System and method for outputting recognition information of workpiece parts

Abstract

The invention relates to a method for outputting identification information of a workpiece part (21) to be identified from a plurality of workpiece parts (20), each having a cut edge, wherein: for each workpiece part (20) of the plurality of workpiece parts (20), at least one first image (6) of at least one portion (25) of the cut edge of the respective workpiece part (20) is generated; the respective first image (6) is assigned identification information (50) of the corresponding workpiece part (20); for the workpiece part (21) to be identified, at least one second image (7) of at least one portion of the cut edge of the workpiece part (21) to be identified is generated; the second image (7) is compared with the plurality of first images (6); a first image (6) having the greatest similarity to the second image (7) is selected, and the identification information (50) assigned to the selected first image (6) is output.

Description

Technical Field

[0001] This invention relates to a method for outputting identification information of workpiece parts.

[0002] The present invention also relates to a system for performing the method according to the invention. Background Technology

[0003] DE 10 2017 001 915 A1 discloses a method for identifying at least one component without using identification marks. Summary of the Invention

[0004] The purpose of this invention is to provide a method for further improving the prior art.

[0005] This objective is achieved by a method for outputting identification information of a workpiece part to be identified from multiple workpiece parts, wherein each workpiece part has a cutting edge, wherein for each of the multiple workpiece parts, at least one first image of at least a portion of the cutting edge of each workpiece part is generated; identification information of the corresponding workpiece part is assigned to each first image; wherein for the workpiece part to be identified, at least one second image of at least a portion of the cutting edge of the workpiece part to be identified is generated; wherein the second image is compared with multiple first images; wherein the first image with the greatest similarity to the second image is selected; wherein the identification information stored together with the selected first image is output.

[0006] In other words, in the first step, first images of the corresponding cutting edges of multiple workpiece parts are taken. These images of the cutting edges are then used as identification features of the workpiece parts. To identify a workpiece part from the multiple workpiece parts, an image of the cutting edge of the workpiece part to be identified is taken, and this image is compared with the first image. The cutting edge can exist on either the outer contour or the inner contour of the corresponding workpiece part. Preferably, the cutting edge of the outer contour or a portion of the cutting edge is used for identification.

[0007] The workpiece part to be identified is one of multiple workpiece parts. This method is particularly suitable for identifying individual workpiece parts among multiple workpiece parts that are geometrically identical or nearly identical.

[0008] This method makes it particularly simple and reliable to identify workpiece parts using the cut edges.

[0009] Preferably, the workpiece is a sheet workpiece cut from sheet (metal sheet) using a laser cutting machine. For workpieces cut from sheet using a laser cutting machine, the cut edge is particularly suitable for identifying the workpiece.

[0010] Preferably, the portion of the cutting edge in each of the first and second images is selected using a predetermined algorithm. By selecting a portion of the cutting edge, the data requirements are significantly reduced compared to imaging the entire cutting edge. The selection is algorithmically implemented, thereby ensuring that the same portion of the cutting edge is reliably selected for both the first and second images. The selection can be based on the characteristics of the workpiece part. In particular, for workpiece parts cut from sheet material using a laser cutting machine, the starting cut (Anschnitt) can be selected as the portion of the cutting edge to be imaged.

[0011] Preferably, the portion of the cutting edge is selected based on the geometry of the workpiece part. By selecting based on geometry, the same portion of the cutting edge is selected for both the first and second images for workpiece parts with the same geometry.

[0012] Preferably, the portion of the cutting edge is determined based on the enclosing rectangle (enclosing the workpiece part) with the smallest area and the centroid of the workpiece part. In particular, for sheet workpieces, a single portion of the cutting edge can be selected very easily and reliably in this manner.

[0013] Preferably, characteristics of the first and second images are determined respectively, and the characteristics of the second image are compared with those of the first image. By determining and comparing characteristics, the comparison of images becomes more robust, and the identification of workpiece parts becomes more reliable. Image characteristics may include, for example, the geometric relationships of salient points in the image. The geometric relationships of salient points in the image are known, for example, from fingerprint recognition or facial recognition.

[0014] Particularly preferably, features are determined by means of trained artificial intelligence, particularly by means of an autoencoder. The trained autoencoder enables the encoding and comparison of features of an image with cut edges very quickly and reliably.

[0015] Preferably, the second and first images are captured at a predetermined distance from the cutting edge. This predetermined distance makes it easier to compare the geometric relationships within the images. This makes the method more robust.

