Funding institution

Abstract

Conveyor equipment (1), in particular for parcel sorting systems, wherein the conveyor equipment (1) comprises: a conveying device (3) which is designed to convey objects (7) along a conveying axis (A); at least one line-scan camera (5) whose line of sight is directed at the conveying device (3) at a specific effective angle of inclination (β) relative to the conveying axis (A); and an evaluation unit which is designed to generate a reconstructed image of each object (7) from the image signal of the at least one line camera (5) which was conveyed through the detection area of ​​the at least one line camera (5) by means of the conveying device (3), characterized in that The evaluation unit is further configured to identify different edges of the test box (7) in the reconstructed image of a cuboid test box (7) and to calculate, store and / or output, the effective inclination angle (β) of the line of sight of the at least one line camera (5) relative to the conveyor axis (A) from the distance of at least two of the identified edges (e1 to e3) along the conveyor axis (A) in a single reconstructed image of the test box (7).

Description

[0001] The present invention relates to conveying equipment, in particular for parcel sorting systems.

[0002] Corresponding conveying systems typically comprise a conveying device, at least one line scan camera, and an evaluation unit. The conveying device is designed to transport objects along a conveying axis. For example, the conveying device can include one or more driven conveyor belts on which the objects rest for transport. The at least one line scan camera has a line of sight which, depending on the specific installation, is aligned with the conveying device at a certain effective angle of inclination relative to the conveying axis. It is aligned with the conveying device in such a way that objects being conveyed along the conveying axis by the device pass through its detection range. The effective angle of inclination is understood to be the angle at which the line of sight intersects the conveying axis.With one or more deflecting mirrors in the optical path from the respective line scan camera to the conveyor axis, the line of sight can change direction. This can result in the line scan camera actually being at an angle to the conveyor axis that differs from the effective tilt angle described here. The evaluation unit is designed to generate a reconstructed image of each object from the image signal of the at least one line scan camera that has been conveyed through the detection range of the at least one line scan camera by the conveyor device.

[0003] It is necessary to know the exact position and / or orientation of the planned line scan cameras to enable improved image processing during normal operation of the conveyor system following calibration. Of particular importance is the effective tilt angle of each line scan camera relative to the conveyor axis. This effective tilt angle is crucial for determining the degree of optical distortion, which must be considered when evaluating the images during normal operation. Traditionally, the tilt angle and / or position of the line scan camera can be measured manually, which is time-consuming and inaccurate.

[0004] The object underlying the invention is therefore to provide a conveying device which enables a reliable and accurate determination of the position and orientation of the different line cameras provided.

[0005] This task is solved by the funding institution according to claim 1. Advantageous further training measures can be found in the following description and the dependent claims.

[0006] The conveying device according to the invention is characterized in that the evaluation unit is configured to identify different edges of the cuboid test box in the reconstructed image and to calculate the effective inclination angle of the line of sight of the at least one line camera relative to the conveying axis from the distance of at least two of the identified edges along the conveying axis in a single reconstructed image of the test box. The evaluation unit can then store and / or output this effective inclination angle for further use. This eliminates the need for manual measurement.

[0007] Such a conveyor system is capable of obtaining the aforementioned effective tilt angle, and thus at least this one essential parameter of the alignment, for the calculation of corresponding distortions, with just a single calibration measurement run. This enables simplified yet reliable calibration of the conveyor system or the respective line scan camera(s). To facilitate the reconstruction and evaluation of the image by the processing unit, the conveyor system can be equipped with a laser barrier that emits a signal as soon as the test box passes it. This, together with information about the conveyor system's speed, allows the positioning of the test box along the conveyor axis to be determined. It is also advantageous if the processing unit is informed about the specific position of the test box perpendicular to the conveyor axis.This positioning can, for example, be measured as the lateral distance of the test box from one of the two side edges of a conveyor belt, which acts as the conveying device, and manually provided to the evaluation unit. Preferably, this distance corresponds to the lateral distance (perpendicular to the conveying axis) of the test box from a designated laser barrier, whereby the position of the laser barrier can serve as the origin of a reference coordinate system in which the positioning and orientation of the respective line scan camera(s) is determined.

[0008] Preferably, the evaluation unit is designed to use information on the dimensions of the test box and / or information on at least the rough positioning and alignment of the respective line scan camera when calculating the effective tilt angle.

[0009] Relevant information can be requested from a user, defined as a calibration parameter and retrieved from memory, or obtained from another device that is part of or coupled to the conveying system, such as a volumetric measuring system. The information does not necessarily have to be complete, i.e., include all side lengths of the test box, but can be limited to those parameters actually required for the implemented calculation. Since different calculation methods exist, the composition of the information may vary between different designs and / or for different line scan cameras.Regarding the dimensions of the test box, for example, to calculate the respective effective tilt angle, it may suffice to know only the length of the side extending between the two edges whose distance is considered for calculating the effective tilt angle. Information on the rough positioning and orientation, as understood here, includes, for example, determining whether the respective line scan camera is positioned above, below, to the left, or to the right of the conveyor when viewed along the conveyor axis, and / or whether the respective line scan camera effectively looks at the conveyor along or opposite the conveyor axis. If a deflecting mirror is provided in the optical path from the respective line scan camera to the conveyor, then, in addition to information on the actual positioning and orientation of the respective line scan camera, the positioning and orientation of the associated deflecting mirror would also be important.Alternatively, the virtual positioning and alignment of the respective line scan camera (i.e., the effective positioning and alignment of the camera taking the deflecting mirror into account) can also be used to calculate the effective tilt angle. It is also possible to include more than one deflecting mirror in the optical path of the respective line scan camera. These deflecting mirrors can also be movable, in particular, pivotable.

