Method for detecting the hanging position of a support bar and flat bed machine

Abstract

A method is disclosed for detecting the suspension position (15) of a support rod (13) in a tray (3) equipped with a plurality of support rods (13) along a longitudinal direction for use with a flatbed machine tool (1), particularly a laser cutting flatbed machine tool or a plasma cutting flatbed machine tool. Each support rod (13) has a plurality of support rod tips (17) along a main extension direction (13a) oriented transversely to the longitudinal direction of the tray (3). The support rod tips (17) of the plurality of support rods (13) define a support plane (17a). The method includes the following steps: - establishing a two-dimensional contrast image data set of the tray (3), the two-dimensional contrast image data set having multiple image pixels, configuring a pixel value and pixel area unit of a support plane (17a) for each image pixel, and the contrast image data set including a region configured for the tip of the support rod (17) as a local pixel value extreme in a uniform pixel value background, - determining the longitudinal position of the region configured for the tip of the support rod (17) of one of the multiple support rods (13) in the contrast image data set by using the local pixel value extreme, and - deriving the suspension position (15) of the support rod (13) in the tray (3) based on the longitudinal position in the contrast image data set and the size of the pixel area unit in the longitudinal direction.

Description

Technical Field

[0001] This invention relates to a method for detecting the suspension position of a support rod in a tray, the tray being configured for use with a platen cutting machine, particularly a laser cutting platen cutting machine or a plasma cutting platen cutting machine. The invention also relates to a platen cutting machine suitable for implementing the method. Background Technology

[0002] Flatbed cutting machines, particularly laser-cutting or plasma-cutting flatbed cutting machines, typically have a machine area for performing (e.g., laser) processing. This machine area is configured with storage areas upstream and optionally downstream for raw materials and / or cutting materials. In the case of a flatbed cutting machine, a panel of material to be processed (e.g., sheet metal) can be fed onto a pallet to the machine area. The pallet typically includes an arrangement of support rods, by which the panel of material to be processed can be mounted in a support plane. According to an embodiment, the support rods may be laid only at specific locations on the pallet.

[0003] The pallet can be, for example, part of a pallet changer, which represents the storage area upstream of the flatbed machine tool. After processing, the pallet may contain a large number of cut workpieces. Workpieces can be removed from the pallet manually or (partially) automatically during the sorting process. The relative position of the produced workpieces on the pallet to the supporting structure is important for both the production and sorting processes.

[0004] Support rods are typically configured as plates with evenly spaced support tips formed along one edge / side of the support rod. The support rods are inserted vertically with their upward-pointing support tips to form multiple support points for the material panel and the workpiece being cut, these support points defining a support plane. In the inserted state, the multiple support rods extend parallel to each other. The support rods are arranged adjustablely along the longitudinal range of the tray at their spacing and can be suspended, for example, on the tray frame at equally spaced (suspended) positions. Therefore, the arrangement of the support rods on the tray can be configured as needed, i.e., the positions of the support rods along the longitudinal range can be selected specifically for the processing sequence. A support area is formed in the region of the support tips, where the support rod can contact the material panel mounted in the support plane.

[0005] Modern feed-in methods also allow the panel of material to be processed to be occupied by workpieces cut according to a given support rod configuration. Specifically, the workpiece is oriented relative to the tip or support area of ​​the support rod. For example, by avoiding workpiece tilting and subsequent collisions between tilted workpieces and the cutting head of the platen machine, or by employing specific (less destructive) piercing locations for lasers, occupying the panel with workpieces based on the support rod or support area location allows for reduced wear on the support rod tips and improved process quality. The starting point for occupying the panel with workpieces based on the support rod or support area location is knowing the support rod configuration to be used, which is particularly determined by the suspension position of the support rods. However, since it is unknown which support rods are installed, support rod positions cannot be considered for all applications of platen machines in some cases.

[0006] The position of the support rod is known to be detected, for example, by means of a spacing sensor system within the machine area (e.g., fixed to the cutting head) or by laser triangulation. However, such methods result in non-productive downtime due to the "scanning" of empty pallets. Summary of the Invention

[0007] One aspect of this disclosure is to provide a method for detecting the suspension position of a support rod and / or its tip. The method should be operable, particularly under adverse conditions frequently encountered in practice, such as in the presence of uncertain lighting or when the support rod tip is worn or affected by slag. The method should provide the most robust, rapid, and accurate results possible, regardless of the surface condition of the support rod.

[0008] One of the tasks is accomplished by the method according to the invention for detecting the suspension position of the support rod in the tray, and by the platen machine tool according to the invention, particularly a laser cutting platen machine tool or a plasma cutting platen machine tool. Further solutions are described below.

[0009] One aspect discloses a method for detecting the suspension position of a support rod in a tray. The tray is equipped with a plurality of support rods along its longitudinal direction for use with a platen machine, particularly a laser-cutting platen machine or a plasma-cutting platen machine. Each of the support rods has a plurality of support rod tips along a main extension direction oriented transverse to the longitudinal direction of the tray, and the support rod tips of the plurality of support rods define a support plane. The method includes the following steps:

[0010] - Establish a two-dimensional contrast image data set for the tray, the two-dimensional contrast image data set having multiple image pixels, wherein each image pixel is assigned a pixel value and pixel area unit of the support plane, and the contrast image data set includes a region configured as a local pixel value extremum in a background of uniform pixel values, which is assigned to the tip of the support rod.

[0011] -The longitudinal position of the region configured for the tip of one of the plurality of support rods in the contrast image data set is determined by using the local pixel value extrema, and

[0012] -The suspension position of the support rod in the tray is derived based on the longitudinal position in the contrast image data set and the size of the pixel area unit in the longitudinal direction.

[0013] In another aspect, platen cutting machines, particularly laser cutting platen cutting machines or plasma cutting platen cutting machines, include:

[0014] A tray having multiple support rods suspended in a suspension position, each support rod having multiple support rod tips that define a support plane.

[0015] A camera, the camera being used to generate a two-dimensional camera image or a two-dimensional dark-field camera image of a tray, and

[0016] Evaluation device.

[0017] Cameras and / or evaluation devices are designed for

[0018] - Establish a two-dimensional contrast image data set for the tray, the two-dimensional contrast image data set having multiple image pixels, assigning each image pixel a pixel value and pixel area unit to the support plane, and the contrast image data set including the region configured for the tip of the support rod as a local pixel value extremum in a background of uniform pixel values.

[0019] The evaluation device is also designed for:

[0020] -The longitudinal position of the region configured for the tip of one of the plurality of support rods in the contrast image data set is determined by using the local pixel value extrema, and

[0021] -The suspension position of the support rod in the tray is derived based on the longitudinal position in the contrast image data set and the size of the pixel area unit in the longitudinal direction.

[0022] In some further embodiments, one of the regions configured for the support rod tip may be configured for a support rod tip and / or configured for a side extending to the support rod tip.

[0023] In some further schemes, a uniform pixel value background may not represent a structure and / or include image pixels with constant pixel values ​​or a random distribution of pixel values ​​around constant pixel values.

[0024] In some further embodiments, the establishment of a set of contrasting image data of the tray may also include:

[0025] - Generate a two-dimensional camera image of a tray in the visible and / or infrared spectrum, the camera image having image pixels arranged in rows and columns, and assigning pixel area units of a supporting plane to each image pixel.

[0026] - Perform a Fourier transform on the camera image to the frequency domain and output the transformed frequency data set.

[0027] - Filter the transformed frequency data set to highlight frequencies in the frequency domain that belong to the region configured for the support rod tip, and

[0028] - The filtered frequency data set is inversely transformed into an inversely transformed image data set using the inverse Fourier transform, and the inversely transformed image data set is output as a comparison image data set.

