FI105419B - Identifier and procedure for identifying objects - Google Patents

Description

105419

Identifier and method for the identification of an object Identifierare och förfarande vid identifiering av föremäl 5

The invention relates to a sensor and a method for identifying an object.

It is known that soft drink, beer and similar bottles are transported and stored in baskets. Baskets are usually the property of a particular brewery or similar and are used on a 10 return basis.

Under certain circumstances, baskets belonging to different owners need to be identified and sorted. Otherwise, the baskets may have the same structure and shape, but differ from each other in the logo on the side surface. The identification and sorting of such baskets had to be performed manually before, which has been time consuming and thus costly.

In the most general embodiment of the invention, the object of the invention is specifically the identification of objects. As used herein, objects generally refer to products, packages, cans, etc. A particular object of the invention is the identification of a basket containing a bottle return and in particular a return bottle in connection with a return basket machine.

It is also an object of the invention to provide an apparatus and method which is as reliable as possible and does not impose requirements on the exact position of the car, i.e. the distance of the car from the side of the conveyor. Another object of the invention is to provide an apparatus and method in which baskets can be categorized on the basis of small details even when. their main logo is similar, but they have some sign or pattern differences on their side which allow them to be distinguished.

30 This application discloses an identification apparatus suitable for color separation of all objects, products, packaging, etc. The invention is particularly suitable for identifying baskets, such as bottle baskets. The invention is particularly suitable for use in connection with 105419 2 return basket dispensers, whereby the dispenser automatically detects not only the basket but also the return bottles in the basket and provides the return receipt and / or return money and / or deposit or other similar document corresponding to the baskets and bottles. The apparatus according to the invention is located protected from light, so that the outside light of sleep is not obstructed in the identification of the body.

According to the invention, the basket is placed on a conveyor and moved at a constant speed past the sensor. The sensor transmits a light strip to the surface of the basket, and the light beam reflected from that light strip back to the device is examined based on the proportions of the three colors 10 red, green, and blue. At each two-millimeter longitudinal point of the longitudinal side wall of the basket, a sample is taken from the surface of the basket to obtain the average color of the basket at that point. The hardware does not need to separate colors separately in the vertical direction. The sample is taken approximately 100 times per second.

In the device solution according to the invention, a beam of light is produced on the surface of the basket and said beam extends over a substantial part of the height of the side surface of the basket. During car color identification, the conveyor is moved at a constant speed such that a light bar wipes a substantial portion of the car's sidewall.

In the method of the invention, after the above number, a comparison is made to the preprogrammed baskets data in the device memory. The counting algorithm preferably performs a test based on a particular color of the base color of the body, the logo of the body and preferably still of the side surface of the body. In the method and device solution according to the invention, one side surface of the basket is considered. The light beam runs vertically 25 and is perpendicular to the plane of the conveyor, preferably a belt conveyor. Preferably, the color of the basket side begins immediately at the front edge of the basket and ends about 100 mm before the second vertical edge of the basket.

! t

To achieve the foregoing and further objects, the object detector of the invention and the method of object identification are characterized in what is set forth in the claims.

3 105419

In particular, the device according to the invention enables the baskets to be reliably identified and distinguished on the basis of their color, which results in considerable cost and time savings.

The device according to the invention also makes it possible to automate the sorting of baskets in various applications.

In the following, the invention will be described in more detail with reference to the figures in the accompanying drawings, in which details, however, are not intended to be limited in any way.

Figure 1A is a side elevational view of a device arrangement according to the invention.

Figure IB shows the device solution of Figure IA, the direction of the arrow k.

15

Figure 1C shows the device of Figure IA, the direction of the arrow k2.

Figure ID shows the device of Figure IA, the direction of the arrow k ^.

Figure 2 is a schematic representation of the main component of the device according to the invention.

9 «

Figures 3A, 3B and 3C show a stepwise operation of the device according to the invention.