[0010] Preferably, the evaluation unit is configured to calculate the effective inclination angle by calculating the actual distance between the two identified edges with the distance between these two edges in the reconstructed image. The evaluation unit is configured to query the actual distance of said edges from a user, retrieve it from a database, or obtain it from an external device for measuring the test box.

[0011] The actual distance between the two identified edges effectively refers to the respective side lengths of the cuboid test box that extend between the two edges used to calculate the effective angle of inclination. In the simplest case, a geometric trigonometric function such as sine, cosine, or tangent, or arcsine, arccosine, or arctangent, can be used for the calculation. For example, with a suitable selection of the respective edges and appropriate reconstruction of the image of the test box, the desired effective angle of inclination can be obtained as the arctangent of the quotient of the respective side length and the distance between the two corresponding edges in the reconstructed image of the test box.

[0012] Preferably, the at least one line scan camera is positioned and aligned such that its line of sight lies in a vertical or horizontal plane with the conveyor axis. The evaluation unit is configured to access relevant information regarding the coarse positioning and alignment of the at least one line scan camera in order to select which distance between the identified edges is to be used for calculating the effective inclination angle.

[0013] Limiting the specific positioning and orientation of each line scan camera significantly simplifies the calculation of the respective angle. Furthermore, these positioning and orientations facilitate the processing of the images from the respective line scan cameras during normal operation of the conveyor system.

[0014] Preferably, the at least one line camera is positioned and aligned such that its line axis, along which the pixels of the respective line camera are arranged, runs essentially perpendicular to the conveyor axis.

[0015] This facilitates the reconstruction and evaluation of the image of the test box during the calibration run, as well as the images during subsequent normal operation of the conveyor system.

[0016] Preferably, the evaluation unit is designed to obtain or determine an optical distance of the respective line camera from the conveyor axis as the optimized focus distance of the respective line camera, or to calculate it from a geometric calculation of the lengths of different identified edges in the reconstructed image with corresponding dimensions of the test box.

[0017] Optical distance is defined as the distance of the respective line scan camera from the conveyor axis along its associated optical path. The optical distance corresponds to the length of the line of sight from the respective line scan camera to the conveyor axis. With one or more deflecting mirrors, the optical distance is the sum of the lengths of the differently oriented segments of the line of sight. Optimized focus distance refers to a distance resulting from an optimization process for the line scan camera's focus. This optimization process can be performed manually or automatically and corresponds to a setting parameter of the line scan camera and / or its associated optics.To calculate the optical distance via the geometric calculation of the lengths of different identified edges, intrinsic properties of the respective line scan camera, such as its viewing angle and / or the specific positioning and orientation of the individual pixels, are regularly required.

[0018] Preferably, the evaluation unit is further configured to identify the edges of the test box in the reconstructed image using reference patterns from a first set of reference patterns provided on the test box.

[0019] The reference patterns in question are designed and arranged on the test box in such a way as to facilitate the identification and evaluation of the different edges of the test box in the reconstructed image. In particular, the reference patterns are clearly structured and preferably comprise only simple geometric shapes, such as triangles or quadrilaterals. They exhibit the highest possible contrast and are preferably pure black and white patterns. Preferably, each edge is assigned its own reference pattern, so that each recognized reference pattern can be uniquely identified as corresponding to a specific edge. Preferably, the reference patterns for different edges are all identical. Preferably, the reference patterns are symmetrically designed and centered on the respective edge of the test box.Preferably, each edge of the test box is provided with at least, and in particular with exactly, one corresponding reference pattern.

[0020] Preferably, the reference patterns of the first set of reference patterns are triangular patterns. These preferably extend along the entire length of an associated edge of the test box. In particular, each triangular pattern is preferably formed by four isosceles triangles whose bases, preferably adjacent to one another, are arranged along a line parallel to the respective edge, and especially directly along the respective edge. The four triangles preferably have the same height. The two bases of the two middle triangles have a first length. The two bases of the two outer triangles have a second length. Preferably, the second length is shorter than the first length.

[0021] Such reference patterns have proven to be particularly suitable in tests for reliably identifying the edges of the test box in the corresponding image.

[0022] Preferably, the evaluation unit is designed to identify the reference patterns of the first set of reference patterns based on a first set of comparison patterns with different distortion and / or compression.

[0023] This enables reliable identification of the edges of the test box even with comparatively large optical distortions.

[0024] Preferably, the evaluation unit is designed to determine a rough positioning and alignment of the respective line scan camera relative to the conveyor axis by identifying reference patterns from a second set of reference patterns on the different surfaces of the test box, and to take this into account when calculating the effective tilt angle. The reference patterns of the second set of reference patterns differ from the reference patterns of the first set of reference patterns.

[0025] This allows the rough positioning and orientation of each line scan camera to be determined automatically, without relying on complex and potentially error-prone user input or trusting that the test box was positioned according to specifications in a database. It is also conceivable to combine the reference patterns of the second set of reference patterns with those of the second set of reference patterns to form a higher-level set of reference patterns. The only important point is that the reference patterns used must still allow the identification of the edges of the test box in the reconstructed image.