[0029] In some further embodiments, adjacent support rod tips in the main extension direction may have substantially the same tip spacing. Filtering the transformed frequency data set may include: identifying a frequency band configured for the tip spacing in the frequency domain and limiting the transformed frequency data set to said frequency band. The frequency band may be formed, for example, around a tip repetition frequency configured for the main extension direction. The tip repetition frequency may be derived, in particular, from:

[0030]

[0031] in,

[0032] h: The number of pixels in the area represented by the supporting plane in the camera image along the main extension direction.

[0033] Δx: The size of the pixel area unit in the main extension direction, and

[0034] d: The distance between the tips of adjacent support rods.

[0035] In some further embodiments, the generation of two-dimensional camera images can be achieved by generating dark-field camera images of the tray in the near-infrared spectrum, and the dark-field camera images can be captured from a shooting direction that forms an angle with the supporting plane, the angle being in the range of 10° to 70°, particularly greater than 45°, for example 60°, and the shooting direction extending at an angle in the range of 70° to 110° with the main extension direction, particularly perpendicular to the main extension direction.

[0036] In some further embodiments, the establishment of the contrast image data set of the tray may also include generating the contrast image data set by producing a dark-field camera image of the tray in the near-infrared spectrum from an shooting direction at an angle to the support plane. This angle can be, for example, in the range of 10° to 70°, particularly greater than 45°, for example, 60°. Furthermore, the shooting direction can extend at an angle in the range of 70° to 110° with the main extension direction, particularly perpendicular to the main extension direction.

[0037] In some further embodiments, the establishment of a set of contrasting image data of the tray may also include: illuminating the tray from above with light in the near-infrared spectrum from one or more illumination directions. The light may, in particular, be in the spectrum of 845 nm to 855 nm and / or 935 nm to 945 nm.

[0038] In some further embodiments, at least one of the irradiation directions may form an angle with the support plane, the angle being less than 30° and particularly in the range of 10° to 20°, for example, about 15°, and at least one of the irradiation directions projecting onto the support plane may form an angle with the main extension direction of the support rod, the angle being less than 30° and particularly in the range of 10° to 25°, for example, about 15°.

[0039] In some further embodiments, the generation of a two-dimensional camera image or a two-dimensional dark-field camera image may include: capturing multiple partial camera images, each of which represents a portion of the tray in two dimensions, and combining the partial camera images into a two-dimensional camera image or a two-dimensional dark-field camera image of the tray.

[0040] In some further embodiments, determining the longitudinal position within a set of contrasting image data may include:

[0041] Template matching image analysis is performed using templates. The template can be formed as a set of image data of support rods with multiple support rod tips, particularly support rods arranged in the internal area of ​​the tray.

[0042] In some further embodiments, determining the longitudinal position in the comparison image data set may include: calculating the pixel values ​​of the image pixels in the comparison image data set row by row according to the main extension direction, and outputting a first distribution of the first sum of pixel values, and determining the local extrema in the first distribution and outputting the row of the local extrema as the longitudinal position.

[0043] In some further embodiments, the method may also include:

[0044] - Corresponding to the longitudinal direction, in the region of local extrema, the pixel values ​​of the image pixels in the contrast image data group are summed column by column, and a second distribution of the second pixel value sum is output.

[0045] - Identify multiple local extrema in the second distribution and output a column of said local extrema, and

[0046] - Determine the image position of the region configured for the tip of the support rod in the contrast image data set based on the corresponding row and column of the determined local extrema.

[0047] In some further embodiments, the support rod can be suspended at a predetermined suspension position on the longitudinal side of the tray, and the main extension direction can extend laterally to the longitudinal side. The tray can have a reference structure that can be identified as a reference pixel arrangement in the comparison image data set. The suspension position of the support rod in the tray can be derived by determining the distance between the reference pixel arrangement and the longitudinal position laterally to the main extension direction in the comparison image data set, and by identifying one of the predetermined suspension positions for the longitudinal position based on the determined distance.

[0048] In some further schemes, pixel values ​​can be grayscale values, color values, or brightness values.

[0049] In some further embodiments, the camera can be configured to:

[0050] - Generate a two-dimensional camera image of a tray in the visible or infrared spectrum, wherein the camera image has image pixels arranged in rows and columns, and each image pixel is assigned a pixel area unit of a supporting plane, or

[0051] - A set of contrasting image data is generated by capturing two-dimensional dark-field camera images of the tray in the near-infrared spectrum from the shooting direction.

[0052] In some further embodiments, the evaluation apparatus may be designed to implement the method and / or may include a processor.

[0053] The processor is configured to:

[0054] - Perform a Fourier transform on the camera image to the frequency domain and output the transformed frequency data set.

[0055] - Filter the transformed frequency data set to highlight frequencies in the frequency domain that belong to the region configured for the support rod tip, and

[0056] - The filtered frequency data set is reverse-transformed using the inverse Fourier transform, and the image data set that has been reverse-transformed is output as the comparison image data set.

[0057] In some further embodiments, at least one lighting device may be provided, designed to illuminate the tray with light in the near-infrared spectrum from one or more illumination directions. At least one of the illumination directions may form an angle with the support plane, the angle being less than 30° and particularly in the range of 10° to 20°, for example, about 15°. At least one illumination direction projected onto the support plane may form an angle with the main extension direction of the support rod, the angle being less than 30° and particularly in the range of 10° to 25°, for example, about 15°.

[0058] In some further embodiments, the method may include: determining the locations within a column or row assigned to the respective support rod tips as local extreme values ​​of the sum of first local pixel values. For this purpose, for example, parallel projection image analysis may be implemented, in which the pixel values ​​of the image pixels assigned to the first local extreme value of the sum of first local pixel values ​​(i.e., the column or row assigned to the support rod) are summed perpendicular to the main extension direction of the support rod (row-by-row or column-by-column) to obtain a second sum of pixel values, which is then assigned to the corresponding row or column. Based on this, rows or columns can be determined, and the sum of pixel values ​​assigned to said rows or columns forms the second local extreme value of the sum of pixel values.

[0059] In some further embodiments, determining the longitudinal position of the region configured for the tip of the support rod in the comparative image data set may alternatively or additionally include: performing sliding window method image analysis, SIFT-based image analysis, or neural network-based image analysis, in which a framework is established around the template matching filter. Attached Figure Description

[0060] This document discloses solutions that allow for at least partial improvements to the prior art. Further features and their effectiveness are derived, particularly from the following description of the embodiments with the aid of the accompanying drawings. In the drawings:

[0061] Figure 1 A schematic spatial view of a flatbed machine tool with a tray equipped with multiple support rods is shown.

[0062] Figure 2 A flowchart illustrating a method for detecting the suspension position of a support rod in a tray is shown.

[0063] Figure 3 A corrected camera image (top view) of an exemplary tray is shown.

[0064] Figure 4 This shows a view of the amplitude spectrum of a frequency data set transformed with respect to the frequency domain from a camera image.

[0065] Figure 5 The illustration shows side-by-side segments of exemplary contrast image data generated by filtered Fourier transform, as well as segments of the camera image to which the tray belongs.

[0066] Figure 6 A flowchart illustrating the calculation of a set of contrasting image data of a tray using Fourier transform and filtering in the frequency domain is shown.

[0067] Figure 7 An exemplary set of contrast image data generated by means of dark field illumination is shown.

[0068] Figure 8 A flowchart illustrating the process of assembling contrasting image data of a tray using dark-field illumination and dark-field imaging is shown, along with...

[0069] Figure 9 An exemplary section of a template is shown for template image analysis. Detailed Implementation

[0070] The aspects described herein are partly based on the understanding that robust, rapid, and highly accurate detection of the suspension position of a support rod can be performed via image recognition, particularly image recognition of the rod tip, if a set of contrasting image data in which the support rod tip is highlighted and other "interference" structures are largely suppressed or not shown, can be implemented. Interference structures may arise from factors such as uncertain lighting, wear, slag, and the surface condition of the support rod. Using such a low-interference set of contrasting image data reduces erroneous support rod detections. Furthermore, the complexity of image analysis can be reduced.