Fig. 4A shows a light strip, or light beam, which is a vertical, preferably 25 parallelogram illuminated area on the side wall of the basket in step 3A, Fig. 4B a light beam in Fig. 3B and Fig. 4C in a light beam in Fig. 3C.

Figure 5 is a voltage curve illustrating a color-coded side of the body.

30 105419 4

Figure 6A shows the number of the color matrix of the basket and the prior measurement of the distance of the basket.

Fig. 6B shows an identification unit for an illustrative color recognition device, which is a single color component having a color filter and comprising detectors.

Figure 6C illustrates a color matrix of the sidewall of the basket.

Figure 6D shows the conversion graphs used to normalize the color codes.

10

Figure 7 illustrates the apparatus in connection with a recovery basket machine.

Figure 1A is a side elevational view of a device arrangement according to the invention. The invention is generally applicable to the identification of objects. However, the invention is particularly applicable to the identification of baskets. The basket 10 is placed on the conveyor 11 and is moved forward by the top of the conveyor belt 11 to the reading area 200 of the color separation unit 14 (direction arrow = Nj). In the figure, the reading area 200 consists of a dashed pyramidic space in which the device 14 can read the light beam 12 produced on the surface of the car 10 It is essential that the light beam 12 is produced on the side surface 10a of the car 10. 20 In a space delimited by the pyramid 200 at any space boundary or within the space, the color separation unit 14 may read a light beam 12 produced from a light source 13 on the side 10a of the basket 10.

Figure IB shows the device solution of Figure IA, the direction of the arrow k, that is, from the top. As shown in Figure IB, the basket 10 has reached the reading area 200 and the beam 12 is produced from the light source 13 of the 25 color separation units 14. The light source 13 produces white light and is led through the optical cables 18ai, 18a2,18a3. Optical fibers! 18aj, 18a2-- - is divided vertically into a beam of light and from which parallel ends of the fibers the light is next directed to the lens 19. It forms a flat light line or beam of light directed to the side surface 10a of the body 10. Thus, a light beam from the light source 30 13 is provided on the surface of the basket 10 and the reflected light beam S2 is received by the detectors 17a, 17b and 17c.

105419 5

Figure 1C shows the equipment of Figure IA, the direction of the arrow k. As shown in the figure, the pyramid-like reading area 200 is adapted to the device solution so that the detector apparatus 17a, 17b, 17c receiving the reflected light beam B2 is located mid-height relative to the height Sj of the basket 10. This provides the best possible numerical accuracy. As illustrated in Figure 1C, the optical fibers 18a1, 18a2, 18a3, 18a4 and l5a4 are arranged at their ends so that they are in the same vertical position and the white light produced therefrom is guided and assembled by the lens 19 to form a flat light strip

Figure 10 ID is shown in Figure IA, the direction of the arrow k3 in a perpendicular to the transport direction of the conveyor with respect to Li. The illustration is illustrative, and as shown in the figure, the detector apparatus 17a, 17b, 17c is located at the tip of a reading area, or reading pyramid. All the beams 12 in the area defined by the side edges and the bottom of the pyramid, as well as the beams 12 reflected thereon, are readable by the detector device 17a, 17b, 17c of the invention. The light beam 12 15 is arranged to be transmitted on the body surface such that it will extend substantially along the length of the side wall of the body in a vertical direction.

In the exemplary embodiment shown in Fig. 2, the conveyor 11 has a basket 10. Against the side edge 10a of the baskets 10, a beam 12 is produced from a light source 13 that emits a white beam of light 20 against the surface of the basket. The light beam Sj is reflected from the side surface 10a of the basket and the reflected light beam S2 is transferred to a color separation unit 14 where a color code is provided. The code is compared in the memory 15a of the central processing unit 15, or so-called. stored color codes in the library, and so if they match, Body Type is recognized. The color separation unit 14 comprises filters 16a, 16b, 16c, followed by detectors 17a, 17b, 17c, which measure the red, blue and green portions of reflected light S2. By generally identifying objects such as packages, cans, packages, etc., the codes of the objects to be recognized prior to identification are programmed into the memory 15a of the central processing unit 15.

