DE102018217582A1 - Method and system for identifying a workpiece - Google Patents

Abstract

The invention relates to a method for identifying a preferably plate - shaped workpiece (8), in particular a sheet metal, comprising: illuminating an identification area (26) on a surface (8a) of the workpiece (8) with illuminating radiation, taking an image of the identification area (26), which contains a workpiece-specific reflection pattern of illumination radiation reflected at the identification area (26), and identifying the workpiece (8) by comparing the recorded image of the identification area (26) with at least one of a plurality of images stored in a storage device (33), at least partially the identification area (26) overlapping reference area (32), each containing a workpiece-specific reflection pattern and which were recorded on the surfaces (8a) of a plurality of workpieces (8). The invention also relates to a system (34) for identifying a workpiece (8).

Description

The present invention relates to a method for identifying a preferably plate-shaped workpiece, in particular a sheet. The invention also relates to an associated system for identifying a workpiece.

If workpieces, for example plate-like metallic workpieces (sheets), are processed on several different processing machines or processing stations, it is advantageous to provide a respective workpiece or sheet with a coding for easier tracking. The coding can be applied, for example, in the form of a linear bar code or a 2D code by laser marking to the surface of the workpiece. The coding can also be applied to the workpiece in other ways.

In the DE102017202628A1 describes a method for coding and a method for identifying a plate-like workpiece, the method for coding comprising: defining a coding region, applying a magnetic layer to the surface of the workpiece, and fusing parts of the magnetic layer with the surface of the workpiece by applying the workpiece with radiation within the coding range. The method for identifying the workpiece comprises the steps of the method for coding and additionally: storing the coding in a memory device, reading the coding by measuring the local distribution of the magnetization in a reading area, and comparing the stored coding and the read coding.

The DE102017202629A1 also describes a method for coding, in which a chemical reaction of a base material of the workpiece is caused by irradiation of the workpiece in a coding area and a reaction material is formed within the coding area which has a magnetic remanence which differs from the base material.

In the DE102017202630A1 describes a processing machine for the shaping processing of a plate-like ferromagnetic workpiece, which has a processing head and a write head for coding the workpiece within a defined coding range. The write head is set up to generate a magnetic field which is used for the uniform alignment of magnetic domains of the workpiece within the coding area.

In the EP2875932A1 describes a method and a device for producing a three-dimensional workpiece which has an information code pattern. The information code pattern is determined by the microstructure of the workpiece, which is generated by irradiating a powder during the additive manufacturing of the workpiece. The information code pattern can, for example, form a linear bar code or a two-dimensional matrix code, for example a QR code.

In the DE102005022095A1 describes a method and a device for determining a lateral relative movement between a machining head and a workpiece during machining of the workpiece. In the method, the surface of the workpiece in the region of the machining head is illuminated with optical radiation, and optical radiation reflected from the surface of the workpiece is repeatedly detected with an optical detector in order to obtain optical reflection patterns of the surface of the workpiece at different times. The lateral relative movement is determined by comparing the successive reflection patterns.

Object of the invention

The object of the invention is to provide a method and a system which enable particularly simple identification of a preferably plate-shaped workpiece, in particular a sheet metal.

Subject of the invention

This object is achieved according to the invention by a method of the type mentioned at the outset, which comprises the following steps: illuminating an identification area on a surface of the workpiece with illuminating radiation, taking an image of the identifying area which contains a workpiece-specific reflection pattern of illuminating radiation reflected at the identifying area, and identifying the Workpiece by comparing the recorded image of the identification area with at least one of a plurality of images stored in a storage device of at least partially overlapping reference areas, the images each containing a workpiece-specific reflection pattern and recorded on the surfaces of a plurality of - typically identical - workpieces were.

In contrast to the methods described above, it is not necessary in the method according to the invention to code onto the Applying the surface of the workpiece, rather the surface of the workpiece itself serves as a code for uniquely identifying a specific workpiece. Here, the one quoted in the introduction DE102005022095A1 described fact exploited that workpieces typically have a surface structure or surface roughness that produce an irregular reflection pattern when illuminated. The workpiece roughness required for this purpose is automatically given in most metal workpieces in particular. For taking a picture, it is advantageous if the reference area is shown enlarged. For example, the magnification can be selected such that a length of approximately 10 μm on the surface of the workpiece corresponds to a pixel on the detector surface, for example a high-resolution camera.

According to the invention, the fact is used that the surface structure and thus also the reflection patterns generated during the illumination are workpiece-specific, i.e. that these make it possible, in the manner of a fingerprint, to identify the workpiece as a rule unambiguously, or to uniquely assign the workpiece to one of the plurality of workpieces for which images of a respective reference area have been stored in the storage device. The majority of workpieces, of which images of the reference area are stored in the storage device, are typically workpieces that are to be processed on several processing machines or processing stations of a group, between which the workpieces have to be transported and, if necessary, stored temporarily, which is why a clear assignment is cheap or necessary.