[0026] Preferably, the evaluation unit is designed to determine a rough positioning and alignment of the respective line camera relative to the conveyor axis by comparing the lengths of the identified edges and their arrangement along the conveyor axis, and to take this into account when calculating the effective tilt angle.

[0027] Specifically, it is possible, for example, to conclude that the respective line scan camera is effectively aligned with the conveyor direction if the shortest edge is the last one captured or the first edge in the reconstructed image along the conveyor direction. This is the case because, when viewed along the conveyor direction, a line scan camera will last capture the edge of a cuboid test box resting on the conveyor, which is therefore furthest from the camera and thus appears shortest. Conversely, one can conclude, for example, that the respective line scan camera is effectively aligned against the conveyor direction if the shortest edge is the first one captured or the last edge in the reconstructed image along the conveyor direction.This makes it possible to automatically determine the rough positioning and alignment of the respective line scan camera without having to resort to a complex and potentially error-prone user input or relying on the fact that the respective line scan camera has been positioned and aligned according to the respective specifications.

[0028] Preferably, the evaluation unit is designed to calculate the determined inclination angle based on the aforementioned single reconstructed image and to verify and / or optimize at least one further reconstructed image of the test box for at least one further transport of the test box through the detection range of the at least one line camera.

[0029] Specifically, the effective tilt angle is calculated from the reconstructed image of the test box from at least one further calibration run and compared with the initially determined value. If a discrepancy is identified, the calculated tilt angle is adjusted. For this purpose, an average of the original tilt angle value and the subsequently determined values ​​can be stored and / or output. Furthermore, in multiple runs, tilt angle values ​​that deviate significantly from the others (and are therefore likely erroneous) can be disregarded when calculating the average. This allows for an even more precise and reliable determination of the effective tilt angle.It would also be possible to capture multiple images in different runs and use a combined optimization algorithm to calculate an optimized value for the effective tilt angle, which on average delivers the best result across all runs.

[0030] Preferably, the conveying device further comprises a volume measurement system and the evaluation unit is designed to access measurement data from the volume measurement system when reconstructing the image of the test box and / or when evaluating the reconstructed image of the test box.

[0031] For example, the edge positions identified by the evaluation unit in the reconstructed image can be compared with edge positions identified by the volumetric measurement system. If several parallel edges exhibit an identical offset perpendicular to their direction of travel, this indicates an incorrectly determined positioning by the respective line scan camera. Conversely, if the distance between several parallel edges varies with unequal offsets perpendicular to their direction of travel, this indicates an incorrectly determined orientation or a tilt angle that needs correction. In such cases, a stepwise optimization algorithm can vary the positioning and orientation until the combined offset and / or each individual offset of the identified edges falls below a predefined threshold.A similar systematic approach can also be applied to the previously described multiple passes through the test box and the combined evaluation of the different images of the test box.

[0032] Preferably, the conveying device comprises more than two, in particular three or six, corresponding line cameras, and the evaluation unit is designed to perform the respective operations separately for each of the line cameras.

[0033] A sufficient number of line scan cameras, when appropriately positioned and aligned, enables the most comprehensive and efficient imaging and analysis of objects conveyed through the cameras' field of view by the conveyor. For example, with three line scan cameras, one camera can be positioned centrally above the conveyor, while the other two are positioned at the same level as the conveyor, on opposite sides (left and right). With six line scan cameras, for example, the first camera can be positioned centrally above the conveyor, while the second camera is positioned centrally below it.A third and a fourth line camera can then be positioned to the left of the conveyor with different orientations, while a fifth and a sixth line camera are positioned to the right of the conveyor with different orientations.

[0034] Preferably, the evaluation unit is designed to determine, in particular with the aid of manual input from a user, the position and orientation of the respective line scan camera in a given reference coordinate system based on boundary conditions and / or parameters.

[0035] Under certain boundary conditions, specifications for the positioning and / or orientation of the different line scan cameras are required. These parameters include, for example, the focus setting of the different line scan cameras or the conveying speed of the conveyor system. The specific positions and orientations of the different line scan cameras within a common reference coordinate system facilitate, for example, the combined evaluation of the images from the different line scan cameras.

[0036] Preferably, the evaluation unit is designed to query and consider one or more of the following pieces of information from a user to determine the position and orientation of the respective line camera in the specified reference coordinate system: rough positioning and / or alignment of the respective line camera relative to the conveyor axis; position and orientation of a deflecting mirror assigned to the respective line camera in the specified reference coordinate system; position and orientation of the conveyor axis in the specified reference coordinate system; approximate optical distance of the respective line camera from the conveyor axis along the line of sight of the respective line camera; position and / or orientation of the test box in the specified reference coordinate system; dimensions of the test box.

[0037] Such information significantly facilitates the determination of the specific position and orientation in the reference coordinate system.