[0071] The inventors also realized that the contrast image dataset can be established through Fourier transform and through dark-field illumination or dark-field shooting. Furthermore, these two methods can also be combined, for example, by subsequently filtering the (camera) image established through dark-field illumination or dark-field shooting before passing it to the contrast image dataset using Fourier transform. This combination of the two methods can achieve a contrast image dataset with lower interference.

[0072] In both cases, the tray (usually unoccupied) is first photographed using a camera, and then its perspective is transformed into a "top view," so that the support rod appears in the image as a (vertical or horizontal) line extending, for example, along the main extension direction of the support rod. The photographs can also be taken from multiple parts of various cameras, and not only from outside the machine ("front") but also from inside the machine (i.e., in the production space).

[0073] In the images to be evaluated, the supporting plane is represented without perspective distortion in the set of contrasting image data. For correction, the image with original perspective distortion can be processed, for example, using the usual homography method. Therefore, each image pixel can be assigned a pixel area unit of the supporting plane, specifically, for example, a square pixel area unit of the same size for all image pixels of 1 mm². After correction in this way, objects of the same size in the supporting plane can be represented with the same size in the set of contrasting image data, regardless of the position of individual objects in the supporting plane or the set of contrasting image data. If the size of the image pixels in the extension direction of the support rod and in its lateral direction is known, correction can simplify determining the longitudinal position of the area assigned to the tip of the support rod (and thus the longitudinal position of the corresponding support rod).

[0074] The disclosed scheme is also based on the use of a set of contrasting image data of the tray, in which the area configured for the tip of the support rod is highlighted. Exemplarily, such a set of high-contrast image data can be obtained through a special imaging arrangement or image analysis method (or a combination of both). The application of filtering and dark-field imaging in Fourier image processing is illustrated in more detail here as an example. The application of filtering and dark-field imaging using Fourier image processing can be used here, for example, by first capturing an image of the tray (camera) using dark-field imaging and then filtering this image using Fourier image processing, or a combination thereof. Modifications to this or other methods will be apparent from the common description of the exemplary embodiments.

[0075] The Fourier transform is a common transformation in the field of image processing from a set of image data to a set of frequency data. The set of image data is called the first reference system, i.e., the image or spatial domain, and the set of frequency data is called the second reference system, i.e., the frequency domain.

[0076] In the Fourier transform, the periodic patterns appearing regularly in the spatial domain, i.e., the geometric shape of the tray, are assigned frequencies in the frequency domain of the transformed image data set. Patterns appearing periodically in the spatial domain along the direction of pixel arrangement are represented as horizontal or vertical (frequency) lines in the amplitude spectrum of the transformed frequency data set, and are therefore easily detected. The tips of the support rods also form repeating patterns in the image data set along the main extension direction of the support rods, and can therefore be assigned frequencies in the transformed frequency data set. These frequencies are also represented as lines in the amplitude spectrum (assuming the corresponding orientation of the support rods).

[0077] Filtering the transformed frequency data set allows the frequencies assigned to the support rod tip to be highlighted, while other frequencies are (relatively) reduced or even removed. A subsequent inverse transformation of the filtered transformed frequency data set allows for the output of an image data set (within the image domain) in which the support rod tip is highlighted with high contrast. In contrast, “disturbing” structures or textures not explicitly assigned to the support rod are at least reduced, such that, in other words, the area of ​​the support rod tip can be represented as a local pixel value extremum within a background of uniform pixel values ​​in the contrasting image data set of the tray.

[0078] In this context, a local pixel extreme can be a group of adjacent image pixels or a single image pixel. If the local pixel extreme forms a group of adjacent image pixels, then the image pixels belonging to the group therefore each have a much higher pixel value than those image pixels surrounding the group of image pixels that form a background of uniform pixel values. Pixel values ​​can be, for example, grayscale values, with low grayscale values ​​corresponding to dark image pixels and high grayscale values ​​corresponding to bright (illuminated) image pixels.

[0079] The contrast image dataset of the tray can also be obtained using dark-field illumination or dark-field imaging (dark-field strategy). Dark-field strategy is used, for example, in automated visual inspection. Dark-field strategy uses a dark-field illumination device and a dark-field camera to capture dark-field images. Illumination is performed at (typically very small) angles, while imaging can be performed at different (steep) angles. Most of the light emitted by the dark-field illumination device is not reflected to the dark-field camera. In the aforementioned visual inspection, only defects in the surface scatter / reflect light, whereas in the current case of imaging a tray with support rods, only the tips of the support rods and the edges extending to the tips scatter / reflect light to the dark-field camera. The result is a mostly dark dark-field image (referred to herein as a uniform pixel value background), with only the tips and edges of the support rods appearing much brighter than their surroundings (referred to herein as local pixel value extrema). Such a dark-field image is another example of a contrast image dataset of the tray, where the area configured for the tips of the support rods is represented as a local pixel value extrema in a uniform pixel value background.

[0080] Using near-infrared light for (dark-field) illumination also allows for reliable suppression of external light effects caused by sunlight or artificial lighting without the need for a housing. In this case, external light effects caused by sunlight can be suppressed, in particular, by using near-infrared light in the frequency range of 845 nm to 855 nm and / or 935 nm to 945 nm, as this corresponds to the band gap in the solar spectrum at the Earth's surface. In the near-infrared spectrum, and especially in the aforementioned spectrum, the acquisition of dark-field camera images of the tray can be achieved, in particular, by using appropriate spectral filters.

[0081] The determination of the suspension position is based on the analysis of a set of contrasting image data. Various analytical methods are presented here with examples.

[0082] Template matching image analysis is a two-stage approach for detecting objects in a dataset of images. First, a template of the object to be identified is constructed. Then, the presence of the template in the image dataset is evaluated using a similarity metric, such as the sum of absolute differences. Template matching image analysis is described in detail in G. Cheng and J. Han, “A survey on object detection in optical remote sensing images,” ISPRS Journal of Photogrammetry and Remote Sensing, Vol. 117, pp. 11–28, 2016.

[0083] Templates that can be formed as image data sets for support rods can be generated in this case, particularly by capturing camera images or dark-field camera images in the near-infrared spectrum of a new (unworn or unaffected by slag) support rod.

[0084] Since the representation of the support rod may change depending on its suspension position along the longitudinal range of the pallet due to perspective effects, support rods already suspended within the inner area of ​​the pallet's longitudinal range can be selected as templates. Such templates can be used to detect support rods suspended more "forward" or more "backward".

[0085] Parallel projection image analysis is also a method for detecting objects that extend along specific portions of rows or columns of an image data set. Specifically, in parallel projection image analysis, support bars (whose main extension direction extends along the image pixel lines or rows) can be easily and quickly identified by adding the grayscale values ​​of image pixels row by row (or column by column). The support bars form a maximum value in this "signal distribution" generated by the projection / summation of pixel values ​​along the support bars. Parallel projection image analysis is described, for example, in P. Du et al., "Improved peak detection in mass spectrum by incorporating continuous wavelet transform-based pattern matching," Bioinformatics, Vol. 22, No. 17, pp. 2059–2065, 2006.

[0086] Other methods for image analysis of contrastive image datasets include: sliding window method image analysis (see, for example, M. Barva et al., “Parallelintegral projection transform for straight electrode localization in 3-D ultrasound images,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 55, No. 7, pp. 1559–1569, 2008); SIFT-based image analysis (see, for example, D. GLowe, “Object recognition from local scale-invariant features,” IEEE International Conference on Computer Vision, Vol. 99, No. 2, 1999, pp. 1150–1157); and neural network-based image analysis (see, for example, D. Erhan et al., “Scalable object detection using deep neural networks,” IEEE Conference on Computer Vision and Pattern Recognition, 2014, pp. 2147–2154, and K. He et al., “Maskr-CNN,” Proceedings of the IEEE International Conference on…). ComputerVision, 2017, pp. 2961–2969.

[0087] Typically, misidentification might be caused by, for example, a highly reflective edge on a support rod.