[0016] The present invention also relates to a system for outputting identification information of a workpiece part to be identified from multiple workpiece parts. The system includes a camera device and a computing unit. The camera device is configured to generate at least one first image of at least a portion of the cutting edge of each workpiece part for the multiple workpiece parts. The camera device is communicatively connected to the computing unit and configured to transmit the first image to the computing unit. The computing unit is configured to assign identification information of the corresponding workpiece part to each of the first images. The camera device is configured to generate at least one second image of at least a portion of the cutting edge of the workpiece part to be identified and transmit the second image to the computing unit. The computing unit is configured to compare the second image with the multiple first images. The computing unit is configured to select the first image with the greatest similarity to the second image and output the identification information assigned to the selected first image via an output interface. The system is capable of identifying a workpiece part to be identified from multiple first workpiece parts and outputting the assigned identification information. The computing unit may be part of a machine tool used to cut out workpiece parts. Alternatively, the computing unit may be part of a factory control system. Alternatively, the computing unit may be part of a remote computing system, such as a cloud. Alternatively, the computing unit may include multiple distributed computing units configured to perform different steps of the method according to the invention.

[0017] The computing unit is preferably configured to compare the second image with a plurality of first images using a predetermined algorithm.

[0018] The system preferably includes a data storage device, wherein the computing unit is configured to store the first image and / or corresponding characteristics of the first image together with correspondingly assigned identification information on the data storage device. By using the data storage device, the information required for identification can be permanently stored.

[0019] The computing unit is preferably configured to determine the corresponding characteristics of the first image by means of a predetermined algorithm, particularly by means of an autoencoder. Attached Figure Description

[0020] The following description of preferred embodiments is provided to illustrate the invention in more detail with reference to the accompanying drawings.

[0021] In the attached diagram: Figure 1 A sequence of methods according to the present invention is shown; Figure 2 A schematic representation of the system is shown; Figure 3 Another schematic representation of the system is shown; Figure 4 This shows the selection of a portion of the cut edge; and Figure 5 A three-dimensional view of the workpiece part is shown. Detailed Implementation

[0022] In all embodiments, elements that are the same or have equivalent functions are labeled with the same reference numerals.

[0023] Figure 1 An exemplary method flow 100 according to the present invention is shown. In a first step 110, a first image 6 of each workpiece part 20 is generated using a camera device 5. In a second step 120, identification information 50 is assigned to each first image 5. In a third step 130, a second image 7 of the workpiece part 21 to be identified is generated. In a fourth step 140, a first image 6 having the greatest similarity to the second image 7 is selected. In a fifth step 150, the identification information 50 assigned to the selected first image 6 is output.

[0024] Combination Figure 2 The schematic diagram of System 1 shown describes the first two steps 110 and 120 in more detail. System 1 includes a camera device 5 and a computing unit 10. Multiple workpiece parts 20 are manufactured by a machine tool 15 (here, a laser cutter). These workpiece parts 20 are nearly identical in geometry and have cut edges on their outer contours. In the first step 110, the camera device 5 of System 1 generates first images 6 of the cut edges or portions thereof. The first images 6 are transmitted from the camera device 5 to the computing unit 10. In the second step 120, the computing unit 10 assigns identification information 50 to each first image 6. In this example, the identification information 50 is transmitted from the laser cutter 15 to the computing unit 10. Alternatively, the identification information 50 can be input into the computing unit 10 by an operator. To enable a simple and reliable comparison between the first images 6 and the second images 7, the computing unit 10 uses an autoencoder 55 in this example. The autoencoder 55 is used to encode the characteristic features of the cut edges in the corresponding images. The computing unit 10 stores the first image 6 and / or the representative features encoded by the autoencoder together with the corresponding identification information 50 in the data storage device 35.

[0025] Combination Figure 3The schematic diagram of System 1 shown illustrates steps 130, 140, and 150 in more detail. In the third step, a second image 7 of the cut edge or a portion of the cut edge of the workpiece part 21 to be identified is generated using the camera device 5 of System 1. The second image 7 is transmitted from the camera device 5 to the computing unit 10. In the fourth step 140, the computing unit 10 compares the second image 7 with a first image 6 and selects the first image 6 that has the greatest similarity to the second image 7. In this example, for comparison using an autoencoder 55, the characteristic features of the second image 7 are encoded and compared with the encoded characteristic features of the first image 6 stored in the data storage device 35. In the fifth step 150, the computing unit reads the assigned identification information 50 from the data storage device for the selected first image 6 and outputs the identification information 50 via an output interface 30. In this example, the output interface 30 is a screen. Alternatively, the output interface 30 can be connected to a machine, particularly a machine tool, such as a bending machine, deburring machine, or marking machine. By connecting directly, the machine can be instructed on how to process the current workpiece part 21 to be identified.