[0038] The invention is described below by way of example only, with reference to the drawings. It shows: Fig. 1 a vertical cross-section of a schematic setup of an exemplary conveyor system with a line camera effectively looking at the conveyor system from an oblique angle above; Fig. 2 a schematic reconstructed image of the test box from Fig. 1 taken with the line camera shown there; Fig. 3 an enlarged view of the reconstructed image of the Fig. 2 with comparative data from a volume measurement system; Fig. 4 a horizontal cross-section of a schematic setup of an exemplary conveyor system with a line camera which effectively looks obliquely to the side at the conveyor system; Fig. 5 an exemplary reconstructed image of a test box taken from an oblique angle above; Fig. 6 an exemplary reconstructed image of a test box taken from an oblique side angle; Fig. 7 an example of a reference pattern from a first set of reference patterns; Fig. 8 to Fig. 10 different comparison patterns to the reference pattern from Fig. 7 with different heights and / or optical distortion; Fig. 11 to Fig. 15 different windows of an exemplary calibration wizard, which can be executed by the evaluation unit to guide a user through the calibration of the conveying equipment; and Fig. 16 A schematic representation of a view from above against the direction of conveyance to clarify the reference values ​​used to calculate the y-position of the associated line camera.

[0039] Fig. Figure 1 shows a side view of an exemplary conveying device 1 with a conveying unit 3 and at least one line-scan camera 5. The conveying unit 3 can be designed as a simple conveyor belt and is configured to convey objects, such as the test box 7 shown here, along a conveying axis A (in Fig. 1 from left to right). The line scan camera 5 shown is positioned above the conveyor device 3 and aligned parallel to the conveyor axis A (see: “real position”).

[0040] The line of sight of the line camera 5 is deflected onto the conveyor 3 by a deflecting mirror 9. In this context, the line of sight of the line camera 5 is therefore not effectively parallel to the conveyor axis A, but rather at an effective angle of inclination β. The effective or virtual position (see: "virtual position") of the line camera 5, determined by the positioning and orientation of the deflecting mirror 9, is still located above the conveyor 3. However, the line camera 5 is effectively oriented from above and against the direction of conveying along the conveyor axis A (from left to right in the image). It is also possible to omit the deflecting mirror 9 and position the line camera 5 directly according to the virtual position and orientation shown. Furthermore, the use of more than one deflecting mirror 9 for the line camera 5 is possible.These deflecting mirrors 9 can be fixedly mounted or provided as movable elements and generally allow for the realization of more compact overall designs.

[0041] The corresponding optical distance f f This corresponds to the distance at which the focus of the line camera must be set in order to sharply image the conveyor axis A.

[0042] In the example shown, both the real and virtual positions of the line scan camera 5 lie in a vertical plane through the conveyor axis A (i.e., centered above the conveyor device 3). The entire line of sight of the line scan camera 5 also lies in this vertical plane. This represents a potential constraint on the positioning and orientation of the line scan camera 5 and the deflecting mirror 9, which simplifies calculations for the positioning and orientation of the respective line scan camera 5.

[0043] The Fig. 2 and Fig. Figure 3 shows a reconstructed image of the test box 7. This image was assembled from the images taken by the line scan camera 5 while the test box 7 was being conveyed through the detection area of ​​the line scan camera 5 by the conveyor device 3. The reconstruction of the image is performed by a suitably designed evaluation unit (not shown separately). For this purpose, the evaluation unit can, for example, also access information about the speed at which the test box 7 was conveyed through the detection area of ​​the line scan camera 5 and information about the sampling frequency of the line scan camera 5. In the example shown, the image of the test box is reconstructed such that the size dx in Fig. 1 of size dx in Fig. 2 corresponds. The size dx indicates the distance between the two edges e1 and e2 from the effective viewing angle of the line camera 5.

[0044] As in the Fig. 2 and Fig. 3 indicated and in the Fig. 5 and Fig. As shown in Figure 6, test box 7 comprises reference patterns 11 of a first set of reference patterns, each positioned along different edges e1 to e3 of test box 7. These reference patterns 11 of the first set could, for example, be triangular patterns composed of four isosceles triangles of equal height and different widths aligned along the respective edges e1 to e3. Other reference patterns, such as those with more or fewer triangles or other geometric and non-geometric structures, are also conceivable. To be reliably identifiable in the reconstructed image, reference pattern 11 requires sufficiently high contrast and adequate illumination. It is also possible for the evaluation unit to binarize the reconstructed image to facilitate the identification of the respective reference patterns 11.The evaluation unit must also take into account different rotations of the reference patterns 11 when searching for them.

[0045] To identify the reference patterns 11 in the reconstructed image and thus to identify the associated edges, the evaluation unit preferably uses corresponding sets of comparison patterns. These sets preferably each contain the reference pattern in its original form as well as the reference pattern with different distortion and / or compression, as is the case, for example, in the Fig. 7, Fig. 8, Fig. 9 to Fig. 10 is shown. While Fig. 7 shows the reference pattern in the original, the comparison pattern of the Fig. 8 a reduced height and the comparison pattern of the Fig. 9 a further reduced height. The comparison pattern of the Fig. Image 10 exhibits reduced height and optical distortion. This comparison pattern of Fig. 10 is, for example, in relation to the reconstructed image of the Fig. 5 significantly more suitable than the original reference pattern of the Fig. 7, the top reference pattern 11 in Fig. 5 to identify.

