Non-contact detection method for loaded large deformation of buffer honeycomb structure for vehicle

Abstract

The invention relates to the technical field of honeycomb structure detection, in particular to a vehicle buffer honeycomb structure loaded large deformation non-contact detection method, which comprises the following steps of: respectively assigning initial seed branch points in a reference image and a deformation image, indicating cell wall expansion directions, respectively carrying out adjacent branch point matching by utilizing three cell walls of each branch point, and finally obtaining the positions of all branch points in the ROI in all the images. And on the basis of the positions of the branch points in the reference image and the deformation image, calculating the displacement of the branch points by using camera calibration parameters to obtain the deformation trend of the honeycomb structure in the three-dimensional space after the honeycomb structure is loaded, and obtaining the deformation trend of the cell wall through the refined cell wall between the branch points. The method is suitable for full-field non-contact deformation detection requirements of the honeycomb structure; and the detection of large contact deformation of the cell wall of the honeycomb structure under load can be met.

Description

Technical Field

[0001] The present invention relates to the technical field of honeycomb structure detection, and particularly to a non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load. Background Art

[0002] In the manufacturing and application process of honeycomb materials, their in-plane deformation characteristics can reflect the mechanical behavior of honeycomb materials under load. As an important means of protecting vehicle collision safety, how to design the shape and size parameters of honeycomb structures and verify them through experiments is crucial; it is difficult to obtain the quantitative deformation of honeycomb structures and the deformation trend under different loading conditions by using the direct observation method. In addition, the traditional method can only obtain the overall trend of the deformation of honeycomb structures under load, and it is difficult to obtain the accurate deformation amount for local deformation, especially in the area with large deformation. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to propose a non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load to solve the problem that the current detection method is difficult to obtain the accurate deformation amount for the area with large deformation.

[0004] Based on the above purpose, the present invention provides a non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load. The execution end of the detection method includes a loading system control end, a fixture, two industrial cameras, a light source, and a measurement system control end;

[0005] The detection method includes:

[0006] S1: Pretreat the surface of the honeycomb structure, cross-color the cell walls of the honeycomb structure with six different colors, use white for the background, and evenly distribute the three colors at the branch points where the cell walls cross;

[0007] S2: Clamp the honeycomb structure through the fixture and apply a predetermined initial load to make the upper and lower surfaces of the honeycomb structure fit the surface of the fixture, and obtain the lighting parameters and the internal and external parameters of the industrial cameras;

[0008] S3: Load the fixture in a stepped loading mode through the loading system control end. Each time a load is applied, the two industrial cameras simultaneously collect images using external trigger hard synchronization until the honeycomb structure reaches the dense stage and stops loading and image collection;

[0009] S4: Filter the collected images, convert the images from the RGB color gamut to the Lab color gamut, and obtain the color of each pixel of the entire image belonging to the Lab color gamut;

[0010] S5: Extract the pixels of different colors belonging to the Lab color gamut in S4 to form 7 new binary images. First, perform morphological closing operations on the binary images, and then perform morphological erosion processing. Synthesize the 7 images after morphological erosion processing to form a new binary image. At this time, the 3 branch endpoints at the branch points of the new binary image are not connected. Connect them pairwise by searching for the nearest endpoints, and then obtain the centers of the 3 line segments, which are the branch points and are connected to the 3 branches to form a complete honeycomb structure;

[0011] S6: Select the ROI where the honeycomb structure is located, and use the Zhang-Suen cell wall thinning algorithm to thin the cell wall of the honeycomb structure to obtain a single-pixel-width cell wall. The intersection of the three single-pixel cell walls is the branch point, which serves as the carrier for expressing the displacement information of the honeycomb structure;

[0012] S7: Specify the initial seed branch points in the reference image and the deformed image respectively, and indicate the cell wall expansion direction. Then, use the three cell walls of each branch point to perform adjacent branch point matching respectively, and finally obtain the positions of all branch points in the ROI in all images;

[0013] S8: Based on the positions of each branch point in the reference image and the deformed image, use the camera calibration parameters to calculate the branch point displacement, obtain the deformation trend of the honeycomb structure in the three-dimensional space after being loaded, and obtain the deformation trend of the cell wall from the thinned cell walls between the branch points.

