A 3D printing-based timber member surface crack repairing method

Abstract

This invention discloses a 3D printing-based method for repairing surface cracks in wooden components. The method first determines the moisture content of the wooden component to be repaired; then, it acquires multi-view images of the surface cracks; next, it filters the acquired images; then, it obtains an average image from the filtered images, forming the final crack image and corresponding crack width information; next, it constructs a crack morphology learning model based on the crack image and the crack width information in the image to predict the crack depth; then, it establishes a three-dimensional model of the crack and uses wood flour / lignin-modified polylactic acid as the printing material to print crack filling strips on a 3D printer; finally, it fills the cracks on the surface of the wooden component with the printed crack filling strips and applies a layer of white glue to the repaired area. This solution combines image processing technology and 3D printing technology to effectively solve the problems of low quality and efficiency in existing in-situ repair methods for surface cracks in wooden components.

Description

Technical Field

[0001] This invention relates to the field of building engineering technology, and more specifically, to a technique for repairing surface cracks in wooden components. Background Technology

[0002] Throughout their entire lifespan, timber structures are inevitably subjected to the continuous effects of the environment and natural disasters, which may cause varying degrees of damage to their components. Given the good repairability of timber structures, demolition and reconstruction are usually unnecessary after damage; instead, reinforcement and repair measures are sufficient.

[0003] Surveys of timber structures in some urban renewal projects have revealed that timber structural components are prone to various types of damage, primarily including component cracking, decay, and connection failure. Cracks in timber structures may be caused by insufficient drying of some wood during fabrication, with the surface drying faster than the interior. The uneven shrinkage of the wood fibers, especially between the inner and outer layers, leads to shrinkage and cracking after seasonal changes. Alternatively, prolonged exposure to loads can reduce the wood's strength, decreasing its tensile, compressive, bending, and shear resistance, thus causing cracking under external forces.

[0004] The reinforcement strategies vary depending on the width of the cracks in the wooden components. Taking a load-bearing wooden column as an example, if the crack width is less than 3 mm, it can be considered that the crack is caused by changes in the moisture content of the wood, i.e., a shrinkage crack. Generally, putty is applied in situ to repair the component, prevent the crack from spreading, extend the service life of the damaged component, and restore its normal function. If the crack width exceeds 3 mm, the surface of the load-bearing component is considered to be damaged and affects safety, requiring reinforcement. Generally, wooden strips are used for patching, and if the crack is greater than 30 mm, clamps should be added to secure it.

[0005] In addition, cracks may also appear in the components of ancient wooden structures such as dougong, gong, and ang. These cracks can usually be fixed with adhesive and then used again. However, because dougong components are numerous, complex in structure, and difficult to disassemble and repair, small damage to components is usually not disassembled and is repaired in place.

[0006] In practical experience with timber structure reinforcement projects, it has been found that cracks on the surface of timber components are generally wider than 3mm. The common method is to fill the cracks with thin wood veneers and then apply putty to the gaps. This method is labor-intensive and time-consuming, and it is difficult to find wood veneers that perfectly match the shape of the cracks, making it difficult to fill them completely. Furthermore, changes in environmental humidity can easily cause the reinforced timber components to crack again. Traditional repair methods cannot guarantee the quality and efficiency of timber structure reinforcement.

[0007] Therefore, how to provide a material suitable for repairing wooden structures, and how to propose innovative construction methods based on this material to solve the problem of low quality and efficiency in repairing surface cracks in wooden components, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a new solution for repairing surface cracks in wooden components, which addresses the problem of low quality and efficiency of existing in-situ repair techniques for surface cracks in wooden components.

[0009] To achieve the above objectives, the present invention provides the following technical solution.

[0010] A method for repairing surface cracks in wooden components based on 3D printing includes the following steps:

[0011] Step (1) Determine the moisture content of the wooden component to be repaired;

[0012] Step (2) Obtain multi-view images of the cracks on the surface of the wooden component to be repaired;

[0013] Step (3) Filter the image obtained in step (2) to remove small details at the boundary and connect small discontinuities in the image to obtain a smooth image at the boundary;

[0014] Step (4) Obtain the average image from the image filtered in step (3) and form the final crack image and the corresponding crack width information;

[0015] Step (5) Based on the crack image obtained in step (4) and the crack width information in the image, construct a crack morphology learning model to predict the crack depth.