[0046] Returning to the Fig. 2 and Fig. The evaluation unit 3 is designed to identify at least edges e1 to e3 in the reconstructed image, particularly based on the provided reference pattern 11. Edge e1 in the reconstructed image corresponds to the front lower edge of the test box 7, while edges e2 and e3 correspond to the front and rear upper edges of the test box 7, respectively. Edges e1 to e3 are those edges that run perpendicular to the conveyor axis A. To clearly define these edges and to facilitate the subsequent calculation of the effective inclination angle β, the test box 7 is aligned along the conveyor axis A and positioned as centrally as possible on the conveyor 3. To ensure that its position and alignment remain unchanged, the test box 7 can be fixed to the conveyor 3, for example, using adhesive tape.

[0047] The evaluation unit can determine the virtual distance between different edges from the reconstructed image, in particular the distance dx between the first two edges e1 and e2. The evaluation unit can then combine this virtual distance with information about the real height h0 of test box 7 (i.e., the real distance between the two edges e1 and e2) to obtain the angle of inclination β. For this purpose, the evaluation unit calculates, for example, the arctangent of the quotient of the real height h0 of test box 7 (as the dividend) and the virtual distance dx between edges e1 and e2 (as the divisor): β=arctan(h0dx)

[0048] The evaluation unit can, for example, query the actual distance h0 between the respective edges from a user, obtain it from a database containing dimensions for test box 7, or receive it from a separate device for measuring test box 7. The calculated value for the inclination angle β can then be stored and / or output by the evaluation unit.

[0049] Taking into account the above-described specifications that the intended line camera 4 and the associated line of sight lie in a common vertical plane with the conveyor axis A, the evaluation unit can be optimized using a focus distance f. f , in which case, in the example of the Fig. 1. The edge e1 is imaged as sharply as possible, at least to the virtual position and orientation of the line camera 5 in a given reference coordinate system (with respect to the origin U and the coordinate axes x, y and z at the bottom left). Fig. 1) Determine. With knowledge of the position and orientation of the provided deflecting mirrors 9, the evaluation unit can even determine the actual position and orientation of the respective line scan camera 5 in the specified reference coordinate system. The origin U for a corresponding reference coordinate system can, for example, be the center of the conveyor device 3 at the level of a provided laser barrier 13. Alternatively, the origin U of the corresponding reference coordinate system can be located not in the center, but on one of the two sides, for example, on the right side, of the conveyor device 3.

[0050] Which of the identified edges e1 to e3 in the reconstructed image correspond to which edges of test box 7 and are to be analyzed, results from information on a rough positioning and orientation of the respective line camera 5. This can be retrieved or specified by a user via reference pattern 15 of a second set of reference patterns (see Fig. 5 and Fig. 6) can be determined directly from the reconstructed image of test box 7 or by comparing the lengths of the different identified edges along the conveying direction. For example, the top of test box 7 can be labeled with a capital T for "top" and / or the right side of test box 7 with an R for "right".

[0051] By adhering to the relevant specifications for the rough positioning and alignment of the line scan camera 5 and the test box 7 relative to or on the conveyor device 3 or to the conveyor axis A, the evaluation unit is able to determine detailed information on the positioning and alignment of the respective line scan camera from a single reconstructed image from a single measurement run. Validation and / or optimization of the determined information based on further measurement runs is possible.

[0052] The evaluation unit can also be configured to interact with a further provided volume measurement system (not shown) and, based on corresponding information, to validate or optimize the information obtained for the respective line scan camera 5. Fig. 3. Position information can be obtained. i1 until e i3to the positions of the different edges e1 to e3, which result from the calculated information on the positioning and orientation of the respective line camera 5 with corresponding position information e v1 until e v3 a further planned volume measurement system is to be compared. If the same offset results for the different edges e1 to e3, this indicates a necessary correction of the information on the position of the line camera 5 along the conveying direction x. If different offsets result for the different edges e1 to e3, this indicates, for example, a necessary correction of the information on the alignment of the line camera 5 with the conveying axis A (in particular the necessary correction of the tilt angle β).

[0053] More than one line scan camera 5, for example 3 or 6 line scan cameras, can also be provided. The respective information for each line scan camera 5 is then acquired independently for the different line scan cameras 5. The different line scan cameras 5 are preferably arranged in different positions and orientations around the conveyor device 3. For example, at least one line scan camera 5 can also be positioned with its line of sight in a horizontal plane through the conveyor axis A and effectively aligned with the conveyor direction, as shown in Fig. Figure 4 illustrates this. The selection of the edges whose distance from each other is analyzed, as well as the corresponding dimensions of the test box 7 used (and, if necessary, the specific geometric calculation), must be adapted to the corresponding rough positioning and orientation of the respective line scan camera 5. The reference point for determining the effective optical distance is also variable and depends, for example, on the choice of the specific edge on which the respective line scan camera 5 is focused. Varying and adjusting the respective reference points and the associated calculations is within the capabilities of a person skilled in the art.

[0054] However, it should be noted that for each measurement run and each calculation, fixed assumptions should apply, such as a constant focus distance, conveying speed, acquisition frequency, and the positioning and alignment of the line scan camera 5 and test box 7 on the conveyor device 3. Certain positioning and alignment of the line scan cameras 5 are particularly easy to evaluate and are therefore preferable. For example, positioning and aligning the line scan cameras 5 according to the previously described horizontal and vertical planes through the conveyor axis A is advantageous. It is also advantageous to align the respective line scan cameras 5 such that their line axes, along which the pixels of the respective line scan camera 5 are arranged, are essentially perpendicular to the conveyor axis A.

