Method for the automated image-based detection of nodes in a contrast line grid

Abstract

The present invention relates to a method for detecting the surface geometry of an object and in particular changes in the surface geometry of an object using a contrast line grid (1) as well as to a method for the automated detection of nodes (7) from a plurality of images of a contrast line grid (1).

Description

[0001] METHOD FOR AUTOMATED IMAGE-BASED CAPTURE OF

[0002] NODES OF A CONTRAST LINE GRID

[0003] BACKGROUND OF THE INVENTION

[0004] AREA OF INVENTION

[0005] The present invention relates to a method for automated image-based detection of nodes of a contrast line grid, as well as a method for detecting the surface geometry and in particular changes in the surface geometry of an object with a contrast line grid.

[0006] STATE OF THE ART

[0007] In a wide variety of applications, it is necessary to capture the surface geometry of an object or to determine any changes or deviations of the geometry from a reference state. One known method is to determine the geometry using photogrammetry techniques from images of the object taken from different positions. In this context, it is also known to apply contrast line grids to the surfaces under investigation, for example by printing, to enable the determination of homologous points and the geometry.

[0008] Such well-known methods are used, for example, in the analysis of forming processes for pressed or deep-drawn components. In this process, a corresponding contrast line grid is applied to a test sheet, and the deep-drawn or pressed component is imaged from several positions after forming in order to determine the component's geometry from the various images.

[0009] However, capturing the various nodes in a contrast line grid representation and corresponding the nodes in multiple images is time-consuming, which can significantly increase processing time. Alternatively, the 3D coordinates of the nodes in a contrast line grid can be captured tactilely using a coordinate measuring machine, but this is also time-consuming.

[0010] REVELATION OF THE INVENTION

[0011] TASK OF INVENTION

[0012] It is therefore an object of the present invention to provide a method for the simpler detection of nodes of a contrast line grid and an improved method for determining the surface geometry of an object or, in particular, changes in the surface geometry, which can be carried out automatically, more easily and quickly, while ensuring high accuracy of detection.

[0013] TECHNICAL SOLUTION

[0014] This problem is solved by a method for the automated detection of nodes of a contrast line grid with the features of claim 1 and a method for detecting the surface geometry of an object and, in particular, changes in the surface geometry of an object with the features of claim 8. Advantageous embodiments of the invention are the subject of the dependent claims.

[0015] To solve the problem, the use of a contrast line grid with contrast lines is proposed, where the contrast lines define nodes at their intersections. The contrast line grid has a first node as its origin. The origin is, for example, the node of the contrast line grid from which the search for further nodes is to begin, and can be part of a first grid cell defined by the contrast lines running through the origin of the contrast line grid and adjacent nodes. The contrast line grid is characterized by the fact that a marker is located adjacent to the node at the grid's origin. This marker can be detected by image recognition and serves to uniquely identify the origin. The marker can, in particular, be located within the first grid cell.This automates and simplifies the capture of the nodes of the contrast line grid. Manual user input to define the origin, for example by clicking on images, is eliminated, thus reducing processing time.

[0016] The marker at the origin can be a circle, as simple methods for capturing such shapes are known in image recognition. However, other marker shapes, such as squares, ellipses, or other forms, are also possible.

[0017] For automated recognition, the marker and the adjacent contrast lines should be clearly aligned with each other, for example by placing the circular marker closer to the origin than to other nodes.

[0018] For example, the marker can be positioned eccentrically in the first grid cell at the origin of the contrast line grid.

[0019] For the automated acquisition of the nodes of a contrast line grid, various parameters can be used by an algorithm or evaluation software to, for example, adjust the size and position of the marker relative to the origin of the contrast line grid. These parameters can include, for example, the size of the marker, in particular the radius of a circular marker or at least one radius of an elliptical marker, or the side length of a square marker, and / or one or more distances of the marker to the contrast lines through the origin of the grid, and / or the asymmetry of the distances of the marker to the contrast lines, and / or the width of one or more contrast lines.

[0020] Another possible parameter could be the selection of the color or brightness relation of the contrast line grid and / or marker to the background, such as a light contrast line grid on a dark background or vice versa.