[0014] Optionally, the preprocessing of the honeycomb structure surface includes removing burrs and bifurcations on the thin cell walls of the honeycomb structure.

[0015] Optionally, the six different colors are black, red, green, blue, yellow, and purple.

[0016] Optionally, the filtering process includes filtering the acquired image using the non-local means filtering algorithm (NL-means).

[0017] Optionally, the conversion of the image from the RGB color gamut to the Lab color gamut to obtain the color of each pixel in the entire image belonging to the Lab color gamut includes:

[0018] Select a pixel area of a certain size for each color and the background color and coincide with the corresponding cell wall, and calculate the average value of the Lab color gamut space corresponding to each color, as shown in the left figure; Solve the Euclidean distance from the Lab space color gamut of each pixel in the entire image to the average value of the Lab color gamut of each color, and then obtain the color of each pixel in the entire image. Figure 2

[0019] ​When this detection method is in use, six different colors are used to cross-color the cell walls of the honeycomb structure. Pixels belonging to different colors in the Lab color gamut in S4 are extracted to form seven new binary images. First, morphological closing operations are performed on the images, and then morphological erosion processing is carried out to separate the cell walls that come into contact with each other after loading. The seven images after morphological processing are synthesized to form a binary image. At this time, the three branch endpoints at the branch points are not connected. By searching for the nearest endpoints, they are connected pairwise, and then the centers of the three line segments are obtained, which are the branch points and are connected to the three branches to form a complete honeycomb structure. The ROI where the honeycomb structure is located is selected, and the Zhang-Suen cell wall thinning algorithm is used to thin the cell walls of the honeycomb structure to obtain cell walls with a single-pixel width. The intersection points of the three single-pixel cell walls are the branch points, which serve as the carriers for expressing the displacement information of the honeycomb structure. Initial seed branch points are specified in the reference image and the deformed image respectively, and the cell wall expansion directions are indicated. Then, the three cell walls of each branch point are used to perform adjacent branch point matching respectively, and finally, the positions of all branch points in the ROI in all images are obtained. Based on the positions of each branch point in the reference image and the deformed image, the camera calibration parameters are used to calculate the displacement of the branch points, and the deformation trend of the honeycomb structure in the three-dimensional space after loading is obtained. The deformation trend of the cell walls can be obtained from the thinned cell walls between the branch points.

[0020] As can be seen from the above, the present invention is applicable to the full-field non-contact deformation detection requirements of honeycomb structures; it can meet the detection of large deformations in contact of honeycomb structure cell walls under load; based on the measured data, the deformation amount and deformation trend of the honeycomb structure are analyzed to assist the structural design of the honeycomb structure. The method of cell wall thinning and branch point reconstruction can avoid the problems of large measurement data volume and long processing time, and improve the detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only those of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 It is the topology diagram of the execution end of the embodiment of the present invention;

[0023] Figure 2 It is the schematic diagram of the colors of each pixel of the honeycomb structure in the embodiment of the present invention;

[0024] Figure 3 It is the complete cell wall diagram of the honeycomb structure before and after loading obtained in the embodiment of the present invention;

[0025] Figure 4 It is the result diagram of the three-dimensional diffusion matching of the branch points in the embodiment of the present invention;

[0026] Figure 5 Schematic diagram of the displacement of the branch point and the deformation of the cell wall of the honeycomb structure according to an embodiment of the present invention. Detailed implementation manners

[0027] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with specific embodiments.

[0028] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be the ordinary meanings understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar terms used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative position relationships, and when the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0029] As Figures 1 - 5 shown, a non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load, the execution end of the detection method includes a loading system control end, a fixture, two industrial cameras, a light source and a measurement system control end;

[0030] The detection method includes:

[0031] S1: Pretreat the surface of the honeycomb structure, cross-color the cell walls of the honeycomb structure with six different colors, use white as the background, and three colors are evenly distributed at the intersections of the cell walls, that is, at the branch points;

[0032] S2: Clamp the honeycomb structure through the fixture and apply a predetermined initial load to make the upper and lower surfaces of the honeycomb structure fit with the surface of the fixture, and obtain the illumination parameters and the internal and external parameters of the industrial cameras;

[0033] S3: Load the fixture according to the step loading mode through the loading system control end. Each time after loading, the two industrial cameras use external trigger hard synchronization to simultaneously collect images to avoid asynchronous image collection of the binocular cameras, and stop loading and image collection until the honeycomb structure is in the dense stage;

[0034] S4: Filter the acquired image to reduce the noise level; convert the image from the RGB color gamut to the Lab color gamut to obtain the color of each pixel in the entire image belonging to the Lab color gamut.