[0055] From the inclination angle β described above, it is possible with the setup made of Fig. 1 For example, it is possible to define the virtual position of the line camera 5 along the conveyor axis A as the sum of the distances traveled x. L of test box 7 since the front edge e1 of test box 7 enters the laser barrier 13 with the product of the optimized focus distance f f to calculate the respective line scan camera 5 with the sine of the tilt angle β: x=xL+ff⋅sin(β)

[0056] However, this only applies as long as the associated line camera 5 is effectively oriented opposite to the conveying direction x. If, on the other hand, the line camera 5 is effectively oriented along the conveying direction x, the formula must be rearranged accordingly: x=xL+dx−ff⋅sin(β)

[0057] The position of the line scan camera 5 along the vertical z-axis results from the product of the optimized focus distance f. f with the cosine of the angle of inclination β: z=ff⋅cos(β)

[0058] In the example shown, the position of the line scan camera 5 along the horizontal y-axis, perpendicular to the conveyor axis A, is directly determined as the y-coordinate of the center point of the recorded images. This can also be determined by the lateral distance yo of the test box 7 from the origin, for example, the position of a laser barrier at the edge of the conveyor 3, and a distance d. y the corresponding side edge P l test box 7 from the image center P m calculable (cf. Fig. 16): y=yo+dy(Pm−Pl)

[0059] The rotational orientation of the line camera 5 about the y-axis is indicated in the example shown by the tilt angle β. The rotational orientation about the conveyor axis A is determined by the constraint that the respective line camera 5 is positioned with its line of sight in a vertical plane. The rotational orientation about the line camera 5's line axis is determined by the perpendicular alignment of the line camera's axis to the conveyor axis A, as described above. Thus, all information regarding the position and orientation of the line camera 5 with respect to the respective reference coordinate system is uniquely defined and determinable.

[0060] In the design of Fig. 4. The angle of inclination β is obtained as the arctangent of the quotient of the distance dx between the two rear edges in the reconstructed image (as dividend) and the real width b0 of the test box 7 perpendicular to the conveying axis A (as divisor): β=arctan(dxb0)

[0061] At the in Fig. 4 plotted optimized focus distance f f The virtual x-position is derived here from the difference in the distance traveled along the shortest edge x. L (as minuend) and the projection length dx (as subtrahend) as well as the product of focus distance f f with the sine of the tilt angle β, if the line camera is effectively aligned in the conveying direction as shown (i.e., looking at test box 7 from behind): x=xL−ff⋅sin(β)

[0062] The position of the line scan camera 5 along the vertical z-axis is determined by the z-coordinate of the center point of the captured images. It corresponds to a distance d. z of the image center px mid from the top edge z1(e2) of test box 7: z=dz(pxmid−z1(e2)

[0063] The y-coordinate of the virtual position of the line scan camera 5 is obtained as the difference between the lateral distance d and the y-coordinate of the lateral distance d. y,O the nearest edge of the test box aligned parallel to the conveyor axis A (left in Fig. 4) from the respective origin (for example, the position of the laser barrier 13) as minuend and the product of the optimized focus distance f f with the cosine of the angle of inclination β as the subtrahend: y=dy,o−ff⋅cos(β)

[0064] The rotational orientation of the line camera 5 about the z-axis is indicated by the tilt angle β. The rotational orientation about the conveyor axis A is determined by the constraint that the respective line camera 5 is positioned with its line of sight in a horizontal plane. The rotational orientation about the line camera 5's line axis is determined by the perpendicular alignment of the line camera's axis to the conveyor axis A, as described above.

[0065] The formulas mentioned above apply to the line camera 5 of the Fig. 4, which effectively looks at test box 7 from the rear right. For other positions and orientations, the formulas must be rearranged accordingly. For example, for an effective orientation from the front left towards test box 7 with the actual width b0, the following formulas apply: x=xL+dx+ff⋅sin(β) and y=dy,o+b0+ff⋅cos(β)

[0066] The specific forms also vary depending on the chosen reference values ​​or parameters. A person skilled in the art can easily make the necessary adjustments based on the present disclosure.

[0067] Finally, attention is drawn to the Fig. 11, Fig. 12, Fig. 13, Fig. 14 to Fig. Reference is made to page 15. These show different steps or windows of an example calibration assistant, which can be executed by the evaluation unit to support the calibration described above.

[0068] In the Fig. The start screen shown in Figure 11 lists various pieces of information about the associated conveyor system. There is an activation button (see: “ICR Tunnel Calibration”) to start the guided calibration process for the designated line scan cameras. 5. When this button is pressed, the following window appears. Fig. 12.

[0069] The IP addresses of the various line scan cameras 5 connected to the evaluation unit are listed there. Each of these IP addresses is assigned three fields to select or enter the rough positioning, rough orientation, and distance between the respective line scan camera 5 and an associated deflection mirror 9. For each line scan camera 5, one of the following can be selected as the rough position: top, if the respective line scan camera 5 is located above the conveyor 3 when viewed along the conveying direction x; bottom, if the respective line scan camera 5 is located below the conveyor 3 when viewed along the conveying direction x; left, if the respective line scan camera 5 is located to the left of the conveyor 3 when viewed along the conveying direction x; right, if the respective line scan camera 5 is located to the left of the conveyor 3 when viewed along the conveying direction x;(“Right”) if the respective line camera 5 is positioned to the right of the conveyor 3 when viewed along the conveying direction x. A corresponding selection is available for the coarse alignment of each line camera 5, based on the positioning of the fan on the respective line camera 5. The distance (see “Mirror Distance”) to the corresponding deflection mirror 9 can be entered manually and / or pre-set with a standard value. If there is no deflection mirror 9, the corresponding field can be omitted. To illustrate the respective information, a diagram of the schematic setup of the conveyor 1 with a line camera 5 positioned above the conveyor 3 is shown on the right side of the window. The line camera's fan points downwards, i.e., towards the conveyor 3 (corresponding to the information for the first line camera in the list on the left).