In one variant, the method comprises: illuminating the reference area on the surface of the or one of the plurality of workpieces with illuminating radiation, recording the image of the reference area which contains the workpiece-specific reflection pattern of illuminating radiation reflected on the surface of the workpiece, and storing the image of the reference area in the storage device. To take the picture, the surface of the workpiece is illuminated at least at the reference area. The position of the reference area on the surface of the workpiece is usually fixed. An image (“template”) is recorded from the reference area, which has a predefined size, e.g. 256 x 256 pixels or 32 x 32 pixels on a spatially resolving detector, for example in the form of a camera. The picture can be a grayscale picture, but it is also possible to take a colored picture. Interference radiation can be eliminated by suitable bandpass filters, which are arranged between the spatially resolving detector and the reference area.

In one variant, when illuminating the reference area and / or the identification area, an illuminating power of the illuminating radiation is set or regulated. In this case, the illuminating radiation is generated by an illuminating source whose output can be adjusted. The illumination or the illumination power can be adapted to the reflection behavior of the respective workpiece, for example with the aid of a regulation of the illumination, in order to generate an illumination intensity that is optimal for the spatially resolving detector. In order to regulate the lighting power, the brightness of the image can be determined on the spatially resolving detector, for example in the form of an average of the brightness or the intensity of all pixels of the recorded image. In this case, the lighting power is set or regulated in such a way that the brightness of the respectively recorded image is always the same size or is within a predetermined brightness interval. Comparing images that have a similar brightness is particularly advantageous if a pattern recognition algorithm or a correlation algorithm is used that does not take into account different brightness levels, as is the case, for example, with the SAD ("sum of absolute differences") algorithm.

The lighting, the recording of the image of the reference area and the storage of the image are carried out for a plurality of workpieces which are to be distinguished from one another using the method.

In a further development, the illumination and the recording of the image of the reference area and the illumination and the recording of the image of the identification area take place with two different, preferably identical sensor devices. The two sensor devices can be formed, for example, on different processing machines in a combination of processing machines, or on two different processing stations on one and the same processing machine. The sensor devices are connected to one and the same storage device in signaling connection, which can be formed, for example, on a central computer or in the cloud and in which the images of the reference areas of the workpieces to be machined are stored. The illumination and recording of the image of the reference area and the illumination and recording of the image of the identification area can alternatively also be carried out with one and the same sensor device.

In one variant, the illumination and recording of the image of the reference area and / or of the identification area takes place through processing optics, in particular through focusing optics, of a processing head for processing the workpiece by means of a processing beam. In this case, the reference area or the identification area is usually illuminated coaxially or only slightly inclined to a machining beam which is aligned with the workpiece. Such illumination has proven to be advantageous, since with this type of illumination, illumination radiation reflected from a respective surface area generates very irregular, workpiece-specific reflection patterns in the image on a spatially resolving detector, which facilitate the comparison between the reflection patterns. Coaxial observation of the surface area, for example with the aid of imaging optics, has also proven to be advantageous for carrying out the comparison. In this case, the detector, which can be a high-resolution camera, for example, is typically also arranged coaxially with the processing beam or with an extension of the beam axis of the processing beam.

In a further variant, the image of the reference area and / or the image of the identification area is / are recorded by means of a telecentric image. When the image is recorded by means of a telecentric image, the imaging scale is retained even when the distance between the surface of the workpiece and the detector or the detector surface changes. This is advantageous since the distance between the workpiece and the imaging optics used to record the image generally has tolerances and is therefore not exactly constant. In this case, the imaging optics are designed such that they are telecentric, at least on the object side. This can be achieved, for example, by arranging an aperture diaphragm in the image-side focal plane of a converging lens of the imaging optics.

In a further variant, when comparing the image of the reference area and the image of the identification area, a similarity measure between the image of the identification area and the image of the reference area is preferably determined by a pattern recognition algorithm. In this variant, the image of the or a respective reference area and the image of the identification area are correlated in order to determine a similarity measure. Known image processing methods are used for this purpose: For example, difference-based similarity functions (Sum of Absolute Differences - SAD, median absolute deviation - MAD, Mean-squared-Difference - MSD), cross correlation, correlation coefficient or phase correlation can be used. With regard to the comparison of two images or the reflection patterns contained therein with the aid of a cross correlation, reference is made to the one cited at the beginning DE102005022095A1 referenced, which is made by reference in its entirety to the content of this application. The workpiece to be identified is assigned to the workpiece whose stored image has the greatest similarity to the recorded image, that is to say the image in which the degree of similarity is at a maximum. The similarity measure can also be determined at the sub-pixel level, as in the DE102005022095A1 is described.

In a further development, the workpiece is only identified if the similarity measure between the image of the identification area and the image of at least one reference area exceeds a minimum value. In the event that the similarity measure does not exceed the minimum value for any of the stored images, this indicates that the workpiece to be identified does not match one of the plurality of workpieces for which images of the reference area have been stored, or that an unambiguous assignment is not possible is. By using the minimum value or the threshold, incorrect assignments can be avoided when recognizing the workpiece.