[0035] S5: Extract the pixels belonging to different colors in the Lab color gamut in S4 to form 7 new binary images. First, perform a morphological closing operation on the images, and then perform a morphological erosion process to separate the cell walls that are in contact with each other after loading; synthesize the 7 images after the morphological processing to form a binary image. At this time, the 3 branch endpoints at the branch points are not connected. Connect them pairwise by searching for the nearest endpoints, and then obtain the centers of the 3 line segments, which are the branch points, and connect them to the 3 branches to form a complete honeycomb structure.

[0036] S6: Select the ROI (region of interest) where the honeycomb structure is located, and use the Zhang-Suen cell wall thinning algorithm to thin the cell walls of the honeycomb structure to obtain a single-pixel-width cell wall. The intersection of the three single-pixel cell walls is the branch point, which serves as the carrier for expressing the displacement information of the honeycomb structure.

[0037] S7: Specify the initial seed branch points in the reference image (the image in the initial stage is used as the reference image) and the deformed image (the image deformed after loading) respectively, and indicate the cell wall expansion direction. Then, use the three cell walls of each branch point to perform adjacent branch point matching respectively, and finally obtain the positions of all branch points in the ROI in all images.

[0038] S8: Based on the positions of the respective branch points in the reference image and the deformed image, calculate the branch point displacement using the camera calibration parameters to obtain the deformation trend of the honeycomb structure in three-dimensional space after loading, and the deformation trend of the cell wall can be obtained from the thinned cell walls between the branch points.

[0039] When this detection method is used for detection, the cell walls of the honeycomb structure are cross-colored with six different colors, and the pixels belonging to different colors in the Lab color gamut in S4 are extracted to form seven new binary images. First, morphological closing operations are performed on the images, and then morphological erosion processing is carried out to separate the cell walls that come into contact with each other after loading. The seven images after morphological processing are synthesized to form a binary image. At this time, the three branch endpoints at the branch points are not connected. By searching for the nearest endpoints, they are connected in pairs, and then the centers of the three line segments are obtained, which are the branch points and are connected to the three branches to form a complete honeycomb structure. The ROI where the honeycomb structure is located is selected, and the Zhang-Suen cell wall thinning algorithm is used to thin the cell walls of the honeycomb structure to obtain cell walls with a single-pixel width. The intersection of the three single-pixel cell walls is the branch point, which serves as the carrier for expressing the displacement information of the honeycomb structure. Initial seed branch points are specified in the reference image and the deformed image respectively, and the cell wall expansion direction is indicated. Then, adjacent branch point matching is performed using the three cell walls of each branch point, and finally, the positions of all branch points in the ROI in all images are obtained. Based on the positions of the branch points in the reference image and the deformed image, the camera calibration parameters are used to calculate the displacement of the branch points, and the deformation trend of the honeycomb structure in the three-dimensional space after loading is obtained. The deformation trend of the cell walls can be obtained from the thinned cell walls between the branch points.

[0040] As can be seen from the above, the present invention is applicable to the full-field non-contact deformation detection requirements of honeycomb structures; it can meet the detection of large deformations in contact of honeycomb structure cell walls under load; based on the measured data, the deformation amount and deformation trend of the honeycomb structure are analyzed to assist the structural design of the honeycomb structure. The method of cell wall thinning and branch point reconstruction can avoid the problems of large amount of measurement data and long processing time, and improve the detection efficiency.

[0041] In some embodiments, the pretreatment of the honeycomb structure surface includes removing burrs and bifurcations on the thin cell walls of the honeycomb structure to prevent the influence of burrs and bifurcations on subsequent image processing.

[0042] In some embodiments, the six different colors can be black, red, green, blue, yellow, and purple, so that they can be distinguished more clearly.

[0043] In some embodiments, the filtering process includes filtering the acquired image using the non-local means filtering algorithm (NL-means).