[0070] On the window of the Fig. For example, the window of 12 follows. Fig. 13. This is used to input or determine at least the approximate focus distance f. f (cf. rough focus distance). This essentially corresponds to the length of the entire line of sight of the respective line camera 5 up to the conveyor axis A or to the test box 7 (cf. calibration target). This distance can be measured manually or predefined. Alternatively, it can also be determined automatically using an "autofocus". For this purpose, the focus distance f can be calculated using a suitable camera model and the dimensions of the test box 7. f The distance between projected edges can be varied stepwise until it matches the corresponding side length. This depends on which edges are referenced and how the focus distance f is set. f If precisely defined, a correction of a model-based obtained focus distance f is possible. fby a corresponding projection length, which results from the quotient of a corresponding dimension of the test box as the dividend and the cosine of the inclination angle β as the divisor. Regardless of how the focus distance f is determined. f Once the focus distance has been determined, it is important to know the value for f. f The focus distance f must not be varied during the recording of an image of test box 7 for evaluation. A sketch on the right of this window shows the specific focus distance f to be entered. f illustrated.

[0071] The evaluation unit can, if necessary, independently infer further information regarding the rough positioning and / or orientation of the respective line scan cameras 5. This works, for example, by comparing the lengths of the different identified edges in the reconstructed image. If the edge e1 with the shortest length is captured first by a line scan camera 5, the respective line scan camera 5 is effectively (i.e., taking into account any deflection mirrors 9) oriented against the conveying direction x (cf., for example, the Fig. 1 and Fig. 2) If the edge e1 with the shortest length is recorded last by a line camera 5, the respective line camera 5 is effectively (i.e. taking into account any deflecting mirrors 9 provided) aligned in the conveying direction x.

[0072] In the following window of Fig. In step 14, the user is prompted to specify the positioning of the test box 7 on the conveyor 3. In this example, this is done by positioning it on the conveyor belt 3 perpendicular to the conveyor axis A (here along the y-axis).

[0073] It should be noted that the calibration assistant described here only requests a specific selection of information that it actually needs to calculate the desired values. The specific combination of information requested depends in particular on the specifications for the positioning and orientation of the line scan cameras 5 and the test box 7. Varying the specific selection of requested information with different assumptions and / or specifications is within the expertise of a specialist and therefore does not need to be discussed further here.

[0074] In the last window accordingly Fig. In step 15, the calculated information for each designated line scan camera 5 is displayed, along with the status of the associated calculations. While only an initial set of information regarding the position and orientation in the defined reference coordinate system has been determined for the topmost listed line scan camera 5 (see the status "initial"), fine-tuning is already underway for the second listed line scan camera 5 (see the status "fine-tune"). If necessary, a recalculation of the respective information can be requested via corresponding fields. The system is awaiting the calculation results for the third and fifth listed line scan cameras 5 (see the status "waiting"). Specific information regarding position or orientation is not yet available for these cameras. The test box 7 has not yet appeared in the images from the fourth listed line scan camera 5, or at least has not yet been identified.Here too, no concrete information regarding the position and orientation is yet available. For the sixth and last listed line scan camera 5, speed information for the conveyor device 3 and / or the sampling frequency of the line scan camera 5 is missing (see: the status "no speed"). This prevents the evaluation unit from assembling the recordings from the associated line scan camera 5 into a corresponding image and thus from beginning the evaluation of a reconstructed image.

[0075] Finally, the window contains the Fig. 15. There are also two selection fields: "Reset All" and "Finish". The first of these two fields deletes all calculated information and restarts the corresponding analysis and calculation steps. The second field ends the calibration assistant, but is currently grayed out because the calibration is not yet complete for all line scan cameras 5.

[0076] Finally, it should be noted that the present description merely includes exemplary embodiments to illustrate the present invention. Modifications and / or further developments of the described embodiments, which are covered by the claims, also fall within their scope of protection, even if they are not explicitly described here. Reference symbol list 1 funding facility 3 Conveyor device 5-line camera 7 Test box / (promotional) object 9 deflecting mirrors 11 reference patterns of a first set of reference patterns 13 laser barriers 15 reference patterns of a second set of reference patterns