[0088] Or it may be caused by cutting burrs. To subsequently eliminate such false identifications, the detected support rod candidates can be examined. For example, it can be checked how much the grayscale value of the image portion assigned to the support rod candidate fluctuates in the direction of the support rod tip. If the fluctuation is higher than a threshold to be fixed, then a falsely identified support rod candidate can be assumed and eliminated. In this way, false identifications can be advantageously avoided quickly and easily when using parallel projection image analysis.

[0089] The following is for reference. Figure 1 Describe the installation of the flatbed machine tool. Figure 2 The flowchart summarizes the steps of the method for detecting the suspension position. Figures 3 to 6This paper describes how to generate contrast image datasets using filtered Fourier transforms. Figure 7 and Figure 8 This paper describes how to generate contrast image datasets using a dark field method. Figure 9 A template for image analysis is shown.

[0090] Figure 1 A flatbed machine tool 1 is shown, which has a tray 3, a camera 5, lighting devices 7 and 7' (shown by dashed lines), and an evaluation device 11, which includes a processor 9 and is connected to the camera 5.

[0091] The support rod 13 can be suspended in the frame 4 of the tray 3 along the longitudinal side 3a of the tray 3 (the longitudinal side defining the longitudinal range) at multiple discrete suspension positions 15, including any type of fixed arrangement and optional fixation. At some suspension positions 15, in Figure 1 In the example of the pallet, the support rod 13 is not suspended. It can be seen that the support rod 13 is mounted more tightly to the machine tool than to the far side 3b of the pallet.

[0092] Support rod 13 extends along a common main extension direction 13a (in Figure 1 Extending from the end side 3b) of the central tray 3, each support rod tip 17 is configured in the form of a plate and has a plurality of support rod tips 17. The support rod tips 17 form / define a common support plane 17a. In the main extension direction 13a, successive / adjacent support rod tips 17 are arranged apart from each other with a tip spacing 17b. Generally, the tip spacing 17b is substantially the same for all support rods 13.

[0093] A material panel (not shown) for processing by the flatbed machine tool 1 can be mounted on a support rod 13 in a support plane 17a. In this case, the tip 17 of the support rod contacts the material panel, such that, for example, during material processing, the laser beam can also interact with the tip 17 of the support rod. In order to take into account the position of the support rod 13, for example, in the arrangement of workpieces to be cut in the material panel, it is preferable to automatically detect the position of the support rod 13.

[0094] To ensure the correct positioning of the material panel, tray 3 has a stop 19, which is arranged on frame 4 and... Figure 1 The diagram is schematic. The design of the stop 19 allows it to be used as an optically detectable reference structure 19a in image processing.

[0095] Camera 5 is connected, for example, to the production housing of the flatbed machine tool 1 via camera arm 5'. With the aid of camera 5, two-dimensional, particularly monochrome or color, camera images of the tray 3 can be generated in the visible or infrared spectrum (see...). Figure 3Image data set 21), or dark-field camera images of tray 3, can be generated in the near-infrared spectrum (see image data set 21). Figure 7 (Comparison image data set 23'). The corresponding image is captured from the shooting direction 5a, from the side upwards toward the tray 3. Due to the tilted shooting direction 5a, optical or digital correction of the camera image can be performed to facilitate image processing. Figure 3 and Figure 7 The image shown is a correspondingly corrected camera image. As already mentioned, this “corrected” camera image of the support plane 17a can be provided, for example, by a homography method.

[0096] In this case, reference can also be made to the image data group 21, the contrast image data group (23), and the dark field camera image or contrast image data group (23') for the corresponding image-based or image data group direction indication, using the (common) main extension direction 13a of the imaging support rod 13 and the longitudinal direction of the imaging tray 3 extending laterally to the main extension direction.

[0097] Camera images and dark-field camera images have image pixels arranged in rows and columns, with pixel values ​​on a supporting plane and pixel area units of the same size, typically square, assigned to the image pixels, respectively. Pixel values ​​are grayscale values, with low grayscale values ​​corresponding to dark pixels and high grayscale values ​​corresponding to bright pixels. For the purposes of this discussion, color resolution is not necessary because image analysis is primarily related to intensity values.

[0098] The shooting direction 5a and the supporting plane 17a form an angle 5b. Figure 1 In the example given, the shooting direction 5a extends perpendicularly to the main extension direction 13a of the support rod 13. The angle 5b is, for example, in the range of 10° to 70°, and in particular, this angle can be greater than 45°, for example, 60°. The deviation of the vertical orientation from the main extension direction is, for example, approximately ±20°.

[0099] With the aid of the illumination device 7 (shown by dashed lines), the tray 3 can be additionally illuminated in the illumination direction 7a by means of infrared light, for example by means of light in the near-infrared spectrum of 845nm to 855nm and / or 935nm to 945nm. The illumination direction 7a forms an angle 7b with the support plane 17a. The angle 7b is, for example, less than 30° and particularly in the range of 10° to 20°, for example, about 15°. The illumination device 7 illuminates the support rod 13 from the "rear," that is, relative to the camera 5, on the rear side of the support rod 13. Accordingly, the illumination direction 7a (perpendicular to the support plane 17a) projected into the support plane 17a forms an angle 7c with the main extension direction 13a of the support rod 13. The angle 7c is, for example, less than 30° and particularly in the range of 10° to 25°, for example, about 15°.

[0100] Other illumination directions can be provided by other lighting devices; for example Figure 1 Additional lighting device 7' is shown. Multiple lighting devices allow the support rod to be illuminated more evenly. It must be considered here that the lighting devices should not shine directly on the camera to avoid camera oversaturation (“camera glare”). In addition, the camera 5 can be used in conjunction with a frequency filter that transmits a specifically selected spectral range.

[0101] The flatbed machine tool 1 is configured to enable a method for detecting the suspension position 15 of the support rod 13. The computations required for this image processing can be implemented, for example, using the processor 9 of the evaluation device 11. Accordingly, the camera 5 and optional lighting device 7 are connected to the processor 9 / evaluation device 11 to receive data (image data sets) and to activate them (trigger image capture or lighting). The evaluation device 11 can be configured as a PC, computing node, or similar suitable hardware, and particularly as part of a control unit that implements the activation of the flatbed machine tool 1. Therefore, the evaluation device 11 can be configured as part of an advanced or local control system of the flatbed machine tool, or as a unit within the flatbed machine tool itself. The evaluation device 11 is specifically designed to perform / maintain checks on the method for determining the suspension position during real-time operation of the flatbed machine tool 1. For this purpose, the underlying computing system has, for example, microprocessor circuitry with a digital processor system with data inputs and control outputs, and a database that operates according to computer-readable instructions stored on a computer-readable medium. These instructions include, for example, computer routines for Fourier transforms, filtering, and image analysis. The evaluation apparatus 11 can provide high computing power for real-time support. The evaluation apparatus can also provide long-term (non-volatile) memory for storing program instructions and very fast short-term (volatile) memory for (buffering) storing acquired data and data sets generated during the implementation of the method.

[0102] The method for detecting the suspension position 15 of the support rods involves a tray 3, which, as described above, is equipped with a plurality of support rods 13 along its longitudinal direction. Each of the support rods 13 has a plurality of support rod tips 17 along a main extension direction 13a, which is oriented transversely to the longitudinal direction. The support rod tips 17 define a support plane 17a. Figure 2 As shown, the method includes the following steps, which can be implemented at least in part as a computer-implemented method.

[0103] In step 100, a two-dimensional contrast image data set of tray 3 is established, the two-dimensional contrast image data set having multiple image pixels. In this case, each image pixel is assigned a pixel value and pixel area unit of the support plane 17a, and the contrast image data set includes a region assigned to the tip 17 of the support rod as a local pixel value extremum in a uniform pixel value background.

[0104] In step 200, the longitudinal position of the region of the support rod tip 17 of one of the plurality of support rods 13 in the comparison image data group is determined based on the local pixel value extreme value.

[0105] In step 300, based on the longitudinal position and the size of the pixel area unit in the longitudinal direction in the comparison image data group, the suspension position 15 of the support rod 13 in the tray 3 is derived.

