (y), and after the gray value sum is generated, using a total of n machines, This is because it can be derived. However, in this case, ``Ray value measurement It is necessary to ensure that only the entrance surface of the symbol is completely covered by the mask.

Brief description of the drawing Embodiments of the invention are shown in the drawings and will be described in detail below. drawing During, FIG. 1 shows an arrangement for carrying out the method according to the invention Figures 2 and 6 are diagrams showing a device for properly recognizing a keg; FIG. 4 is a top view of a keg or barrel with a symbol according to the invention; FIG. 5 shows the symbol according to the invention, and FIG. 6 shows the scanning of the symbol using a mask. Diagrams to briefly illustrate; FIG. 7 is a diagram showing the numerical letters and symbols used in the method according to the invention; FIG. 8 is a diagram showing radially eccentrically arranged symbols for a cylindrical object; Figure 9 shows a centrally located code field with orientation marks for a cylindrical object. Figure 10 shows a cylinder with two possible positions for the rotational M orientation. a diagram illustrating an eccentric arrangement of a code field with orientation symbols for an object of 11 and 12 are diagrams showing an apparatus for bottle recognition; Figures 13 and 14 show a code-symbol arrangement in which the symbols are provided in a raised shape. diagram showing, 15 and 16 are diagrams showing a device for recognizing a bottle box; Figures 17, 18 and 19 show the arrangement of codes on kegs. Figure, Figures 20, 2!421 and 22 are for use in the method of the invention. A diagram showing the various codes that can be used, and FIG. 26 shows a medal or medal with a code, an orientation symbol and a further orientation symbol. It is a figure showing a board piece.

Identical parts in the drawings are provided with the same reference symbols.

FIG.6 FIG.8 international search report

Claims (1)

[Claims] 1. In a method for recognizing objects, such as documents, workpieces and means of transport, a) No segmented, clear character symbols that are readable to the naked eye or machine. A permanent code made of clear text is attached to an object, b) Further setting an orientation symbol on said object with a predetermined position for encoding. hair, c) Photograph the code and orientation symbol by an electronic camera and image the output symbol of the camera. Supplied to image processing computer, d) passing through the predicted position of the orientation symbol in the photographed image and increasing the area of the orientation symbol; generate a scanning mask having a rotating area; e) Recognize the symbol present in the mask as an orientation symbol and calculate its coordinates, f) Calculate the coordinates of the individual segments of the code from the coordinates of said orientation symbol, followed by Next generate another mask, g) deriving code information from the presence or absence of segments within said another mask; A method characterized by: 2. Before generating a mask for the orientation symbol to recognize rotationally symmetric objects , generate a circular ring-shaped mask for the rotational position symbol of the object, and the ring-shaped mask The center point of the object is located on the predicted center point of the object. The method described in box 1. 3. of the center point drawn on the object before generating the mask for the rotation position symbol. 3. A method as claimed in claim 2, characterized in that the method comprises the step of generating a mask for. 4. Area of the sign field containing the sign before generating the mask for the orientation sign Find the centroid and use the coordinates of the area centroid to calculate the mask for the orientation symbol. A method according to claim 1, characterized in that: 5. It is characterized by generating a mask for the orientation symbol after determining the center of gravity of the area. The method according to claim 4. 6. A large number of orientation symbols are arranged at different intervals from the area centroid, and generating a plurality of masks for the orientation symbols at intervals of at least one first mask; A claim characterized in that the ring has a circular ring shape and is concentric with the center of gravity of the area. The method described in item 4. 7. The method according to claim 1, wherein the other mask has a disk shape. Law. 8. 2. The method of claim 1, wherein the further mask is rectangular. 9. Claim characterized in that the character symbol is formed from seven elongated segments. The method described in item 1. 10. Claim 9, characterized in that the segments are connected to each other. How to put it on. 11. A segment of a character symbol represents a selected point in a point matrix. A method according to claim 1, characterized in that: 12. In the code field, add the character symbol using the same manufacturing method as the character symbol and orientation symbol. At the same time as the gray value regulation symbol and the orientation symbol, an additional gray value regulation symbol is provided, and the gray value regulation symbol is The symbol allows for the measurement of the contrast of the symbol against the background and also for the orientation symbol. The method according to claim 1, characterized in that the method has a predetermined position. . 13. Segmented text on objects that is both visually and mechanically readable A permanent code consisting of a symbol or clear text is provided and the object is further marked with a permanent code. A product characterized by having an orientation symbol having a predetermined position for encoding. Body identification device. 14. Contrast the symbol (98) and orientation symbol (96, 97) to the material of the object. Provided as a hole in a plate (94) made of a material having a strike, The plate (94) has a curved portion as another orientation symbol (coarse orientation symbol, rotational position orientation symbol). (95), and the plate (94) is embedded in the surface of the object. 14. The apparatus according to claim 13, characterized in that: 15. Signs and orientation symbols are provided in the form of holes formed in the metal plate. 14. The device according to claim 13, characterized in that: 16. The metal plate (82) forms a hollow chamber (83) with the object, and the metal plate (82) forms a hollow chamber (83) with the object. Claims characterized in that it is filled with a synthetic material having a brightness contrast with respect to the Apparatus according to clause 15. 17. Composite with the same or different sign and orientation symbol within the surface of the element made of composite material A request characterized in that the material is embedded as a geometric island region with different colors. The device according to item 13. 18. Device according to claim 17, characterized in that the synthetic material element is a plate. Place. 19. The device according to claim 17, characterized in that the synthetic material element is an object. Place. 20. Identification marks for objects made of synthetic materials produced by injection casting or foaming in molds. Materials that have a contest test against synthetic materials in the method for The code and orientation symbol are provided as holes in the plate piece formed from the plate, and the plate piece is injection molded or or into the mold before foam molding, with the surface representing the sign and orientation symbol against the wall of the mold. contact, and the remaining surface of the board, including the edges, during injection molding or foam molding of synthetic materials. A method characterized in that it is brought into contact with a synthetic material and causes said synthetic material to penetrate into the pores. 21. The object is preferably a polyurethane foam-coated container and a piece of board. as claimed in claim 20, characterized in that it is formed from another synthetic material. Method. 22. The object is preferably a polyurethane foam-coated container and a piece of board. is made of metal and is connected to the container before the foam coating. The method according to item 20. 23. Whether a bright symbol exists on a dark background or a dark symbol exists on a light background. Before the method step to recognize the orientation symbol to determine the object with a large mask Scan and find the number of dark pixels (picture elements) and the number of bright pixels within the large mask. and when evaluating a signal, always compare the symbol and the signal level representing the symbol. The signal inverter is controlled so that the same correlation is given between The method according to claim 1. 24. Next, generate a circular ring-shaped mask for the rotational position orientation symbol, and Find the coordinates of the orientation symbol and create another map that essentially encompasses the region of the sign and orientation symbol. generate a risk, determining the number of dark pixels (picture elements) and the number of bright pixels in the other mask and comparing them with each other; And when evaluating a signal, there is always the same correlation between the symbol and the signal level representing that symbol. Claim 23, characterized in that the signal inverting device is controlled so that The method described in section. 25. a feeding device (roller path, transfer belt) (12) arranged above the feeding device; An electronic camera (1), a sensor (18) that detects a position set in advance for recognition ), and a bearing (92, 93) for setting the position of the object and a pressing device (91) and a positioning device (91, 92, 93) configured from An apparatus for carrying out the method described in item 2.