In the exemplary embodiment of Fig. 2, the basket 10 is on a conveyor 11. At a suitable distance from the light source 13, a beam of white light Sj is transmitted through a fiber optic cable 18, i.e. lenses 19aj, 18a2 --- and 6 105419, to the side surface 10a of the basket 10 and its color separation unit 300. The color of the basket is measured and the color codes (R, G, B) of the basket 10 are compared with the pre-stored color codes (R ', G', B ') of the various baskets in the memory 15a of the central unit 15. If the 5 measured data corresponds to the stored data, the body type is recognized and feedback, i.e. the receipt / money or the like, corresponding to the body type is given.

As shown in Figure 2, a beam of white light S i is transmitted from the light source to the surface of the basket 10. Light Sj reflected through the fiber optic cable is processed in its width direction D so as to obtain three partial samples. Samples aj, 82 ^ 3 are filtered through filters 16,16a, 16b, and 16c, where samples are filtered for red (R), green (G), and blue (B). The filter 16a passes through the red (R) light, the filter 16b the blue (B) light and the filter 16c the green (G) light. Filter 16a is followed by detector 17a, filter 16b 15 is followed by detector 17b and filter 16c is followed by detector 17c. Detectors 17a, 17b, 17c detect the proportion of each light component (R, G, B) in the sample, e.g., by measuring the intensities of the filtered light components. This gives a measure of the proportion of each color and thus produces a trivalent color code for the measured color. The resulting color code is compared with the known color codes in the memory 15a of the central processing unit 15 and as a result of the comparison 20, the type of the basket 10 is determined on the basis of its colors. In this way, the basket 10 can be identified and sorted.

Figures 3A, 3B, 3C show the movement of the basket through a reading area 200 (the so-called reading pyramid). In Fig. 3A, the identification of the car from the vertical edge of the car is started. The car body 25 from area "A" to A "" can identify the overall color of the car.

•

Fig. 3B shows a light strip, or light beam 12, already in the center of the basket where the logo is usually located, and in Fig. 3C there is a longitudinal direction of the end of the car, where reading can be stopped before the end of the car.

30 105419 7

Fig. 4A is a reading step corresponding to Fig. 3A, in which the bar 12 is located at the inlet edge of the basket.

Fig. 4B shows the position of the beam 12 corresponding to Fig. 3B at the 5 colored logos (LO) on the surface of the body and Fig. 3C shows the position of the beam 12 in the end region of the side wall 10a of the body 10.

When a device according to the invention uses a pyramid or truncated pyramid-shaped reading area and an elongated light strip or light beam 12 produced from a light source 13 emitting white light 13 through fiber optic cables 18a, 18a2, 18a3 · .. to the side surface 10a. An advantage of the solution is that it is not very accurate as to the position of the basket 10 relative to the belt run 11a of the conveyor 11. Most importantly, the basket is located along its longitudinal axis parallel to the longitudinal axis X2 of the conveyor 11, thereby avoiding the so-called. angular errors that would reduce readability.

15

The apparatus 14 reads the side edge 10a of the basket 10 from the color reading point A 1 to the reading point A2. The reading is done by reading the colors of the lightbar at regular intervals with a dense sampling, for example, every two millimeters from the end of the reading range A 1 to A 2 of the basket. In this way, a substantial portion of the side surface of the basket is read and an average color code combination R, G, B corresponding to each of the read longitudinal positions of the basket is produced. The code in question: (R, G, B) can be produced as a voltage value and further as a numerical value.

The logo recognition (LO) in the basket 10, for example, in the position of Figures 3B to 4B, may be performed to detect colors R, G, B from a particular length 25 of the sidewall 10a of the basket e.g. from A 'to A ". predetermined by the LO position of the logo so that the test area, or sample, represents the most typical location of the logo. Finally, the average of the color values of the color bars measured from the range A 'to A "representing the" LO "logo code of that basket type is calculated. The measurement procedure is similar for example for the overall basket color (range A'" to A "").