In one variant, the identification area is chosen to be larger than the reference area. It has proven to be advantageous if the identification area is larger than the reference area, since the workpiece may not be arranged exactly at the same position and / or may not be exactly aligned when the image of the identification area is taken when the image of the identification area is taken , especially if two different sensor devices are used. By selecting a larger identification area, slight lateral displacements of the reference area to the identification area can be corrected. However, it is not absolutely necessary for the identification area to be larger than the reference area, in particular the identification area and the reference area can have an identical size. The positions of the identification area and the reference area on the surface of the workpiece are typically selected such that the reference area lies within the identification area, i.e. that the (larger) identification area includes the respective reference area.

In one variant, when the image of the reference area is compared with the image of the identification area, a rotation and / or a lateral displacement of the reference area relative to the identification area and / or becomes different magnification of the image of the reference area and the image of the identification area taken into account. A lateral offset between the image of the reference area and the image of the identification area can be taken into account, for example, with the aid of a cross-correlation, as was mentioned in the introduction DE102005022095A1 is described, since this is a pattern recognition algorithm which is tolerant of a lateral displacement. Alternatively, a (slight) lateral offset can be taken into account by displacing the image of the reference area laterally for comparison with the identification area and determining a similarity measure for different values of the amount of the lateral offset. The image of the reference area can be rotated relative to the image of the identification area by means of a rotation-tolerant pattern recognition algorithm, for example optical flow. Alternatively, it is possible to rotate the reference area relative to the identification area into different angular positions using a suitable rotation algorithm and to determine a similarity measure for a respective angular position. To identify the workpiece, the maximum degree of similarity of a respective reference area is used, which was determined during the displacement or rotation.

In particular, in the event that the image of the reference area and the image of the identification area are recorded with two different sensor devices, the respective image can optionally be recorded with a different imaging scale. The information about the sensor device, more precisely about the imaging scale with which the respective image is recorded, can also be stored in the storage device. The different magnification of the images can be taken into account, for example, by suitable scaling of the image of the reference area and / or of the image of the identification area, so that two images with the same magnification are compared with one another.

Another aspect of the invention relates to a system for identifying a workpiece, comprising: at least one sensor device with: an illumination source for illuminating an identification area on the surface of the workpiece with illuminating radiation, and a spatially resolving detector for recording an image of the identification area, which has a workpiece-specific reflection pattern contains illumination radiation reflected on the surface of the workpiece, a storage device in which a plurality of images of at least partially overlapping, preferably smaller, reference areas of surfaces of a plurality of workpieces which contain workpiece-specific reflection patterns and an evaluation device for identifying are stored of the workpiece by comparing the reflection pattern of the recorded image of the identification area with at least one of the plurality of images rn the reference ranges is formed or programmed.

While the sensor device is arranged or can be arranged in the vicinity of the workpiece, the evaluation device and in particular also the storage device can be arranged at a basically arbitrary location. In particular, in the event that the system has more than one sensor device in order to record images of workpieces, it is necessary that the recorded images are stored in a common memory device which can be accessed by a respective evaluation device or possibly an evaluation device common to all sensor devices . The workpiece can alternatively be identified in a local evaluation device in or in the vicinity of a respective sensor device which can access the memory device. The system can form a closed network of processing machines which are arranged at one location, but it is also possible that the processing machines or processing stations of the network are distributed over several locations and have to be loaded for transport.

The system can have two or more sensor devices, and the illumination and recording of the image of the reference area and the lighting and recording of the image of the identification area can take place on two different sensor devices. This is particularly favorable or necessary if the workpieces to be identified are transported to different locations in order to enable clear assignment in this case too.

The or a further sensor device can serve or be designed to illuminate the reference area on the surface of the workpiece with illuminating radiation, to take an image of the reference area which contains the workpiece-specific reflection pattern of illuminating radiation reflected on the surface of the workpiece, and the image of the reference area to be transferred to the storage device in order to store it there. The sensor device can, for example, be permanently connected to a processing head of a processing machine, which can be moved relative to a respective workpiece, so that the identification area or the reference area can be reached in a particularly simple manner by means of the sensor device.

In one embodiment, the sensor device is designed to illuminate the identification area on the surface of the workpiece and to take the image of the identification area through processing optics, in particular through focusing optics, of a processing head for processing the workpiece using a processing beam.

In this case, the sensor device can be a sensor device that is already available for monitoring a machining process, for example a laser welding or laser cutting process, and can also be used to identify workpieces. It is not absolutely necessary that the sensor device for illuminating the identification area on the surface of the workpiece and for taking the image of the identification area is designed essentially coaxially to the processing beam by means of processing optics. Rather, both the illumination source and the spatially resolving detector on the processing head can also be arranged outside the processing optics or outside the housing of the processing head.

In a further embodiment, the evaluation device is designed or programmed to preferably determine a similarity measure between the image of the identification region and the image of a respective reference region by means of a pattern recognition algorithm in order to compare the image of the reference region and the image of the identification region. As described above in connection with the method, the workpiece is identified by assigning it to the image of the reference area that has the maximum similarity to the image of the identification area. As described above, it is advantageous if the workpiece is only identified if the degree of similarity exceeds a minimum value.