[0106] The following describes two methods for creating contrastive image datasets by way of example (see step 100).

[0107] First, combined Figures 3 to 6 This section explains how to use Fourier transform to establish a set of contrasting image datasets 23. Next, based on... Figure 7 and Figure 8 The method described here utilizes dark-field illumination and dark-field shooting. The advantage of Fourier transform is its flexibility in application and integration into existing shooting systems, as readily available image data can often be used. In contrast to this computationally intensive implementation, the dark-field method essentially requires a separate camera system, which optionally has its own illumination. On the other hand, data processing is significantly simplified in the case of the dark-field method. The common objective of the methods presented here is to provide a set of contrasting image data in which the support rod tip 17 is highlighted as much as possible, in the example represented as the “brightest possible area” in the darkest possible image under other circumstances (or in a random distribution of pixel values ​​around a constant pixel value).

[0108] When the pallet is fixed in the machine tool or on a pallet changer, both methods start with an image of the pallet. The pallet is also preferably as empty as possible. The initial image is transformed by perspective so that the support rod appears as a line in one of the directions in which pixels are arranged in rows (vertically or horizontally in the image). The initial image can also be a combination of multiple camera images.

[0109] Figure 3 A two-dimensional camera image 21 of the tray in the visible spectrum, generated by camera 5, is shown, wherein a plurality of support rods are present along a main extension direction 13a. Figure 1The middle corresponds to the y-direction. To some extent, the area of ​​the pallet frame can also be seen, with the longitudinal side 3a of the pallet frame along the longitudinal direction (according to...). Figure 1 (Indicates the x-direction) extends and the end side 3b of the pallet frame extends along the main extension direction 13a.

[0110] Camera image 21 has multiple image pixels, each configured with a grayscale value. Camera image 21 is taken from an upward angle, such that as the distance between the support rod and camera 5 increases, not only the upper edge but also increasingly the camera-facing sides of the support rod are visible in camera image 21. In other words, for perspective reasons, the individual support rods 13 are represented differently in camera image 21, depending on their position in the support plane 17a, particularly their suspension position in the longitudinal direction: due to the relatively steep shooting angle, the support rods in the left edge region of camera image 21 are presented almost in a top view and therefore appear as lines (only the upper side is perceived). Due to the relatively shallow shooting angle, the support rods in the right edge region of camera image 21 are shown in an oblique view, thus perceiving their plate-like sides.

[0111] However, since the camera image 21 is corrected, for example by means of the homography method described above, each image pixel is configured with a square pixel area unit of the same size supporting plane 17a, so that the camera image 21 corresponds in this respect to a distortion-free top view (bird's-eye view), even if the image content deviates from the top view due to the tilted shooting direction.

[0112] also, Figure 1 The reference structure 19a shown is clearly visible as a reference pixel arrangement 19b in the camera image 21. The reference pixel arrangement 19b can be identified using image analysis methods (e.g., template matching image analysis, see below).

[0113] Due to lighting conditions, the supporting plane 17a may be unevenly illuminated, which manifests as some areas of the supporting rod 13 appearing much brighter than others. Typically, a highly structured grayscale value distribution can be seen in the camera image 21.

[0114] Furthermore, each support rod 13 may be in different states in terms of wear, slag, and surface condition. For example, the tips of each support rod may be melted and / or deformed.

[0115] During image processing, processor 9 will Figure 3 The camera image 21 in the spatial domain is Fourier transformed into frequency data group 21a in the frequency domain.

[0116] Figure 4The transformed frequency data set 21a is shown in a so-called amplitude-frequency representation, where the brightness of a single image point represents a measure of the amplitude (square of the absolute value) of the corresponding frequency. The frequency axes (fx-axis and fy-axis) represent the spatial frequencies in two orthogonal analysis directions (x-direction, y-direction). Figure 4 In the diagram, the two frequency axes fx and fy have their origins (coordinate origins) at the center.

[0117] Due to the properties of the Fourier transform and the regular periodic arrangement of the support rod tip 17 and the support rod 13 relative to each other, a significantly brighter region with local brightness extrema (peaks) can be seen in the square of the absolute value of the amplitude. For example, in Figure 4 In the diagram, the horizontal line is marked Lx in the fx direction, and the vertical line is marked Ly in the fy direction. The brighter areas are attributed to the geometry of the support rod 13, particularly the repeating structure of the support rod tip 17 in the main extension direction 13a and / or the regularity of the suspension position 15 in the longitudinal direction.

[0118] Taking into account the possible geometry of the support rod for the installation of the tray, binary filtering of the frequency conversion data group 21a can be implemented, for example.

[0119] The support rod tips 17, spaced apart by a tip pitch 17b, form a periodic, regularly repeating pattern in the camera image 21 (in the spatial domain) corresponding to the main extension direction 13a. This allows the tip repetition frequency (spatial frequency in the y-direction) to be configured in the frequency conversion data set 21a. In this case, the tip repetition frequency is... Figure 4 The center is represented by two horizontal lines Ly1 that are closest to the origin and equidistant from it. The repetition frequency of the tip is derived from the following:

[0120]

[0121] in,

[0122] h: The number of pixels in the region of the supporting plane 17a shown in the camera image 21 along the main extension direction 13a.

[0123] Δx: The size of the pixel area unit in the main extension direction 13a, and

[0124] d: The tip distance 17b between adjacent support rod tips 17.

[0125] An example value is: Δx = 1 mm, for example h = 1600 image pixels and d = 16 mm, thus obtaining a tip repetition frequency of 100 image pixels.

[0126] To highlight the tip repetition frequency, the transformed frequency data group 21a can be filtered, i.e., the transformed frequency data group 21a is limited to a frequency band. The filtered transformed frequency data group is, for example, extended within a fixed frequency bandwidth around the tip repetition frequency. Figure 4 In this example, two frequency bands along Ly1 were selected.

[0127] In this regard, the inventors also realized that if the high-frequency, fx spatial frequency is removed from the transformed frequency data group 21a, the amplitude contribution of the spatial frequency can be ignored (in Figure 4 The frequencies shown as dashed lines S1 and S2 on the left and right sides can be quickly and reliably used to determine the downstream position of the region configured for the support rod tip in the contrast image data set. Figure 2 Step 200 in the middle.

[0128] Transformed frequency data set 21a is filtered and limited in this way, for example, by setting the amplitude contribution of the remaining frequencies to zero and including only the amplitude values ​​in, for example, rectangle B. Similar to... Figure 4 In the representation, the area outside rectangle B is represented in black.

[0129] Then, the processor 9 uses inverse Fourier transform to inversely transform the filtered transform frequency data group 21a into the spatial domain.

[0130] In fact, the transform frequency data group 21a has been limited to a frequency band around the tip repetition frequency and therefore all frequencies outside the band are suppressed (filtered), which means that the contrast image data group 23 in the image domain can be generated by the subsequent inverse transform of the filtered transform frequency data group 21a (see...). Figure 5 In the contrast image data set 23, the support rod tip 17 is highlighted and other "disturbing" structures or textures are at least significantly suppressed. The contrast image data set 23 accordingly represents the region of the support rod tip 17 as a local pixel value extremum 23a within a uniform pixel value background 23b. Thus, an inversely transformed image data set is obtained, which can be output as the contrast image data set 23 for further image analysis.

[0131] Figure 5 By showing side by side from according to Figure 3 The “interference reduction” effect is illustrated by a segment of the two-dimensional camera image 21 of the tray (right side) and a segment of the contrast image data group 23 generated from the two-dimensional camera image by means of Fourier transform (left side). Multiple well-structured linear rows of bright regions can be seen, each corresponding to a tip of the support rod. Notably, in the illustrated segments, these rows extend equidistantly from each other (i.e., each suspension position is utilized).