[0044] In some embodiments, the conversion of the image from the RGB color gamut to the Lab color gamut to obtain the color of each pixel in the entire image belonging to the Lab color gamut includes:

[0045] Select a pixel area of a certain size for each color and background color (a total of 7 kinds) and coincide with the corresponding cell wall, and obtain the average value of the Lab color gamut space corresponding to each color, as Figure 2 shown in the left figure; solve the Euclidean distance from the Lab color gamut of each pixel in the entire image to the average value of the Lab color gamut of each color, and then obtain the color to which each pixel in the entire image belongs.

[0046] Those of ordinary skill in the art should understand that the discussion of any above embodiment is only exemplary and is not intended to imply that the scope of the present invention (including the claims) is limited to these examples; under the concept of the present invention, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the present invention as above, which are not provided in detail for the sake of brevity.

[0047] The present invention is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load, characterized in that: The execution end for executing the detection method includes a loading system control end, a fixture, two industrial cameras, a light source and a measurement system control end; The detection method comprises: S1: Pre-treat the surface of the honeycomb structure, use six different colors to cross-color the cell walls of the honeycomb structure, use white as the background, and evenly distribute the three colors at the intersection of the cell walls, i.e., the branching points; S2: Clamping the honeycomb structure with the fixture and applying a predetermined initial load so that the upper and lower surfaces of the honeycomb structure fit with the fixture surface, and obtaining illumination parameters and internal and external parameters of the industrial camera; S3: The fixture is loaded by the control end of the loading system in a step loading mode, and each time the fixture is loaded, two industrial cameras are hard-synchronized by external triggering to simultaneously collect images, and loading and image collection are stopped until the honeycomb structure is in a dense stage; S4: Filter the collected image, convert the image from the RGB color domain to the Lab color domain, and obtain the color of the Lab color domain to which each pixel of the entire image belongs; S5: Extract pixels of different colors in the Lab color domain in S4 to form 7 new binary images, and first perform morphological closing operation on the binary images, and then perform morphological corrosion processing, and synthesize the 7 images after morphological corrosion processing to form a new binary image. At this time, the three branch endpoints at the branch point of the new binary image are not connected. They are connected in pairs by searching for the nearest endpoints, and then the centers of the three line segments are obtained, which are the branch points, and connected with the three branches to form a complete honeycomb structure; S6: Select the ROI where the honeycomb structure is located, and use the Zhang-Suen cell wall thinning algorithm to thin the cell wall of the honeycomb structure to obtain a single-pixel width cell wall. The intersection of three single-pixel cell walls is a branch point, which serves as a carrier for the displacement information of the honeycomb structure. S7: specifying the initial seed branch points in the reference image and the deformed image respectively, and indicating the cell wall extension direction, and then using the three cell walls of each branch point to match adjacent branch points respectively, and finally obtaining the positions of all branch points in the ROI in all images; S8: Based on the positions of each branch point in the reference image and the deformed image, the branch point displacement is calculated using the camera calibration parameters to obtain the deformation trend of the honeycomb structure in three-dimensional space after being loaded, and the deformation trend of the cell wall is obtained from the refined cell wall between the branch points.

2. A non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load according to claim 1, characterized in that: The pretreatment of the surface of the honeycomb structure includes removing burrs and bifurcations on the thin walls of the honeycomb cells.

3. A non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load according to claim 1, characterized in that: The six different colors are black, red, green, blue, yellow and purple.

4. A non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load according to claim 1, characterized in that: The filtering process includes using a non-local mean filtering algorithm (NL-means) to perform filtering on the collected image.

5. The non-contact detection method for large deformation of a vehicle buffer honeycomb structure under load according to claim 1, characterized in that: The converting of the image from the RGB color gamut to the Lab color gamut to obtain the color of the Lab color gamut to which each pixel of the entire image belongs includes: For each color and background color, a pixel area of ​​a certain size is selected and overlapped with the corresponding cell wall, and the average value of the Lab color space corresponding to each color is obtained, as shown in the left figure of Figure 2; the Euclidean distance from the Lab space color space of each pixel in the entire image to the Lab color space mean of each color is solved, and then the color of each pixel in the entire image is obtained.

Images