The system, in particular a respective sensor device, can be designed to select the identification area larger than the reference area. The image of the identification area can cover, for example, the entire detector area of the spatially resolving detector, while the image of the reference area covers only a partial area of the detector area. The evaluation device can also be designed to take into account a rotation and / or a lateral displacement of the reference area relative to the identification area when comparing the image of the reference area with the image of the identification area.

Further advantages of the invention result from the description and the drawing. Likewise, the features mentioned above and those listed further can be used individually or in combination in any combination. The embodiments shown and described are not to be understood as an exhaustive list, but rather have an exemplary character for the description of the invention.

• 1 eine schematische Darstellung eines Ausführungsbeispiels einer Laserbearbeitungsmaschine zum schneidenden Bearbeiten eines Werkstücks mit einem Identifikationsbereich und einem Referenzbereich,
• 2a,b Darstellungen eines Laserbearbeitungskopfs der Laserbearbeitungsmaschine von 1 mit einer Sensoreinrichtung zum Beleuchten sowie zum Aufnehmen eines Bildes des Identifikations- bzw. des Referenzbereichs,
• 3a,b eine Darstellung eines Bildes des Identifikationsbereichs sowie eines Bildes eines Referenzbereichs eines ersten Werkstücks,
• 4a,b eine Darstellung des Bildes des Identifikationsbereichs sowie eines Bildes eines Referenzbereichs eines zweiten Werkstücks, und
• 5a,b eine Darstellung des Bildes des Identifikationsbereichs sowie eines gedrehten und lateral versetzten Bildes des Referenzbereichs des zweiten Werkstücks.

Show it:

• 1 1 shows a schematic representation of an exemplary embodiment of a laser processing machine for cutting processing a workpiece with an identification area and a reference area,
• 2a, b Representations of a laser processing head of the laser processing machine from 1 with a sensor device for illuminating and for taking an image of the identification or reference area,
• 3a, b a representation of an image of the identification area and an image of a reference area of a first workpiece,
• 4a, b a representation of the image of the identification area and an image of a reference area of a second workpiece, and
• 5a, b a representation of the image of the identification area and a rotated and laterally offset image of the reference area of the second workpiece.

In the following description of the drawings, identical reference numerals are used for identical or functionally identical components.

1 shows a laser processing machine 1 with a laser source 2nd , a laser processing head 4th and a workpiece support 5 . One from the laser source 2nd generated laser beam 6 is by means of a beam guide 3rd with the aid of deflecting mirrors (not shown) to the laser processing head 4th guided and focused in this and with the help of mirrors, also not shown, perpendicular to the surface 8a of a workpiece 8th aligned, ie the beam axis (optical axis) of the laser beam 6 runs perpendicular to the workpiece 8th . In the example shown, the laser source is used 2nd a CO ₂ laser source. Alternatively, the laser beam 6 for example generated by a solid-state laser.

For laser cutting the workpiece 8th is with the laser beam 6 first pierced, ie the workpiece 8th is melted or oxidized at a puncture position and the resulting melt is blown out. Below is the laser beam 6 about the workpiece 8th moves so that a continuous cutting contour 9 emerges along the laser beam 6 the workpiece 8th severed.

Both grooving and laser cutting can be supported by adding a gas. As cutting gases 10th oxygen, nitrogen, compressed air and / or application-specific gases can be used. Particles and gases can be removed with the help of a suction device 11 from a suction chamber 12th be sucked off.

The laser processing machine 1 also includes a moving device 13 for moving the laser processing head 4th and the workpiece 8th relative to each other. In the example shown, the workpiece is at rest 8th during machining on the workpiece support 5 and the laser machining head 4th is moved during machining along two axes X, Y of an XYZ coordinate system. For this purpose, the movement device 13 a portal displaceable in the X direction with the aid of a drive indicated by a double arrow 14 on. The laser processing head 4th can with the help of another drive indicated by a double arrow of the movement device 13 in the X direction to be moved to any machining head positions P in the X direction and in the Y direction in one due to the displaceability of the laser processing head 4th or through the workpiece 8th given working area to be moved. At a respective machining head position P points the laser beam 6 an (instantaneous) feed rate V on.

As in 2a can be seen, the laser beam 6 for performing cutting work on the workpiece 8th by means of a focusing device in the form of a focusing lens 15 on the workpiece 8th focused. With the focusing lens 15 In the example shown, a lens made of zinc selenide is used to hold the laser beam 6 through a laser machining nozzle 16 , more precisely through their nozzle opening 16a , on the workpiece 8th focused, in the example shown on a focus position F on the top 8a of the workpiece 8th . The laser beam 6 forms an interaction area there 17th with the workpiece 8th , behind the opposite of the machining direction V or contrary to the cutting direction of the laser cutting process the in 1 shown cutting contour 9 is produced. In the case of a laser beam 6 a focusing lens made of quartz glass, for example, can be used from a solid-state laser.