[0132] Figure 6 Step 100 of establishing the contrast image data set 23 using Fourier transform is shown. The image processing employed as the basis includes the following steps, which can be at least partially implemented as a computer-based method:

[0133] In step 110', a two-dimensional camera image 21 is captured on the tray 3. Part of the capturing process is correction, which is performed in such a way that the image pixels of the camera image 21, arranged in rows and columns, are respectively configured with the same pixel area units on the support plane 17a. The correction can be calculated, for example, by means of the processor 9 of the evaluation device 11 based on the image data of the initial image, or at least partially optically.

[0134] In step 120', the camera image 21 is Fourier transformed to the frequency domain and output as transformed frequency data group 21a.

[0135] In step 130', the transformed frequency data group 21a is filtered to highlight frequencies in the frequency domain that belong to the region assigned to the support rod tip 17.

[0136] In step 140', the filtered transform frequency data group 21a is subjected to inverse Fourier transform by means of inverse Fourier transform, so that the inversely transformed image data can be output as the comparison image data group 23.

[0137] In the second method described below, the contrast image dataset is established using dark-field shooting, optionally with dark-field illumination. For this purpose, as... Figure 1 As shown, tray 3 can be illuminated by one or more lighting devices 7, 7' from the illumination direction 7a with light in the near-infrared spectrum.

[0138] The implementation of the dark field method (dark field strategy) for establishing the contrast image data set for tray 3 can take into account that the space above tray 3 should be available for loading and unloading processes. For example, Figure 1 As shown, camera 5 can be attached to the rear wall of flatbed machine tool 1 via camera arm 5'.

[0139] The lighting device 7 can then illuminate the support rod tip 17 from the side and above, so as to deflect the reflected and scattered light upwards to the camera 5. Furthermore, the lighting device 7 should not be pointed at the tray from above at too steep an angle, because other objects located below the tray 3 (e.g., material scraps) may be illuminated and become part of the image. Additionally, the lighting device 7 should not illuminate the side of the support rod 13 facing the camera 5. The lighting device 7 can preferably illuminate laterally at a small angle from the top and rear (i.e., from the side opposite the camera 5). The lighting device 7 can be positioned, for example, at a distance of 40 cm from the tray. For the lighting device 7, dark field illumination is incident at a small angle onto the support rod tip 13, which is farther from the lighting device 7.

[0140] It should also be noted that if light can enter the camera 5 directly from the lighting device 7, strong lens flare ("camera glare") can sometimes occur. This may be the case if the lighting device 7 is located on the side of the tray 3 opposite to the camera 5. To avoid this, the lighting device 7 can be provided with a corresponding cover. As mentioned above, multiple lighting devices 7, 7' can be provided to evenly illuminate the entire support plane of the tray 3.

[0141] Figure 7 The result of dark-field imaging is illustrated exemplarily in the form of a dark-field camera image on tray 3, which is captured in the near-infrared spectrum from shooting direction 5a by means of camera 5. The dark-field camera image can be directly (after correction) used as a contrast image data set 23' for further image analysis. Alternatively, the dark-field camera image can be filtered using a Fourier transform according to the method further described above, and only the filtered and inversely transformed image data set can be used as the contrast image data set, as the basis for steps 200 and 300.

[0142] In the example shown, camera 5 is equipped with a narrowband filter that has a maximum transmittance of over 90% in the near-infrared spectrum. Figure 7 As can be seen, (basically) only the tip of the support rod 17 and some sides extending to the tip of the support rod 17 can be seen as bright areas in the contrast image data set 23', while the rest of the image is as black as possible.

[0143] Figure 8 The flowchart illustrates step 100 of establishing a set of contrasting image data using images from a dark-field camera. The image capture employed as a basis includes the following steps, which can be implemented at least in part as a computer-implemented method:

[0144] In step 105 (optional), the tray 3 is illuminated from above by light in the near-infrared spectrum, particularly in the spectrum of 845 nm to 855 nm and / or 935 nm to 945 nm, from one or more illumination directions 7a. At least one of the illumination directions 7a may be at the angles 7a, 7b indicated above with respect to the support plane.

[0145] In step 110", the comparison image data set 23' is generated by producing a dark-field camera image of the tray 3 in the near-infrared spectrum from the shooting direction 5a. Dark-field shooting can be performed at angles 5a, 5b indicated above with respect to the support plane and the main extension direction.

[0146] In other words, in a preferred embodiment, the flatbed machine tool can be equipped with a near-infrared (NIR) illuminator for the pallet used for processing, and also with a camera having a narrowband filter. Using a narrowband filter prevents the influence of ambient light. If the camera is mounted on the housing of the flatbed machine tool, and the illuminator is installed along the longitudinal direction of the pallet such that the support rod is illuminated at an angle from behind, this produces a dark-field image in which the tip and sides of the support rod are represented as bright, while the rest of the image is dark.

[0147] If a set of contrasting image datasets 23, 23' is established, the location of the region configured for the support rod tip 17 within the set of contrasting image datasets 23, 23' is determined by means of image analysis. This can be done, for example, by means of parallel projection of the image analysis of the set of contrasting image datasets 23, 23' or by template matching image analysis. Other more sophisticated image analysis methods can also be used, such as sliding window method image analysis, SIFT-based image analysis, or neural network-based image analysis.

[0148] The results of the parallel projection along the main extension direction 13a are as follows Figure 7 The diagram schematically indicates the lower edge of the image. The intensity distribution 31 is obtained in the x-direction, i.e., along the longitudinal direction of tray 3, by summing the pixel values ​​in the y-direction (the grayscale values ​​of the image are added column by column in the main extension direction 13a). Multiple maximum values ​​31a can be seen in the intensity distribution.

[0149] Strongly reflective edges or cut burrs on the support rod may lead to false identification. False identification can be detected, for example, by examining how much the grayscale value of the candidate (the corresponding area of ​​intensity distribution 31) fluctuates along the main extension direction 13a. For example, the grayscale values ​​in the x-axis region surrounding the candidate (e.g., in the region surrounding a maximum value + / - 20 image pixels) can be summed (a parallel projection transverse to the main extension direction). Alternatively or additionally, to quickly classify a candidate as a false identification, the average value fluctuation along the y-axis can be calculated. If this average value is below a previously fixed threshold, it is a false identification.

[0150] exist Figure 7 The diagram also illustrates that the support rods can be suspended at equal intervals. If one of the suspension positions 15 coincides with the maximum strength in the strength distribution, the corresponding suspension position is marked with "1" (the suspension position is equipped). If there is no maximum strength at suspension position 15, the suspension position is marked with "0" (no support rod 13 is suspended). Figure 7 As can be seen from the situation, not every suspension position 15 is used, that is, not every suspension position 15 is equipped with a support rod 13.

[0151] A similar intensity distribution of the parallel projection transverse to the main extension direction 13a, confined to a narrower region (e.g., + / - 4 image pixels) around the support tip, can also be used to identify individual support tips in the y-direction. For this purpose, the maximum value of the signal (the sum of grayscale values) can be determined again.

[0152] This is also in Figure 7 The center indicates the position. For support rod 13, a parallel projection in the x-direction is performed to determine the accurate position of the support rod tip 17. Figure 7 The corresponding strength distribution 35 in the y-direction has been shown using several maximum values ​​35a. The x-position of the suspension position (maximum value 31a) and the y-position of one of the maximum values ​​35a in the strength distribution 35 are used to obtain the two-dimensional position of the corresponding support rod tip 17. Figure 7 The x and y positions in the diagram indicate the support rod tip 37 by way of example. The position of the support rod tip 17 can, for example, serve as the basis for the cutting process by taking into account the planned arrangement of the workpiece to be cut.

[0153] The parallel projection method associated with dark field camera images can be similarly applied to image data set 23.

[0154] Figure 9Template 25 for template matching image analysis is shown. Template 25 shows a region 27 with slightly elevated pixel values, which in the template belongs to the support tip and optionally to the side extending to the support rod tip. Typically, the template can be extracted from multiple images. Template 25 can be obtained, for example, from a set of image data of support rods 13 having multiple support rod tips 17. For example, the template can be obtained from the internal region of the tray 3, since the support rods 13 arranged there can be considered as the average of all support rods 13. Template 25 can also be constructed by means of forward and inverse Fourier transforms, optionally through corresponding filtering.