105419 8 Thus, when the central processing unit 15 is transferred to the memory 15a, the so-called "central processing unit 15". basket identifiable baskets are programmed to identify that baskets by entering the characteristics associated with each baskets in specific regions, and thus, when identifiable baskets are being tested, the received / measured aliasing algorithm-5 rhythm compares with the identification information in memory 15a; the average color color codes to match, within a given tolerance, some color code combination of a particular basket type in the memory of the central processing unit, accepting the above identification and marking the identification of the basket 10 as completed.

10

Preferably, three different tests are performed to identify the body type. From the measured basket, the algorithm identifies a) the basic color of the basket. This can be observed, for example, by observing the front edge of the vertical edge of the basket. The second test is performed at the logo (LO), that is, a recognized area of the body is tested and compared to the so-called "memory" of the machine 15. (c) as a test, a color matrix of a character, a letter, a series of numbers from a particular distance region can be tested and compared to a set of values of the corresponding region in the library, memory 15a. If a basket that meets all three tests is found in memory 15a, i.e. a library, said Basket Type is identified as that type stored in the library memory, and the basket returner is notified and / or the basket is further processed based on the authentication performed. If the basket is not recognized, the transport direction of the conveyor belt can be reversed and the basket returned to the starting point. On the basis of the basket identification, the basket returner may be given a receipt or refund or the like corresponding to the basket return price.

In the apparatus solution and method of the invention, the basket is supplied with a color recognition point. through the belt at constant speed by moving the conveyor belt 11a at constant speed. The identification of the basket takes place by continuously moving the belt and thus the basket on it. Thus, in determining the color matrix in the measuring step, the basket is not separately stopped, but the measuring step is uniform and the movement of the basket in the measuring process 30 is smooth.

11Ι54Ί9 9

Figure 5 shows the signal information for one of the measured basket types by color. Breaking curve f | represents the measured red signal strength over the distance of the basket wall. The breaking curve f2 represents the measured green signal strength and the breaking curve the measured blue signal strength over the length of the sidewall of the body.

5

Fig. 6A illustrates a detector apparatus according to the invention, in which the light beam 12 and the light reflected therefrom are illustrated by means of a detector unit 300 of a color recognition device 14 consisting of a single unit comprising adjacent filters 16a, 16b, 16c and detectors 17a. , 17b and 17c. The detection unit 300 is one integral component.

Fig. 6A illustrates a measuring device at the front of the 500 color recognition device 14 with respect to the direction of car movement. The measuring device 500 measures the distance of the basket from the recognition unit 300. Thus, the distance of the light beam 12 on the surface of the basket from the detection unit 300 is measured specifically. Each machine is separately calibrated to determine the conversion curves for each machine.

In the process according to the invention, the so-called. the voltage values from the various detectors 20 17a, 17b and 17c of unit 300 are normalized to reflection percentages scaled from 0 to 100.

•

Fig. 6B shows a color separation unit 300 of a color separation device 14 comprising adjacent filters 16a, 16b, 16c and each detector 17a, 17b and 17c associated therewith, from which the voltage indicative of each color can be immediately transferred to further processing, where the voltage can be changed as a numeric value. Filters 16a, 16b, 16c are coupled with associated detectors 17a, 17b and 17c into a single integrated chip.

Figure 6C illustrates the formation of a color matrix from the color values of the side 10-nanometer 10a of the basket 10. The device 500, which measures the distance of the color recognition unit 300 from the side surface 10a of the basket 10, is located in front of the unit 300 with respect to the carriage direction.

105419 10

The meter can work on any principle. It can be ultrasonic, electronic, light transmitting / receiving, etc.