[0026] Figure 4 The selection of a portion 25 of the cut edge is shown. To avoid imaging the entire cut edge of the workpiece part 20 in the first image 6 or the second image 7, it is preferable to select a portion 25 of the cut edge. The selection of the portion 25 of the cut edge is preferably based on the geometric characteristics of the workpiece part 30. In this example, the portion 25 is selected based on the centroid 45 of the workpiece part 20 and the bounding rectangle 40 with the smallest area. For the bounding rectangle 40 with the smallest area, the bounding rectangle with the smallest area is selected from all possible bounding rectangles. Then, the workpiece part 20 is oriented relative to the bounding rectangle 40 with the smallest area such that the centroid 45 of the workpiece part 20 is located below and to the left of the center point of the bounding rectangle 40. From the thus oriented portion, a cut edge of a predetermined length in the lower left is selected as the portion 25 of the cut edge for the first image 6 or the second image 7. For workpiece parts 20 for which the method shown here cannot unilaterally select the portion 25 of the cut edge, other criteria can be used for selection. A preferred criterion is the starting point of the profile.

[0027] Figure 5 It shows Figure 4 A perspective view of workpiece part 20. The thickness of the cut edge is visible in the perspective view. The dashed line surrounding the cut edge portion 25 is selected for the image in either the first image 6 or the second image 7.

[0028] List of reference numerals 100 Methods and Procedures 110 First Step 120 Second Step 130 Third Step 140 Fourth Step 150 Fifth Step 1 System 5. Camera device 6 First Image 7 Second Image 10 Computing Units 15 Laser cutting machine 20 workpiece parts 25. A portion of the cut edge. 21. Workpiece parts to be identified 30 Output Interfaces 35 Data storage devices 40 Enclosing Rectangle 45 Center of mass 50 Identify Information 55 self-encoder

Claims

1. A method for outputting identification information (50) of a workpiece part (21) to be identified from multiple workpiece parts (20), in, Each of the workpiece parts (20) has a cutting edge. For each of the plurality of workpiece parts (20), at least one first image (6) of at least one portion (25) of the cutting edge of each workpiece part (20) is generated. In this process, each first image (6) is assigned identification information (50) for the corresponding workpiece part (20). Specifically, for the workpiece part (21) to be identified, at least one second image (7) of at least one portion (25) of the cutting edge of the workpiece part (21) to be identified is generated. The second image (7) is compared with a plurality of first images (6); Among them, the first image (6) that has the greatest similarity to the second image (7) is selected. The output is the identification information (50) assigned to the selected first image (6).

2. The method according to claim 1, characterized in that, The workpiece parts (20, 21) are sheet workpiece parts cut from sheet material by means of a laser cutting machine (15).

3. The method according to any one of the preceding claims, characterized in that, For each of the first image (6) and the second image (7), the portion (25) with the cutting edge is selected by a predetermined algorithm.

4. The method according to claim 3, characterized in that, The portion (25) of the cut edge is selected based on the geometric characteristics of the workpiece parts (20, 21).

5. The method according to claim 4, characterized in that, The portion (25) of the cutting edge is determined based on the centroid (45) of the workpiece part (20, 21) and the enclosing rectangle (40) with the smallest area.

6. The method according to any one of the preceding claims, characterized in that, The characteristics of images (6, 7) are determined for the first image (6) and the second image (7) respectively, and the characteristics of the second image (7) are compared with the characteristics of the first image (6).

7. The method according to claim 6, characterized in that, The above characteristics are determined by means of trained artificial intelligence, in particular by means of an autoencoder (55).

8. The method according to any one of the preceding claims, characterized in that, The second image (7) and the first image (6) are captured at a predetermined distance from the cut edge.

9. A system (1) for outputting identification information (50) of a workpiece part (21) to be identified from a plurality of workpiece parts (20), the system comprising at least one camera device (5) and a computing unit (10). in, The camera device (5) is configured to generate at least one first image (6) of at least one portion (25) of the cutting edge of each of the plurality of workpiece parts (20). The camera device (5) is communicatively connected to the computing unit (10) and is configured to transmit the first image (6) to the computing unit (10). The computing unit (10) is configured to assign identification information (50) for the corresponding workpiece part (20) to the first image (6). The camera device (5) is configured to generate at least a second image (7) of at least a portion (25) of the cutting edge of the workpiece part (21) to be identified, and to transmit the at least second image to the computing unit (10). The computing unit (10) is configured to compare the second image (7) with a plurality of first images (6). The computing unit (10) is configured to select a first image (6) that has the greatest similarity to the second image (7), and output identification information (50) assigned to the selected first image (6) via the output interface (30).

10. The system according to claim 9, characterized in that, The computing unit (10) is configured to compare the second image (7) with the plurality of first images (6) by means of a predetermined algorithm.

11. The system according to claim 9 or 10, characterized in that, The system (1) includes a data storage device (35), wherein the computing unit (10) is configured to store the first image (6) and / or the corresponding characteristics of the first image (6) together with the corresponding assigned identification information (50) on the data storage device (35).

12. The system according to any one of claims 9 to 11, characterized in that, The computing unit (10) is configured to determine the corresponding characteristics of the first image (6) by means of a predetermined algorithm, particularly by means of an autoencoder (55).

Images