[0155] For example Figure 9 As shown, for example, a template can be used asymmetrically with respect to region 27 to account for existing structural information and prevent "rod indentations" from being incorrectly detected as "rod tips." "Rod tips," in particular, are those where the support rod is viewed laterally in the initial image, i.e., they do not form a line. An asymmetrical template can then be selected such that a region without a Fourier bandpass filter response is required next to the detected signal. In the case of a rod indentation, the tip will be visible there, ensuring that only the support rod tip produces a height overlap with template 25.

[0156] In template matching image analysis, determining the position of the region configured for the support rod tip 17 in the comparison image data sets 23, 23' involves performing a matching operation using template 25 with respect to portions of the comparison image data sets 23, 23' of comparable size. The x-position of the template that results in a high overlap with the comparison image data sets 23, 23' is primarily obtained as a result of position determination using template 25. This position then corresponds to the suspension position 15 of the support rod 13. Based on the quality of the template and the contrast between template 25 and the comparison image data sets 23, 23', the corresponding position of the support rod tip 17 along the main extension direction 13a can also be detected.

[0157] Similarly, in template matching image analysis, the actual positions of possible hanging positions 15 in tray 3 configured for comparison image data groups 23 and 23' can also be included for determining the hanging position 15, as already combined with... Figure 7 The explanation regarding the suspension position 15.

[0158] In other words, if the position of the region assigned to the tip of the support rod 17 in the comparison image data set 23, 23' is determined based on the local pixel value extreme value 23a (step 200), then the corresponding suspension position 15 of the support rod 13 in the tray 3 can be derived based on the position of the region assigned to the tip of the support rod 17 in the comparison image data set 23, 23' and the pixel area unit (step 300).

[0159] This can be achieved, for example, by first determining the reference pixel arrangement 19b in the contrast image data set 23 (see...). Figure 3 The number of image pixels between the x-position of the region of the support rod tip 17, which is arranged transversely to the main extension direction 13a, is used to implement the support rod 13. Then, by means of the known range (e.g., 1 mm x 1 mm) of the pixel area units arranged to the image pixels, the corresponding spacing between the reference structure 19a arranged to the reference pixel arrangement 19b and the associated support rod 13 is determined according to the determined number of image pixels. Based on the obtained spacing (between the reference structure 19a and the associated support rod 13) and the known spacing between the reference structure 19a and the (discrete) possible suspension positions 15, the associated support rod 13 can be configured with a suspension position 15, and thus the suspension position 15 of the support rod 13 in the tray 3 can be determined.

[0160] Comparing the two methods used to generate the set of contrasting image data, presented as an example, the dark-field method has the advantage of not relying on favorable ambient light. Another advantage is that the ground beneath tray 3 and the front side of support rod 13 are not illuminated. This results in fewer distractions in the images. Accordingly, dark-field shooting makes the assessment of whether the support rod has been placed in the suspension position 15 more robust. Almost no misclassification occurs. Dark-field shooting can also assess the position of the individual support rod tips 17 on each image with an acceptable error rate. Furthermore, implementing the methods using both the dark-field and Fourier methods is rapid because the dark-field / Fourier analysis can be performed within fractions of a second.

[0161] In the case of Fourier image analysis, the current support rod occupancy of the pallet can be extracted directly from the initial image without any additional aids, such as specific dark-field illumination. Also in this case, unlike the scanning method, there is almost no time loss because, as mentioned earlier, the pallet is empty shortly before the new metal sheet is placed. All that is required in this implementation of Fourier image analysis is a suitable data processing environment for calculating the described transform. The calculation can be performed on the machine tool's controller or outsourced to an external server. Accuracy of 96% or higher has been achieved.

[0162] It is explicitly emphasized that all features disclosed in the specification and / or claims should be considered separate and independent of each other for the purposes of the original disclosure, and similarly for the purposes of limiting the claimed invention independently of the combinations of features in the embodiments and / or claims. It is explicitly stated that all scope indications or unit group indications disclose any possible intermediate values ​​or unit subgroups for the purposes of the original disclosure, and similarly for the purposes of limiting the claimed invention, particularly as limitations of scope parameters.

Claims

1. A method for detecting the suspension position (15) of a support rod (13) in a tray (3), the tray being equipped with a plurality of support rods (13) along a longitudinal direction (x) for use with a flatbed machine tool (1), each of the support rods (13) having a plurality of support rod tips (17) along a main extension direction (13a) oriented transverse to the longitudinal direction of the tray (3), and the support rod tips (17) of the plurality of support rods (13) defining a support plane (17a), the method comprising the steps of: - Establish (100) a two-dimensional contrast image data set (23, 23') of the tray (3), the two-dimensional contrast image data set having multiple image pixels, wherein, Each image pixel is configured with a pixel value and pixel area unit of the support plane (17a), and the contrast image data group (23, 23') includes the region configured for the tip of the support rod (17) as a local pixel value extremum (23a) in a uniform pixel value background (23b). - The longitudinal position of the region of the tip (17) of one of the plurality of support rods (13) configured in the contrast image data set (23, 23') is determined (200) by using the local pixel value extreme (23a). - The suspension position (15) of the support rod (13) in the tray (3) is derived (300) based on the longitudinal position in the comparison image data group (23) and the size of the pixel area unit in the longitudinal direction (x).

2. The method according to claim 1, wherein, One of the regions configured for the support rod tip (17) is configured for a support rod tip and / or is configured for a side extending to the support rod tip (17).

3. The method according to any one of the preceding claims, wherein, The uniform pixel value background (23b) does not represent a structure and / or includes image pixels with constant pixel values ​​or random pixel value distributions around constant pixel values.

4. The method according to claim 1 or 2, wherein, The establishment (100) of the contrast image data set (23, 23') of the tray (3) also includes: - Generate (110') a two-dimensional camera image (21) of the tray (3) in the visible and / or infrared spectrum, the camera image (21) having image pixels arranged in rows and columns, and assigning pixel area units of the support plane (17a) to each image pixel. - The camera image (21) is Fourier transformed (120') to the frequency domain and the transformed frequency data set (21a) is output. - The transformed frequency data set (21a) is filtered (130') to highlight the frequencies in the frequency domain that belong to the region configured for the tip (17) of the support rod, and - The filtered frequency data set is inversely transformed (140') into an inversely transformed image data set by means of inverse Fourier transform, and the inversely transformed image data set is output as the comparison image data set (23).

5. The method according to claim 4, wherein, Adjacent support rod tips (17) in the main extension direction (13a) have the same tip spacing (17b) to each other, wherein filtering (130') of the transformed frequency data group (21a) includes: - Identify in the frequency domain a frequency band configured for the tip spacing (17b), the frequency band being formed around a tip repetition frequency configured for the main extension direction and derived from the following: ， in, h: The number of pixels in the area represented by the supporting plane (17a) in the camera image (21) on the main extension direction (13a). Δx: The size of the pixel area unit in the main extension direction (13a), and d: The tip spacing (17b) between adjacent support rod tips (17), and - Limit the transformed frequency data group (21a) to the frequency band.

6. The method according to claim 4, wherein, The generation (110'') of the two-dimensional camera image (21) is performed by generating a dark-field camera image of the tray (3) in the near-infrared spectrum, wherein the dark-field camera image is captured from a shooting direction (5a) that forms an angle (5b) with the supporting plane (17a) in the range of 10° to 70°, and the shooting direction extends at an angle in the range of 70° to 110° with the main extension direction.

7. The method according to claim 1 or 2, wherein, The establishment (100) of the contrast image data set (23) of the tray (3) also includes: - The contrast image data set (23) is generated (110'') by producing a dark field camera image of the tray (3) in the near-infrared spectrum from the shooting direction (5a), the shooting direction forming an angle (5b) with the supporting plane (17a) in the range of 10° to 70°, and the shooting direction extending at an angle in the range of 70° to 110° with the main extension direction.