[0016] Step (6) Based on the crack morphology obtained in step (5), a three-dimensional model of the crack is established, and the corresponding printing control parameters are formed using the three-dimensional model. Wood flour / lignin modified polylactic acid material is used as the printing material to print crack filling strips on a 3D printer.

[0017] Step (7) Fill the cracks on the surface of the wooden component with the crack filler strip formed in step (6) and apply a layer of white glue to the repaired area.

[0018] In some embodiments of the present invention, when obtaining multi-view images of the surface cracks of the wooden component to be repaired in step (2), the images are obtained from five perspectives: a fully overhead view, a north-facing overhead view, a east-facing overhead view, a south-facing overhead view, and a west-facing overhead view, with the shooting plane parallel to the surface of the component.

[0019] In some embodiments of the present invention, when the image is filtered in step (3), a BilateralFilters nonlinear filter is used to process the crack edges in the image. At the same time, the intensity of the crack edge pixels is represented by a weighted average of the brightness values ​​of the surrounding pixels based on a Gaussian distribution. The weights of the Euclidean distance of the pixels and the radiation difference in the pixel range are introduced to preserve the crack edges in the image and to smooth and reduce noise.

[0020] In some embodiments of the present invention, obtaining the average image in step (4) includes:

[0021] (4.1) Perform three-segment linear grayscale transformation on the obtained multi-view images respectively;

[0022] (4.2) Perform Fourier transform on the multi-view images after grayscale processing to correct each image to a completely top-down view;

[0023] (4.3) The multiple images processed in step (4.2) are superimposed together according to different weights using a superposition algorithm;

[0024] (4.4) Based on deep learning methods, grayscale recognition is performed on the cracks in the superimposed image to obtain the crack image;

[0025] (4.5) Edge detection processing is performed on the crack image obtained in step (4.4) to obtain the crack image edge gradient information, and the distance between the intersection of the normal of the crack boundary pixel and the crack boundary is determined as the crack width.

[0026] In some embodiments of the present invention, step (5) extracts the crack image obtained in step (4) and the morphological information such as the edge shape of the crack, wood texture, moisture content and depth direction in the image, and constructs a crack morphology learning model to predict the internal morphology of the crack through the crack image on the surface of the component.

[0027] In some embodiments of the present invention, in step (6), a three-dimensional model of the crack is established based on the internal and external morphology of the crack obtained in step (5), and the corresponding printing control parameters are formed by the three-dimensional model. In order to obtain better printing effect, the printing control parameters need to be appropriately adjusted and optimized for the three-dimensional model of the crack.

[0028] In some embodiments of the present invention, the repair method further includes applying a layer of white glue and wood powder mixture to the repaired area of ​​the component to make the repair material bond tightly to the wooden component, and applying a tinted water-based paint to the repaired area of ​​the component.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0030] (1) High repair quality. Traditional repair methods rely on the experience of carpenters to select wood chips that are close to the thickness of the crack and fill the crack with wood chips. Compared with traditional repair methods, the present invention uses image recognition and machine learning methods to obtain crack morphology with high accuracy. The crack filling strip obtained by 3D printing technology matches the crack shape better, which can make the repair material bond more tightly to the wooden component.

[0031] (2) It conforms to the principle of reversibility in the reinforcement and protection of ancient buildings. Even if it remains in its original state before and after the repair, the historical and cultural information it carries is completely preserved. The solution of this invention does not damage the wooden structure of the ancient building itself, so that the wooden structure of the ancient building can still be reprocessed after the repair. When the repair materials and repair technology are more advanced, it can still be better repaired and reinforced.