In 2a a partially transparent deflection mirror can also be seen 18th a sensor device 20th which is the incident laser beam 6 (for example with a wavelength of approx. 10.6 µm) reflected and observation radiation relevant for process monitoring to a further partially transparent deflection mirror 19th transmits, which also part of the sensor device 20th forms. The deflecting mirror 18th is partially transparent in the example shown for observation radiation in the form of thermal radiation at wavelengths λ of approximately 700 nm to 2000 nm. A source of lighting 21 the sensor device 20th is used for coaxial illumination of the workpiece 8th with illuminating radiation 22 .

The illuminating radiation 22 is from the further partially transmissive deflecting mirror 19th as well as from the deflecting mirror 18th transmitted and through the nozzle opening 16a the laser processing nozzle 16 through to the workpiece 8th directed. The lighting performance P _B the illuminating radiation 22 the lighting source 21 is adjustable.

As an alternative to the semi-transparent deflecting mirrors 18th , 19th Scraper mirrors or perforated mirrors, which only reflect incident radiation from an edge area, can also be used to illuminate the radiation 22 the workpiece 8th feed. Also at least one laterally in the beam path of the laser beam 6 inserted mirror can be used to enable observation.

As a lighting source 21 diode lasers or LEDs or flash lamps can be provided, as in 2a shown coaxial, but also off-axis to the laser beam axis 24th can be arranged. The lighting source 21 can for example also outside (in particular next to) the laser processing head 4th arranged and on the workpiece 8th be directed; alternatively, the lighting source 21 inside the laser machining head 4th arranged, but not coaxial to the laser beam 6 on the workpiece 8th be aligned. If necessary, the laser processing head 4th even without a lighting source 21 operate.

In an observation beam path 23 the sensor device 20th behind the other partially permeable deflecting mirror 19th is a spatially resolving detector in the form of a geometrically high-resolution camera 25th arranged, also to the sensor device 20th belongs. At the camera 25th can be a high-speed camera that is coaxial to the laser beam axis 24th or to extend the laser beam axis 24th and is thus arranged independent of direction. In the example shown, pictures are taken by the camera 25th in the incident light process in the NIR / IR wavelength range to record the process's own lighting or a thermal image of the cutting process. At the in 2a The example shown can be a filter in front of the camera 25th be arranged if further radiation or wavelength components from the detection with the camera 25th should be excluded. The filter can be used, for example, as a narrowband bandpass filter with a half width of 15 nm, for example, of the wavelengths A transmitted in the range around 800 nm.

To record one in 3a shown image BI one in 1 shown circular identification area 26 on the surface 8a of the workpiece 8th on a detector surface 25a the camera 25th shows the laser processing head 4th an imaging optics 27 on. In the example shown, the imaging optics 27 an aperture 28 on that around a central axis of rotation D is rotatably mounted, so that the position of an eccentrically arranged aperture opening during rotation 28a on a circular arc around the axis of rotation D moved (cf. 2 B) .

Due to the arrangement of the panel 28 in which by means of a lens 29 focused beam path of the imaging optics 27 only part of the observation beam path occurs 23 , which has an edge region of the focusing lens 15 passes through and in the convergent beam path after the focusing lens 15 at an angle β to the beam axis 24th of the laser beam 6 is aligned by the eccentric to extend the beam axis 24th of the laser beam 6 arranged aperture 28a and forms an observation beam 23a which is on the detector surface 25a is mapped. At the in 2a shown example runs an observation direction R1 of the observation beam 23a in the projection in the XY plane or in the workpiece plane parallel to the direction of the machining vector V along which the laser beam 6 and the workpiece 8th are moved relative to one another in the XY plane in order to form the desired cutting contour, that is to say a sharp observation takes place. The angle β under which the direction of observation R1 to the beam axis 24th of the laser beam 6 in the example shown is between approximately 1 ° and approximately 5 °, for example approximately 4 °. The aperture 28 can in particular in the focal plane of the lens 29 or the imaging optics 27 can be arranged to produce a telecentric image.

Instead of a mechanically adjustable cover 28 an electrically adjustable diaphragm, for example in the form of an LCD array, can also be used, in which individual pixels or pixel groups are electronically switched on or off in order to produce the diaphragm effect. The mechanical cover can also be used 28 different from in 2a, b shown across the observation beam path 23 , for example in the YZ plane, are moved or shifted around different parts of the observation beam path 23 shade or open for observation. The aperture 28 can also be realized in the form of one or more mechanical elements that can be opened and closed. Other than this in 2a, b is shown on the bezel 28 can also be completely dispensed with, ie the observation beam path 23 as a whole is on the detector surface 25a pictured.