The transformation from that shown in Figure 6D is as follows. The voltages output from the unit 300 from its detectors 17a, 17b and 17c are read. Let us consider, for example, the red light, i.e. the voltage output from the detector 17a. The voltage value is read from the horizontal coordinate system X of Figure 6B and when the distance of the basket measured by the device 500 from the color sensor unit 300 is known, a R or G or B related graph corresponding to the distance read from the diagram in Figure 6B is selected. to the normalized scaling value in vertical coordinate Y.

Thus, the voltage always coming from the detector 17a or 17b or 17c indicative of a particular color corresponds to a curve corresponding to a certain measured distance and the normalized value 15 read through it. The figure shows the conversion of voltage y 'through the graph fQ' associated with the green color (G) and the distance 1 to the code value X '.

When the baskets 10 are read into the device memory 15a, in the reading step, the distance of the baskets is measured to the measuring device 500 and a conversion of the color values from the detector in said 20 read mode is performed to the conversion graphs fG'fB tuned to that machine.

. using color coded values normalized to 0 to 100.

Similarly, by reading the color of a restored basket or otherwise treated sidewall and producing a color matrix of the sidewall 10a color, denoting R, G, B values at different points on the 25 sidewalls, i.e., the length of the sidewall, these values are converted to normal values by the following conversion.

The conversion curves can be calculated for the machine at certain regular distances that occur in the transport of 30 cabs, and the distances and intermediate values are obtained mathematically, for example, by interpolation.

105419 11 Thus, regardless of the distance of the basket, a comparison of the color values of the restored basket with the color values in the basket memory can be performed based on the following modification and measurement of the distance of the basket.

The sensor 14 according to the invention is ideally suited for use with recovery basket trucks in their light-protected basket tunnel. Thus, the device of the invention can be used such that the return basket automaton calculates the returnable bottles in the basket and the detector device of the invention detects the type of the basket. It is clear that the device can also be used separately for sorting baskets alone and for separating 10.

Fig. 7 shows a detector according to the invention and its color separation apparatus 14 in connection with a recovery basket automaton 600. The detector 14 is mounted in the basket passageway T of the apparatus 600 and the apparatus P for detecting the number of returnable bottles 15 in the basket 10 and their number, also measuring the distance of the basket from the detector 14 on the conveyor side.

Claims (16)