[0032] (3) Fast construction speed. Traditional repair methods use wood chips to fill cracks on the surface of wooden components, which requires finding wood chips with a thickness close to the crack width, resulting in a slow construction speed. Compared with traditional repair methods, the present invention only requires taking 5 photos, which are automatically transmitted to the computer's graphics processing program to automatically generate 3D printing files. No manual operation is required, and the whole process takes no more than 1 minute, which can greatly shorten the material selection time. At the same time, the repair material is tightly bonded to the wooden component, and there is no need to consider adding or reducing the filling material, so the construction speed is faster.

[0033] (4) Good durability. The 3D printed repair material in this solution has high strength and strong weather resistance, and the bond between it and the wooden components is tighter. The repair material is a lignin-modified polylactic acid composite material, which has good interfacial performance with the wood of the wooden components. Applying a layer of white latex to the repair area makes the bond between the repair material and the wooden components tighter, reducing the risk of water seepage caused by interface cracking in the later stage. Applying tinted water-based paint to the repair area further reduces the chance of the wooden components and repair materials coming into contact with the external environment, thus improving the durability of the repaired wooden components.

[0034] (5) Reduce manual experience-based operations. The present invention uses a combination of computer vision image processing technology and 3D printing technology to obtain crack filling strips, replacing manual selection of repair pieces for wooden components, thus eliminating the reliance on professional workers for repairing wooden components. Attached Figure Description

[0035] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0036] Figure 1 This is a flowchart illustrating the surface crack repair process for wooden components based on 3D printing technology in an example of the present invention.

[0037] Figure 2 This is a schematic diagram of a crack on the surface of a wooden component in an example of the present invention.

[0038] In the image: 1-Wooden component to be repaired; 2-Crack on the surface of the wooden component; 31-Southern overhead view; 32-Western overhead view; 33-Complete overhead view; 34-Eastern overhead view; 35-Northern overhead view; 4-Shooting plane; 5-Sampling camera. Detailed Implementation

[0039] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below with reference to specific illustrations.

[0040] To address the problems faced by existing methods for repairing surface cracks in wooden structures, this invention organically integrates image recognition processing and 3D printing control, and further utilizes lignin-modified polylactic acid composite material as the repair material. This results in a new in-situ repair solution for surface cracks in wooden components, which not only improves repair quality but also significantly increases repair efficiency.

[0041] Specifically, this invention provides a method for repairing surface cracks in wooden components based on 3D printing. This method mainly includes the following steps:

[0042] Step (1) Determine the moisture content of the wooden component to be repaired;

[0043] This step involves determining the moisture content of the wooden component to be repaired, thereby determining the appropriate time to repair the cracks and preventing the repaired area from cracking again due to decreased ambient humidity.

[0044] Step (2) Obtain multi-view images of the cracks on the surface of the wooden component to be repaired;

[0045] To reduce sampling errors and avoid incomplete image information acquisition from a single angle, this solution innovatively acquires multi-view images by taking 24-bit true-color images of the surface of wooden components using a camera.

[0046] Step (3) Filter the multi-view image obtained in step (2) to remove small details at the boundary and connect small discontinuities in the image to obtain a smooth image at the boundary.

[0047] Step (4) Perform trapezoidal correction on the multi-view images filtered in step (3) to process them into images from a completely top-down view. Then, perform weighted averaging on the five corrected images to obtain a fully informative average crack image and the corresponding crack width information.

[0048] Step (5) Extract the crack image obtained in step (4) and the morphological information such as the edge shape of the crack, wood texture, moisture content and depth direction in the image, construct the crack morphology learning model, and predict the internal morphology of the crack through the crack image on the surface of the component.

[0049] Step (6) Based on the crack morphology obtained in step (5), a three-dimensional model of the crack is established, and the corresponding printing control parameters are formed by the three-dimensional model. In order to obtain better printing effect, the printing control parameters need to be appropriately adjusted and optimized for the three-dimensional model of the crack, and wood flour / lignin modified polylactic acid material is used as the printing material to print crack filling strips on the 3D printer.

[0050] Step (7) Fill the cracks on the surface of the wooden component with the crack filling strip formed in step (6), and scrape a layer of white glue and wood powder mixture at the repair site to make the repair material bond tightly with the wooden component.

[0051] The following details the implementation process and corresponding technical features of the present invention.