To do that in 1 shown workpiece 8th The machining head is used to identify before machining 4th and thus also the sensor device firmly connected to it 20th with the help of a control device 30th which are the control tasks of the laser cutting machine 1 takes over and among other things also the motor drives of the movement device 13 controls, over the circular identification area in the example shown 26 positioned. The location of the identification area 26 or the machining head position PI at the identification area 26 relative to the workpiece support 5 and thus also relative to the stationary on the workpiece support 5 positioned workpiece 8th is fixed. With the help of the imaging optics 27 will that in 3a shown image BI of the illuminated identification area 26 on the detector surface 25a generated. As in 3a the machining head position can be seen PI here in the center of the nozzle opening 16a , which is the center of the circular identification area 26 corresponds to that through the processing nozzle 16 through by means of the imaging optics 27 on the detector surface 25a is mapped.

One in 2a shown evaluation device 31 that with the detector 25th has a signal connection, is used to compare the recorded image BI of the identification area 26 with a plurality of images BR ₁ , BR ₂ , ... of reference areas 32 that advance on a particular surface 8a a plurality of workpieces 8th were recorded and which in one in 1 shown storage device 33 are saved. The storage device 33 can be formed, for example, in a central computer or in the cloud, with the evaluation device 31 is in signaling connection. Other than this in 1 is shown, the storage device 33 also in the evaluation device 31 be integrated.

Examples are in 3b and in 4b two pictures BR ₁ , BR ₂ of a respective reference range 32 shown on a first and second workpiece 8th were recorded. Taking the pictures BR ₁ , BR ₂ of a respective reference range 32 of a workpiece 8th can by means of the 1 shown sensor device 20th respectively. Another sensor device is usually used 20th to take the picture (s) BR ₁ , BR ₂ , ... of the respective reference range 32 . The further sensor device 20th can be provided on another processing machine on which a respective workpiece 8th is processed for the first time and to identify the workpiece 8th a picture BR ₁ , BR ₂ , ... of the respective reference range 32 is recorded. Taking a picture BR ₁ , BR ₂ , ... of the respective Reference range 32 can also on a sensor device provided specifically for this purpose 20th be carried out on which no machining of workpieces 8th he follows. The further sensor device 20th forms together with the in 1 shown laser cutting machine 1 , the evaluation device 31 and the storage device 33 a system 34 to identify workpieces 8th .

In the 3b and in 4b shown images BR ₁ , BR ₂ of the reference range 32 were in advance on the surfaces 8a a plurality of workpieces 8th at the same machining head position PI recorded on which also the picture BI of the identification area 26 has been recorded. In the example shown there is a respective reference range 32 smaller than the identification area 26 and due to the matching machining head position PI is the reference range 32 on the workpiece 8th concentric to the identification area 26 arranged.

As in 3a can be seen, contains the recorded image BI of the identification area 26 one for that on the workpiece support 5 resting workpiece 8th specific reflection pattern RI which in 3a is symbolically indicated by several lines and a period. To the workpiece 8th the reflection pattern will be clearly identified RI or the captured image BI with the plurality of captured images BR ₁ , BR ₂ , ... of a respective reference range 32 compared in advance on several workpieces 8th , including the one on the workpiece support 5 workpiece to be identified 8th were recorded. When comparing a similarity measure is first A between the picture BI of the identification area 26 and the picture BR ₁ a reference range 32 of a first workpiece 8th certainly. As in 3b can be seen, contains the image BR ₁ of the reference range 32 of the first workpiece 8th a workpiece-specific reflection pattern R ₁ , for which a similarity measure A with the reflection pattern RI of the picture BI of the identification area 26 is determined.

The measure of similarity A is determined by a pattern recognition algorithm that the two images BI , BR ₁ and thus the workpiece-specific reflection pattern RI of the identification area 26 with the workpiece-specific reflection pattern R ₁ of the reference range 32 the first of the plurality of workpieces 8th compares. With the similarity measure A can be, for example, a difference-based similarity function (Sum of Absolute Differences - SAD, median absolute deviation - MAD, Mean-squared-Difference - MSD), a cross-correlation, a correlation coefficient or a phase correlation.

As with a comparison between 3a and 3b can be seen, shows in 3b reflection patterns shown R ₁ of the first workpiece 8th significant of that in 3a shown reflection pattern RI of the workpiece to be identified 8th from. When comparing the two reflection patterns RI , R ₁ therefore becomes a value A ₁ of the similarity measure A determined, which is very low. As with a comparison between 4a and 4b can be seen, shows the reflection pattern R ₂ of in 4b shown image BR ₂ of the reference range 32 the second of the plurality of in the storage device 33 saved images BR ₁ , BR ₂ , ... a very high correspondence with the reflection pattern RI of the identification area BI on. The value A ₂ of the similarity measure A is therefore larger than that in 3b shown first picture BR ₁ , ie A ₂ > A ₁ applies. In the event that as in 4b is shown a very large value A ₂ for the measure of similarity A identification of the workpiece, if necessary 8th be completed, ie the workpiece to be identified 8th becomes the second saved image BR ₂ of the reference range 32 or the associated second workpiece 8th assigned without comparing with the pictures BR ₃ , BR ₄ , ... of reference areas 32 of more of the majority of workpieces 8th he follows. Alternatively, the comparison in the manner described above with other images BR ₃ , BR ₄ , ... of the reference range 32 continued on a third, fourth, ... workpiece 8th were recorded and this is a respective value A ₃ , A ₄ , ... for the measure of similarity A certainly. The workpiece 8th becomes that of the saved images BR ₁ , BR ₂ , ... of the reference range 32 assigned at which the similarity measure A has a maximum value. At the in 3a, b respectively. 4a, b The example shown is the value A ₂ of the similarity measure A at the in 4b shown second picture BR ₂ maximum, ie that on the workpiece support 5 resting workpiece 8th is called that workpiece 8th identified, on which the second image in advance BR ₂ of the reference range 32 recorded and in the storage device 33 was saved.