8. The method according to claim 1 or 2, wherein, The establishment (100) of the contrast image data set (23) of the tray (3) also includes: - The tray (3) is illuminated from above (105'') by light of the near-infrared spectrum from one or more irradiation directions (7a).

9. The method according to claim 8, wherein, At least one of the irradiation directions (7a) forms an angle (7b) with the support plane (17a) less than 30°, and at least one of the irradiation directions (7a) projected onto the support plane (17a) forms an angle (7c) with the main extension direction (13a) of the support rod (13) less than 30°.

10. The method according to claim 4, wherein, The generation (110') of the two-dimensional camera image (21) or the generation (110'') of the two-dimensional dark field camera image includes: - Capture multiple partial camera images, each of which represents a portion of the tray (3) in two dimensions. - Combine the partial camera images into a two-dimensional camera image (21) or a two-dimensional dark field camera image of the tray (3).

11. The method according to claim 1 or 2, wherein, The determination (200) of the longitudinal position in the comparison image data set (23, 23') includes: - Template matching image analysis is performed by using a template (25), which is formed as a set of image data of a support rod (13) with multiple support rod tips (17).

12. The method according to claim 1 or 2, wherein, The determination (200) of the longitudinal position in the comparison image data set (23, 23') includes: - The pixel values ​​of the image pixels in the comparison image data group (23, 23') are summed row by row in accordance with the main extension direction (13a), and a first distribution of the first pixel value sum is output, and - Determine the local extrema in the first distribution and output the rows of the local extrema as the vertical positions.

13. The method of claim 12, further comprising: - The pixel values ​​of the image pixels in the contrast image data set (23, 23') are summed column by column in the region of the local extrema, corresponding to the longitudinal direction (x), and a second distribution of the second pixel value sum is output. - Identify multiple local extrema in the second distribution and output a column of said local extrema, and - Determine (200) the image position of the region configured for the tip (17) of the support rod in the contrast image data set (23, 23') based on the corresponding row and column of the determined local extrema.

14. The method according to claim 1 or 2, wherein, The support rod (13) is capable of being suspended at a predetermined suspension position (15) on the longitudinal side (3a) of the tray (3), and the main extension direction (13a) extends transversely to the longitudinal side (3a), and The tray (3) has a reference structure (19a) that can be identified as a reference pixel arrangement (19b) in the comparison image data set (23, 23'), wherein, The suspension position (15) of the support rod (13) in the tray (3) is derived (300) by determining the distance between the reference pixel arrangement (19b) and the longitudinal position transverse to the main extension direction (13a) in the comparison image data group (23, 23') and identifying one of the predetermined suspension positions (15) of the longitudinal position based on the determined distance.

15. The method according to claim 1 or 2, wherein, The pixel value is a grayscale value, a color value, or a brightness value.

16. The method according to claim 1, wherein, The flatbed machine tool (1) is a laser cutting flatbed machine tool or a plasma cutting flatbed machine tool.

17. The method according to claim 6, wherein, The angle (5b) is greater than 45°.

18. The method according to claim 17, wherein, The angle (5b) is 60°.

19. The method according to claim 6, wherein, The shooting direction (5a) extends perpendicular to the main extension direction.

20. The method according to claim 7, wherein, The angle (5b) is greater than 45°.

21. The method according to claim 20, wherein, The angle (5b) is 60°.

22. The method according to claim 7, wherein, The shooting direction (5a) extends perpendicular to the main extension direction.

23. The method according to claim 8, wherein, The near-infrared spectrum is in the range of 845 nm to 855 nm and / or 935 nm to 945 nm.

24. The method according to claim 9, wherein, At least one of the irradiation directions (7a) forms an angle (7b) with the supporting plane (17a) that is in the range of 10° to 20°.

25. The method according to claim 24, wherein, The angle (7b) is 15°.

26. The method according to claim 9, wherein, The angle (7c) formed by at least one of the irradiation directions (7a) projected onto the support plane (17a) and the main extension direction (13a) of the support rod (13) is in the range of 10° to 25°.

27. The method according to claim 26, wherein, The angle (7c) is 15°.

28. The method according to claim 11, wherein the support rod (13) is arranged in the internal region of the tray (3).

29. A flatbed machine tool (1), comprising: The tray (3) has a plurality of support rods (13) suspended at a suspension position (15), each of the support rods (13) having a plurality of support rod tips (17) that define a support plane (17a). Camera (5), the camera being used to generate a two-dimensional camera image (21) or a two-dimensional dark field camera image of the tray, and Evaluation apparatus (11), said evaluation apparatus being designed to perform the method according to any one of claims 1 to 28, in, The camera (5) and / or the evaluation device (11) are designed for - Establish (100) a two-dimensional contrast image data set (23, 23') of the tray (3), the two-dimensional contrast image data set having multiple image pixels, assigning each image pixel a pixel value and pixel area unit of the support plane (17a), and the contrast image data set (23, 23') including a region configured as a local pixel value extremum (23a) in a uniform pixel value background (23b) for the tip of the support rod (17). And the evaluation device (11) is designed for - The longitudinal position of the region of the tip (17) of one of the plurality of support rods (13) configured in the contrast image data set (23, 23') is determined (200) by using the local pixel value extreme (23a). - The suspension position (15) of the support rod (13) in the tray (3) is derived (300) based on the longitudinal position in the comparative image data set (23) and the size of the pixel area unit in the longitudinal direction (x).

30. The flatbed machine tool (1) according to claim 29, wherein, The camera (5) is configured to: - Generate a two-dimensional camera image (21) of the tray in the visible or infrared spectrum, wherein the camera image (21) has image pixels arranged in rows and columns, and the image pixels are respectively configured with pixel area units of the supporting plane (17a), or - A set of contrast image data (23') is generated by capturing two-dimensional dark field camera images of the tray (3) in the near-infrared spectrum from the shooting direction (5a).

31. The flatbed machine tool (1) according to claim 30, wherein, The evaluation device (11) includes a processor (9), the processor being configured to: - The camera image (21) is Fourier transformed to the frequency domain and the transformed frequency data set (21a) is output. - The transformed frequency data set (21a) is filtered to highlight frequencies in the frequency domain belonging to the region configured for the tip (17) of the support rod, and - The filtered transform frequency data set is reverse-transformed by means of inverse Fourier transform, and the reverse-transformed data set is output as the comparison image data set (23).

32. The flatbed machine tool (1) according to any one of claims 29 to 31, the flatbed machine tool further comprising at least one lighting device (7) designed to illuminate the tray (3) with light in the near-infrared spectrum from one or more illumination directions (7a), wherein, At least one of the irradiation directions (7a) forms an angle (7b) with the support plane (17a) less than 30°, and at least one of the irradiation directions (7a) projected onto the support plane (17a) forms an angle (7c) with the main extension direction of the support rod less than 30°.

33. The flatbed machine tool (1) according to claim 29, wherein, The flatbed machine tool (1) is a laser cutting flatbed machine tool or a plasma cutting flatbed machine tool.

34. The flatbed machine tool (1) according to claim 32, wherein, At least one of the irradiation directions (7a) forms an angle (7b) with the supporting plane (17a) that is in the range of 10° to 20°.

35. The flatbed machine tool (1) according to claim 34, wherein, At least one of the irradiation directions (7a) forms an angle (7b) with the supporting plane (17a) of 15°.

36. The flatbed machine tool (1) according to claim 32, wherein, The angle (7c) formed by the projection of at least one of the irradiation directions (7a) onto the support plane (17a) and the main extension direction of the support rod is in the range of 10° to 25°.

37. The flatbed machine tool (1) according to claim 36, wherein, The angle (7c) formed by at least one of the irradiation directions projected onto the support plane (17a) in the irradiation direction (7a) and the main extension direction of the support rod is 15°.

Images