105419 12 A sensor comprising a light source (13) for producing light on the surface of an object to be detected and for detecting light produced on the surface of the object in a device solution, characterized in that the sensor comprises a light source (13) for producing a uniform light bar, preferably a light bar and the light beam (12) produced on the object surface is detected by means of a color separator (17) which receives a light beam (B2) reflected from the object surface and that the signals produced by the color separator (17) provide different color surface areas of the object; comparing the color codes of the identifiable objects stored in the memory (15a) of the central unit (15) of the device, the comparison identifying the object and the apparatus comprising a conveyor (11) for passing the identifiable object (10) past the color recognition point a slider (500) for measuring the distance of the surface of the object to be detected from the 15 detector units (300) of the color separation unit (14). Sensor according to Claim 1, characterized in that the object to be recognized (10) is a basket. Sensor according to claim 1 or 2, characterized in that it is a light strip. i.e., the light beam (12) is produced from the light source (13), preferably from the light source producing the white light, through the optical fibers (18a1, 18a2, ISa1 ...). Apparatus according to the preceding claim, characterized in that the light fibers (18a1, 18a2 -) are fitted at their ends into the light beam (G) so that the ends of the optical fibers are parallel and the light beam emanating therefrom forms a light strip formed by the lens (19). as a uniform beam (12) on the side surface (10a) of the object (10). 105419 13 A basket sensor according to any one of the preceding claims, characterized in that the light bar or light beam (12) extends vertically on the surface of the object (10) and that said light beam (12) has a constant width (D i) ranging from 5 mm to 20 mm. . Basket detector according to one of the preceding claims, characterized in that the light receiving color separation unit (14) comprises filters (16a, 16b and 16c) followed by detectors (17a, 17b, 17c), wherein the filters (16a, 16b, 16c) located adjacent to each other and adapted to take a sample (aj, 82 ^ 3) from each light beam at each measurement, which sample is passed through a filter (16a, 16b, 16c) to a detector (17a, 17b, 17c) associated with each of the 10 filters; is adapted to pass through red light to an associated detector (17a) which produces a voltage value and a further numerical value corresponding to said signal, and that a filter (16b) which transmits green light is associated with a detector (17b) which measures and changes the intensity of said light component. voltage, and that 15 filters (16c) that pass through the blue light are associated with that light component, a detector (17c) which converts the proportion of said light component of the sample into a voltage value which can be further converted to a numerical value, thereby generating an average color code (R, G, B) for each side of the basket corresponding to that side of the basket. 20th Body detector according to one of the preceding claims, characterized in that the filters (16a, 16b, 16c) of the color separation unit (14) and the associated detectors (17a, 17b, 17c) are arranged side by side (relative to the conveying direction of the object to be detected) . 25 Basket detector according to any one of the preceding claims, characterized in that the detector unit (16a, 16b, 16c, 17a, 17b, 17c) of the color separation unit (14) is adapted to receive / read a light strip produced on the surface of the object (10) shape and at the tip of which is a sensor unit (300; 16a, 16b, 16c, 17a, 17b, 17c) of a color separation-30 bone unit (14). 14 A basket detector according to any one of the preceding claims, characterized in that the apparatus comprises a central unit (15) and a memory (15a) therein. a library preprogramming the various color codes (R ', G', B ') of the object type recognition areas comprising the logo (LO) on the surface of the object on the side surface of the object (10) (area A' - A "); and preferably further, an area indicative of the basic color of the article on the side surface of the article, and further preferably a third region on the side surface of the article, which area comprises a feature of a side surface of any of said identifiable objects such as letter Body detector according to one of the preceding claims, characterized in that the distance measuring device (500) is located before the color separation unit (14) and before the point where the light beam (12) hits the side surface of the body. 10 E 4 Ί 9 An object identification method, characterized in that method 15 detects an average color of the side surface of an object to be recognized at different longitudinal points of the object side (10a) over a certain region (A '- A') and calculates color codes from said region (A '- A) ") the average color code and comparing the average color code thus obtained with the codes of the corresponding region of the object to be identified by algorithm or other means in the memory 20 (15a) of the CPU (15) storing information on different object types recognizing their color codes - A ") on the surface of the object to be identified is selected so that it corresponds to the range of the logo (LO) on the surface of the object to be identified (10). Method according to claim 11, characterized in that one reading area on the surface of the object (10) is selected to correspond to the average base color of the side surface of the object (10) on which the color is read. 13. A method for detecting an object (10) according to claim 11 or 12, characterized in that the method transforms voltages indicating different colors produced from the detectors (17a, 17b, 17c) by means of the conversion curves (^ R ^ G ^ B ···) · 105419 15 normalized color code values, whereby the conversion curves are assigned to each machine separately and that there are / created conversion curves (% \ ίο '>%' ^ R '^ G'%) for each color (R, G, B) and for each basket (10 ) from the detector unit (300), the normalized color code being produced by means of the distance measured in the detection event and 5 when storing the color code values in the machine memory (15a) and reading the identifiable basket (10). Method according to claim 11, characterized in that one reading area on the surface of the object is taken from the distance of the side surface of the object, which has a specific character 10 such as a letter, a number, a pattern or the like. A method according to any one of claims 11 to 14, characterized by an identifiable object being a basket (10). Method according to Claim 15, characterized in that the method moves the identifiable basket (10) and that the basket is placed on the conveyor so that its longitudinal axis (Xj) is substantially parallel to the longitudinal axis (X2) of the conveyor (11) and providing a uniform light bar or light beam (12) on the surface (10a) of the identifiable body (10), and that the light beam (12) is arranged to run vertically 20 on the side surface of the body and that light beam (12) is detected. by means of a color separation unit (14). 16 iut »4 iy