To the number of in the storage device 33 saved images BR ₁ , BR ₂ , ... of reference areas 32 and thus the data processing or the number of comparisons with the image BI of the identification area 26 To keep it manageable, those in the storage device 33 saved images BR ₁ , BR ₂ , ... of the reference range 32 be deleted again. The deletion can take place, for example, when the storage period of a respective reference image BR ₁ , BR ₂ , ... in the storage device 33 exceeds a predetermined period of time when a respective workpiece 8th the system 34 has left and / or if a quality control on the respective workpiece 8th was carried out.

To avoid misclassifications when identifying the workpiece 8th To avoid, the example described here also checks whether the comparison with all workpieces 8th certain maximum value A ₂ of the similarity measure A above a minimum or threshold As lies. Only if this is the case will the workpiece 8th identified. Will the threshold As undershot, this is an indication that the workpiece 8th with none of the majority of workpieces 8th matches for the pictures BR ₁ , BR ₂ , ... of the reference range 32 in the storage device 33 are saved.

When taking the picture BI of the identification area 26 and a respective reference range 32 can be due to incorrect positioning of the workpiece 8th the case occur that the reflection pattern R ₂ ' in the picture BR ₂ of the reference range 32 versus the reflection pattern RI in the picture BI of the identification area 26 is slightly twisted, as shown by a comparison of 5a With 5b one can see: Although the reflection pattern R ₂ ' of the picture BR ₂ of the reference range 32 from 5b practically identical to the reflection pattern RI of the identification area 26 from 5a is not a high value if the rotation is not taken into account A ₂ for the measure of similarity A determined so that there may be a misassignment or no identification of the workpiece 8th comes because of the threshold A _s is not exceeded.

To the in 5b to take into account shown rotation, the picture BR ₂ of the reference range 32 in different angular positions relative to the image RI of the identification area 26 can be rotated, for example into several predetermined angular positions, each at an angle of, for example, 2 ° or 5 ° from the in 5a shown orientation differ. For each angular position there can be a value for the similarity measure A be determined and the maximum value of the similarity measure A the angular positions are used for the comparison described above. Alternatively, a rotation-tolerant pattern recognition algorithm can be used to take the rotation into account, for example optical flow.

Also an in 5a indicated lateral offset between the identification area 26 and the reference range 32 that leads to both areas 26 , 32 are not arranged concentrically when comparing the respective images BI , BR ₂ be taken into account by a suitable pattern recognition algorithm which is tolerant of a relative shift, for example by a cross-correlation algorithm. This has a beneficial effect on the reference range 32 is smaller than the identification range 26 and is completely surrounded by it. Even with the in 5a shown lateral displacement is the reference range 32 still completely within the identification range 26 so the entire reference range 32 is available for comparison. Analogous to that in connection with the rotation of the reference range described above 32 may be in different angular positions for several values of the lateral offset between the reference range 32 and the identification area 26 one value each for the similarity measure A determined and a maximum value for the similarity measure A with different lateral misalignments.

Especially when the picture BR ₁ , BR ₂ , ... of a respective reference range 32 and the picture BI of the identification area 26 with two different sensor devices 20th can be recorded with different image scales, if necessary. In this case, the different magnification of the images can be compared BR ₁ , BR ₂ respectively. BI are taken into account by scaling them appropriately for the comparison so that the difference in the imaging scales is compensated for.

Basically, the pictures should BR ₁ , BR ₂ , ... of the respective reference range 32 as well as the picture BI of the identification area 26 if possible with a comparable brightness level to simplify the comparison. For this purpose, the lighting performance P _B the lighting source 21 be set appropriately, for example taking into account the reflectivity of a respective material type of the workpiece 8th . Alternatively, regulation of the lighting output is also possible P _B the illuminating radiation 22 or the (average) brightness of the respective recorded image BR ₁ , BR ₂ , ..., BI to a target brightness possible to take pictures BR ₁ , BR ₂ , ..., BI with a brightness that is as uniform as possible.

Although the process or system 34 to identify workpieces 8th in connection with a laser cutting machine 1 This method can also be used in other machining processes, for example in the welding machining of a workpiece 8th with the help of a processing machine designed for this purpose 1 , for example in the form of a laser welding machine, a processing machine for forming the workpiece 8th Etc.