[0021] These parameters facilitate the automated detection of contrast line grid nodes that are adjacent to the origin and / or to other previously detected nodes. The parameters can be determined automatically or provided by the user. When determining the parameters automatically, a process can be implemented in which the parameters are evaluated through incremental adjustments and repeated analysis. Regardless of whether the parameters are provided manually or automatically, they can be saved as needed, for example, to be available for reusing a corresponding contrast line grid.

[0022] After identifying the origin using the marker, starting from the origin of the contrast line grid, the adjacent nodes can be successively detected, using the parameters mentioned above, such as line width and / or brightness relation. The detection of nodes can continue until a predetermined number of nodes have been detected or no further nodes can be found.

[0023] The term "recording the nodes" can be understood in particular as determining the coordinates of the nodes in a coordinate system.

[0024] The captured nodes of the contrast line grid can be assigned to a reference grid, for example a planar two-dimensional or undistorted contrast line grid, in order to make it easier to compare changes or deviations of a surface geometry.

[0025] The contrast line grid can be a rectangular grid, especially with square grid cells, although other grid shapes, such as grids with triangular or other polygonal grid cells, are also conceivable.

[0026] The contrast line grid can also be applied to a three-dimensional surface while preserving its topology. Its nodes can be determined based on the spatially oriented images and the parameters set by the user.

[0027] In the method for capturing the surface geometry of an object, and in particular changes to its surface geometry, a grid of contrast lines is applied to the object's surface. To determine the geometry, several images of the object's surface with the grid are taken from multiple different positions. These surface images are then analyzed using photogrammetry methods to determine their inner and outer orientation. Subsequently, the nodes of the grid can be automatically captured using the method described above.To detect surface deformation, the contrast line grid can be applied to or attached to the undeformed surface. At least several images of the deformed surface of the object can be taken from multiple positions using the contrast line grid to determine the geometry. The contrast line grid applied to the surface of the undeformed object can serve as a reference grid against which the changes or deformations can be determined. Deformation detection focuses on the change from an initial state (undeformed state) to a deformed state. Accordingly, the initial state itself could have been the result of a deformation.

[0028] Accordingly, the contrast line grid applied to a surface can also be recorded in two or more deformation states and used for deformation analysis.

[0029] Using the various images from different positions, especially if these have been previously oriented in space, coordinates of the nodes and / or normal vectors on the surface can be determined at the nodes and / or vectors along the directions of the intersecting contrast lines in order to characterize the surface geometry.

[0030] Optionally, a captured, orthonormalized coordinate system for each node can be output. This is derived from the position and directions of the intersecting line segments. The resulting surface normal can also be output. The calculation results can be dynamically displayed to the user on a graphical user interface, for example, as a 3D scene or in a dialog box, as soon as they are available.

[0031] BRIEF DESCRIPTION OF THE FIGURES

[0032] The attached drawings show in a purely schematic way in

[0033] Figure 1 shows an example of a contrast line grid according to the invention.

[0034] Figure 2 shows a representation of several images for the automated determination of a geometry; Figure 3 shows a representation of an image of an object with a contrast line grid in the form of a crash test dummy with a vest on which the contrast line grid is arranged.

[0035] Figure 4 shows a detailed representation of the origin of the contrast line grid from Fig. 3, with a representation of the input of parameters for the contrast line grid to capture the nodes.

[0036] Figure 5 shows a further detailed representation of the origin of the contrast line grid, similar to Figure 4, with a representation of the input of different parameters for capturing the nodes of another contrast line grid and in

[0037] Figure 6 shows a representation of an image of part of the object from Figure 3, showing the determined surface normals at the nodes and vectors with directions to the adjacent nodes.

[0038] EXAMPLES OF EXECUTION

[0039] Further advantages, characteristics, and features of the present invention will become apparent in the following detailed description of the exemplary embodiments. However, the invention is not limited to these exemplary embodiments.

[0040] Figure 1 shows a contrast line grid 1 according to the invention, as it can be used in a method for the automated determination of a geometry or for the detection of deformations of a surface using a corresponding algorithm or evaluation software. The corresponding contrast line grid 1 is applied to the surface to be detected, for example by printing, etching, laser engraving or the like.