Claims

[1] Conveyor (1), in particular for parcel sorting systems, wherein the conveyor (1) comprises: a conveying device (3) which is designed to convey objects (7) along a conveying axis (A); at least one line-scan camera (5) whose line of sight is directed at the conveying device (3) at a specific effective angle of inclination (β) relative to the conveying axis (A); and an evaluation unit which is designed to generate a reconstructed image of each object (7) from the image signal of the at least one line camera (5) which was conveyed through the detection area of ​​the at least one line camera (5) by means of the conveying device (3), characterized by , that The evaluation unit is further configured to identify different edges of the test box (7) in the reconstructed image of a cuboid test box (7) and to calculate, store and / or output, the effective inclination angle (β) of the line of sight of the at least one line camera (5) relative to the conveyor axis (A) from the distance of at least two of the identified edges (e1 to e3) along the conveyor axis (A) in a single reconstructed image of the test box (7). [2] Conveying device (1) according to claim 1, characterized by , that the evaluation unit is designed to use information on the dimensions of the test box (7) and / or information on at least the rough positioning and alignment of the respective line scan camera (5) when calculating the effective tilt angle (β). [3] Conveying device (1) according to claim 2, characterized by , that the evaluation unit is designed to calculate the effective inclination angle (β) by calculating the real distance between the two identified edges (e1 and e2) with the distance between these two edges in the reconstructed image, wherein the evaluation unit is designed to query the actual distance (h0) of said edges from a user, to obtain it from a database or from an external device for measuring the test box (7). [4] Funding institution (1) according to any of the preceding claims, characterized by , that at least one line camera (5) is positioned and aligned such that its line of sight lies in a vertical or horizontal plane with the conveyor axis (A), wherein the evaluation unit is designed to access relevant information on the rough positioning and alignment of the at least one line camera (5) in order to select which distance of the identified edges is to be used to calculate the effective tilt angle (β). [5] Funding institution (1) according to any of the preceding claims, characterized by , that the at least one line camera (5) is positioned and aligned such that its line axis, along which the pixels of the respective line camera (5) are arranged, is substantially perpendicular to the conveyor axis (A). [6] Funding institution (1) according to any of the preceding claims, characterized by , that the evaluation unit is designed to determine an optical distance of the respective line camera (5) from the conveyor axis as the optimized focus distance (f f) of the respective line camera (5) or to determine, or to calculate from a geometric calculation of the lengths of different identified edges in the reconstructed image with corresponding dimensions of the test box (7). [7] Funding institution (1) according to any of the preceding claims, characterized by , that the evaluation unit is further configured to identify the edges of the test box (7) in the reconstructed image using reference patterns (11) of a first set of reference patterns provided on the test box (7). [8] Conveying device (1) according to claim 7, characterized by , that the reference patterns (11) of the first set of reference patterns are triangular patterns which preferably extend along the entire length of an associated edge of the test box (7); wherein each triangular pattern is formed in particular by preferably four isosceles triangles, the bases of which are arranged, preferably adjoining one another, along a line parallel to the respective edge; wherein the four triangles preferably have the same height, the two bases of the two middle triangles have a first length and the two bases of the two outer triangles have a second length; where the second length is preferably smaller than the first length. [9] Conveying device (1) according to claim 7 or 8, characterized by , that the evaluation unit is trained to identify the reference patterns (11) of the first set of reference patterns using a first set of comparison patterns with different distortion and / or compression. [10] Funding device (1) according to any of the preceding claims, characterized by , that the evaluation unit is designed to determine a rough positioning and alignment of the respective line camera (5) relative to the conveyor axis (A) by identifying reference patterns (15) of a second set of reference patterns on the different surfaces of the test box (7) and to take this into account when calculating the effective tilt angle (β), wherein the reference patterns (15) of the second set of reference patterns differ from the reference patterns (11) of the first set of reference patterns. [11] Funding facility (1) according to any of the preceding claims, characterized by , that the evaluation unit is designed to determine a rough positioning and alignment of the respective line camera (5) relative to the conveyor axis (A) by comparing the lengths of the identified edges and their arrangement along the conveyor axis (A) and to take this into account when calculating the effective tilt angle (β). [12] Funding facility (1) according to any of the preceding claims, characterized by , that the evaluation unit is designed to calculate the determined effective inclination angle (β) on the basis of the above-mentioned single reconstructed image and to verify and / or optimize on the basis of at least one further reconstructed image of the test box (7) for at least one further transport of the test box (7) through the detection range of the at least one line camera (5). [13] Funding facility (1) according to any of the preceding claims, characterized by , that the conveying device (1) further comprises a volume measurement system and the evaluation unit is designed to use measurement data from the volume measurement system when reconstructing the image of the test box (7) and / or when evaluating the reconstructed image of the test box (7). [14] Funding facility (1) according to any of the preceding claims, characterized bythat the conveying device (1) comprises more than two, in particular three or six corresponding line cameras (5) and the evaluation unit is designed to perform the respective operations for each of the line cameras (5) separately. [15] Funding facility (1) according to any of the preceding claims, characterized by , that the evaluation unit is designed to determine, in particular with the aid of manual input from a user, the position and orientation of the respective line camera (5) in a given reference coordinate system based on boundary conditions and / or parameters. [16] Conveying device (1) according to claim 15, characterized by, that the evaluation unit is designed to query and consider one or more of the following pieces of information from a user in order to determine the position and orientation of the respective line camera (5) in the specified reference coordinate system: rough positioning and / or orientation of the respective line camera (5) relative to the conveyor axis (A); position and orientation of a deflecting mirror (9) assigned to the respective line camera (5) in the specified reference coordinate system; position and orientation of the conveyor axis (A) in the specified reference coordinate system; approximate optical distance of the respective line camera (5) from the conveyor axis (A) along the line of sight of the respective line camera (5); position and / or orientation of the test box (7) in the specified reference coordinate system; dimensions of the test box (7).

Images