QUOTES INCLUDE IN THE DESCRIPTION

This list of documents listed by the applicant has been generated automatically and is only included for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 102017202628 A1 [0003]
• DE 102017202629 A1 [0004]
• DE 102017202630 A1 [0005]
• EP 2875932 A1 [0006]
• DE 102005022095 A1 [0007, 0010, 0018, 0021]

Claims (13)

A method for identifying a preferably plate-shaped workpiece (8), in particular a sheet metal, comprising: illuminating an identification area (26) on a surface (8a) of the workpiece (8) with illuminating radiation (22), recording an image (B1) of the identification area (26 ), which contains a workpiece-specific reflection pattern (RI) of illumination radiation (22) reflected on the identification area (26), and identifying the workpiece (8) by comparing the recorded image (B1) of the identification area (26) with at least one of a plurality of in a memory device (33) stored images (BR ₁ , BR ₂ , ...) of an at least partially overlapping with the identification area (26) reference area (32), each containing a workpiece-specific reflection pattern (R ₁ , R ₂ , ...) and which have been received on the surfaces (8a) of a plurality of workpieces (8). Procedure according to Claim 1 , further comprising: illuminating the reference area (32) on the surface (8a) of a respective workpiece (8) with illuminating radiation (22), taking an image (BR ₁ , BR ₂ , ...) of the reference area (32), which the contains workpiece-specific reflection patterns (R ₁ , R ₂ , ...) of illuminating radiation (22) reflected on the surface (8a) of the workpiece (8), as well as storing the image (BR ₁ , BR ₂ , ...) of the reference area ( 32) in the storage device (33). Procedure according to Claim 2 , in which an illuminating power (P _B ) of the illuminating radiation (22) is set or regulated when illuminating the reference region (32) and / or the identification region (26). Procedure according to Claim 2 or 3rd , in which the illumination and recording of the image (BR ₁ , BR ₂ , ...) of the reference area (32) and the illumination and recording of the image (B1) of the identification area (26) are carried out with two different sensor devices (20) . Method according to one of the preceding claims, in which the illumination and recording of the image (BR ₁ , BR ₂ , ...) of the reference area (32) and / or the identification area (26) by processing optics, in particular by focusing optics (15) , a processing head (4) for processing the workpiece (8) by means of a processing beam (6). Method according to one of the preceding claims, in which the image (BR ₁ , BR ₂ , ...) of the reference area (32) and / or the image (BI) of the identification area (26) is recorded by means of a telecentric image. Method according to one of the preceding claims, in which, when comparing the image (BR ₁ , BR ₂ , ...) of the reference area (32) and the image (B1) of the identification area (26), a similarity measure (A ₁ , A _2,. ..) between the image (B1) of the identification area (26) and the image (BR ₁ , BR ₂ , ...) of a respective reference area (32) is preferably determined by a pattern recognition algorithm. Procedure according to Claim 7 , in which the workpiece (8) is only identified if the similarity measure (A ₁ , A ₂ , ...) exceeds a minimum value (As). Method according to one of the preceding claims, in which the identification area (26) is chosen to be larger than the reference area (32). Method according to one of the preceding claims, in which when the image (BR ₁ , BR ₂ , ...) of the reference area (32) is compared with the image (B1) of the identification area (26), a rotation and / or a lateral displacement of the reference area (32) relative to the identification area (26) and / or a different enlargement of the image (BR ₁ , BR ₂ , ...) of the reference area (32) and the image (B1) of the identification area (26) is taken into account. System (34) for identifying a workpiece (8), comprising: at least one sensor device (20) with: an illumination source (21) for illuminating an identification area (26) on a surface (8a) of the workpiece (8) with illumination radiation (22) , and a spatially resolving detector (25) for recording an image (B1) of the identification area (26), which contains a workpiece-specific reflection pattern (RI) of illumination radiation (22) reflected on the surface (8a) of the workpiece (8), and a storage device (33), in which a plurality of images (BR ₁ , BR ₂ , ...) of at least partially overlapping with the identification area (Bl), preferably smaller reference areas (32) of surfaces (8a) of a plurality of workpieces (8) is stored, which contain workpiece-specific reflection patterns (R ₁ , R ₂ , ...), and an evaluation device (31) for identifying the workpiece (8) by comparing the reflection pattern rs (RI) of the recorded image (B1) of the identification area (26) with at least one of the plurality of images (BR ₁ , BR ₂ , ...) of the respective reference area (32). System according to Claim 11 , in which the sensor device (20) is designed, the illumination of the identification area (32) on the surface (8a) of the workpiece (8) and the recording of the image (B1) of the identification area (32) by processing optics, in particular by focusing optics (15), a processing head (4) for processing the workpiece (8) by means of a processing beam (6). System according to Claim 11 or 12th , in which the evaluation device (31) is designed to compare the image (BR ₁ , BR ₂ , ...) of the reference area (32) and the image (B1) of the identification area (26) a similarity measure (A ₁ , A ₂ , ...) between the image (B1) of the identification area (26) and the image (BR ₁ , BR ₂ , ...) of a respective reference area (32) preferably to be determined by a pattern recognition algorithm.

Images