[0041] In the embodiment shown in the following figures, the contrast line grid 1 is printed onto a vest 10 of a crash test dummy 9 in order to record the change in surface shape or the deformation of the vest 10 after the crash test. The vest 10 can also be a sensor vest equipped with sensors to additionally record, for example, the forces acting on the dummy. The contrast line grid 1 has a plurality of vertical contrast lines 2 and horizontal contrast lines 3, which define nodes 7 at their intersection points. The contrast lines 2, 3 and nodes 7 define grid cells 4. In the illustrated embodiment of the contrast line grid 1, the contrast lines intersect at right angles. Accordingly, the contrast line grid 1 is a rectangular grid, in this example even a square grid.However, the contrast line grid 1 can also be shaped differently, for example from triangular, hexagonal or polygonal grid cells 4. Accordingly, the contrast lines 2, 3 no longer intersect at the nodes 7 at a right angle, but at other angles, such as 60° or the like.

[0042] The contrast line grid 1 has an origin 6, which can be referred to as the first node 6. The origin 6 is the node of the contrast line grid 1 from which the search for further nodes 7 is to be started. In particular, the origin 6 can be the upper left corner of the contrast line grid 1. Accordingly, the first grid cell 5 with the first node 6 or origin 6 can be referred to as the origin cell, wherein the contrast line grid 1 is formed by the corresponding arrangement of adjacent grid cells 4. In the contrast line grid 1 according to the invention, a marker 8 in the form of a circle is arranged in the first grid cell 5 adjacent to the origin 6, and is arranged eccentrically to the center of the grid cell 5. The marker is arranged closer to the origin 6 than to the other nodes 7.This marker 8 serves to automatically detect the origin 6 of a contrast line grid 1 and consequently the further nodes 7 of the contrast line grid.

[0043] In the following embodiment of detecting the deformation of the surface of a sensor vest 10 of a crash test dummy 9, a contrast line grid 1, as shown in Figure 1, is printed onto the sensor vest 10 of the crash test dummy 9. If the geometry of the surface to be examined is known, this surface represents the reference surface relative to which the deformation of the sensor vest 10 after the crash test is to be detected. However, if the geometry of the surface to be examined is unknown, the reference geometry can first be determined. For this purpose, the crash test dummy 9 with the sensor vest 10 and the contrast line grid 1 can first be captured by a multitude of photographic images, which are shown, for example, in Figure 2, to determine the geometry.The geometry of the sensor vest 10 of the crash test dummy 9 can be captured by taking multiple images from different positions. For this purpose, the nodes 7 of the contrast line grid 1 are first captured and their position in space determined. Using the contrast line grid 1 according to the invention, the origin 6 can be easily captured, particularly automatically, by detecting the marker 8 via image recognition. Starting from the origin 6, the contrast lines 2, 3 can then be traced to the next nodes 7 and these can be determined. The size or radius of the marker 8, the distance of the marker 8 to the adjacent contrast lines 2, 3, and the width of the contrast lines 2, 3 can be used to capture the origin 6 and the further nodes 7.Instead of absolute values, relative values ​​can also be used, such as the asymmetry of the marker to the adjacent contrast lines 2, 3 passing through the origin 6, i.e., the ratio of the distances of the marker 8 to the adjacent contrast lines 2, 3. Furthermore, contrast information such as light to dark or vice versa can be included in the evaluation. The corresponding data can be entered into the evaluation software by the user, for example, via a dialog box in a graphical user interface. Alternatively, the evaluation software can also determine these parameters itself from the images and evaluate them iteratively by modifying the parameters and comparing the acquired data. The successive determination of the further nodes 7 of the contrast line grid 1 can be carried out according to the known number of horizontal and vertical contrast lines.However, it is advantageous to search for nodes 7 using successive procedural steps until no further nodes 7 can be found.

[0044] Figure 3 shows an image of the crash test dummy 9 with sensor vest 10 and a contrast line grid 1 printed on it, taken from the majority of the images shown in Figure 2. After capturing or providing the original geometry of the sensor vest 10 of the crash test dummy 9, a crash test can be performed in which the sensor vest 10 is deformed. After the surface of the sensor vest 10 of the crash test dummy has been deformed by the crash test, the deformed geometry can now be captured by several images of the crash test dummy 9 from different positions.

[0045] Figure 4 shows an enlarged view of the origin 6 of a contrast line grid 1 with marker 8 near the origin 6. In the evaluation software window shown in Figure 4, marker 8 is visible at the origin 6, and the parameters of the circle radius of marker 8 and the width of the contrast lines 2, 3 are adjusted by moving the sliders. Additionally, a further slider ("Offset") can be used to adjust any asymmetry in the distances between marker 8 and the contrast lines 2, 3. This allows the evaluation software to identify and detect the origin 6. These parameters are adjustable relative to each other. Furthermore, the user can define the pattern contrast (light to dark or vice versa).

[0046] Figure 5 shows another illustration corresponding to the illustration of Figure 4 with a different contrast line grid 1, so that the parameters are chosen accordingly.

[0047] Figure 6 shows a representation of an image with a highly deformed area after complete evaluation, i.e., after determining the geometry or deformation by comparison with the reference geometry. In Figure 6, the corresponding surface normals at the nodes 7, which characterize the geometry, are shown. Furthermore, the evaluation shows vectors from the respective nodes 7 directed in the direction of the intersecting line segments, thus providing a comprehensive characterization of the geometry.

[0048] Although the present invention has been described in detail with reference to the exemplary embodiments, it is obvious to those skilled in the art that the invention is not limited to these exemplary embodiments, but rather that modifications are possible in such a way that individual features can be omitted or different combinations of features can be implemented without departing from the scope of protection of the appended claims. In particular, the present disclosure includes all combinations of the individual features shown in the various exemplary embodiments, so that individual features that are only described in connection with one exemplary embodiment can also be used in other exemplary embodiments or in combinations of individual features not explicitly shown.

[0049] REFERENCE MARK LIST

[0050] 1 Contrast line grid

[0051] 2 vertical contrast line 3 horizontal contrast line

[0052] 4 grid cells

[0053] 5 first grid cell at the origin

[0054] 6 first node or origin

[0055] 7 nodes 8 markers

[0056] 9 Crash test dummy

[0057] 10 Vest

Claims

PATENT CLAIMS 1. Method for automated image-based detection of nodes (7) of a contrast line grid (1) with a plurality of nodes (7) which are connected to each other via contrast lines (2,3), wherein the contrast line grid (1) has a first node (6) as its origin, characterized in that a marker (8) is arranged adjacent to the node at the origin (6) of the grid, which can be detected by image recognition and serves to identify the origin.

2. Method according to claim 1, characterized in that the marking (8) is a circle or an ellipse or a square.

3. Method according to one of the preceding claims, characterized in that the size of the marking (8), in particular the radius of a circular marking (8) or at least one radius of an elliptical marking (8), or the side length of a square marking (8), and / or one or more distances of the marking to the contrast lines (2, 3) through the origin (6) of the grid and / or an asymmetry of the distances of the marking (8) to the contrast lines (2, 3) and / or the width of one or more contrast lines (2, 3) are automatically detected or provided and are used to determine nodes (7) adjacent to the origin (6) and / or other detected nodes (7).

4. Method according to one of the preceding claims, characterized in that starting from the origin (6) of the contrast line grid (1) the respective adjacent nodes (7) are successively detected, in particular coordinates of the nodes (7).

5. The method of claim 4, characterized in that the adjacent nodes are detected until a predetermined number of The nodes have been recorded, or no further nodes can be found.

6. Method according to claim 4 or 5, characterized in that the detected node points (7) are assigned to a reference grid.

7. Method according to one of the preceding claims, characterized in that the contrast line grid (1) is a rectangular grid, in particular with square grid cells (4).

8. Method for recording the surface geometry of an object (10) and in particular changes in the surface geometry of an object, wherein a contrast line grid (1) is applied to the surface of the object and at least several images of the surface of the object with the contrast line grid are taken from several different positions to determine the geometry, wherein the images of the surface are evaluated automatically, characterized in that Nodes (7) of the contrast line grid (1) are automatically detected, in particular according to a method according to one of the preceding claims.

9. Method according to claim 8, characterized in that, in order to detect a deformation of the surface, the contrast line grid (1) is applied to or arranged on the undeformed surface and at least of the deformed surface of the object (10) with the contrast line grid (1) several images are taken from several different positions to determine the geometry and the contrast line grid (1) applied to the surface of the undeformed object corresponds to a reference grid.

10. Method according to one of claims 8 or 9, characterized in that coordinates of the node points (7) and / or are derived from the various images. Normal vectors on the surface at the nodes and / or vectors in the direction of the intersecting contrast lines are determined.

Images