[0052] In some embodiments of the present invention, in step (1), the moisture content of the wooden component is measured using an inductive wood moisture meter. When the moisture content of the wooden component is between 9% and 15%, it is determined that the crack repair work of the wooden component can be carried out to prevent the repaired part of the wooden component from cracking again due to the decrease of ambient humidity.

[0053] In some embodiments of the present invention, when obtaining multi-view images of cracks on the surface of the wooden component to be repaired in step (2), the crack images on the surface of the wooden component are captured by a corresponding camera component to achieve accurate sampling of the surface of the wooden component.

[0054] like Figure 2 As shown, when taking specific photographic samples of the surface crack 2 on the wooden component 1 to be repaired, the shooting plane 4 of the sampling camera 5 is set to be parallel to the surface of the wooden component 1 to be repaired. In this state, five surface crack images of the wooden component surface crack 2 from different angles are accurately sampled according to the following angles: top view 33, north-south top view 35, east-south top view 34, south-south top view 31, and west-south top view 32.

[0055] Specifically, when shooting multi-angle images, the camera is set on the shooting plane 4. When shooting a completely overhead view 33, the camera is placed at the center point of the shooting plane 4. When shooting a north-south overhead view 35, the camera is placed 5 cm north of the center point of the shooting plane 4. When shooting an east-south overhead view 34, the camera is placed 5 cm east of the center point of the shooting plane 4. When shooting a south-south overhead view 31, the camera is placed 5 cm south of the center point of the shooting plane 4. When shooting a west-south overhead view 32, the camera is placed 5 cm west of the center point of the shooting plane 4. Under each shooting angle, the camera's field of vision is at the center of the crack in the wooden component.

[0056] In some embodiments of the present invention, in step (3), image filtering processing is performed on the next five surface crack images obtained from different perspectives obtained by shooting and sampling.

[0057] Unlike conventional linear filtering algorithms, this invention employs Bilateral Filters to improve the filtering effect. Furthermore, it uses a weighted average of the brightness values ​​of surrounding pixels based on a Gaussian distribution to represent the intensity of pixels at the crack edge. At the same time, it introduces weights for the Euclidean distance between pixels and the radiation difference in the pixel range domain, so that the crack edge in the image can be preserved and the noise can be reduced smoothly.

[0058] When performing image filtering, the weighted average method based on Gaussian distribution is first used to calculate the weight of each pixel based on the spatial proximity of each pixel to the center point. Then, the weight of each pixel is calculated based on the similarity of the pixel values. Multiplying these two values ​​gives the optimized weight of each point. To speed up the calculation, the Raised cosines function is used to approximate the Gaussian range function. Then, the optimized weights are convolved with the image to achieve the effect of edge preservation and noise reduction.

[0059] In some embodiments of the present invention, when obtaining the average image of the image after the filtering process in step (3) in step (4) and forming the final crack image and the corresponding crack width information, the following steps are taken:

[0060] (4.1) For the five surface crack images from different perspectives after filtering, three-segment linear grayscale transformation was used to process the images to improve the contrast of each crack image and enhance the details of the crack.

[0061] (4.2) After grayscale processing of the five surface crack images from different perspectives in step (4.1), Fourier transform is performed on each attached image to correct the image to an ideal perspective. In this step, the grayscale image is first read in, the image is binarized to show the image boundary, then opening and closing operations are performed to remove burrs and noise, then the vertex positions of the image contour are identified, and finally perspective correction is performed to process the image into a completely top-down perspective.

[0062] (4.3) For the five surface crack images with different perspectives after the perspective correction process in step (4.2), the image overlay algorithm is used to overlay the five images together with different weights to obtain the complete average crack image and the corresponding crack width information.

[0063] (4.4) The crack images obtained, along with the edge shape of the cracks, wood texture, moisture content, and depth direction in the images, are used to construct a crack morphology learning model. The cracks are then identified by grayscale to finally obtain crack images. Based on deep learning, the internal morphology of the cracks is predicted from the surface crack images of the components.

[0064] (4.5) For the crack image obtained in step (4.4), the Sobel operator is used to obtain the crack image edge gradient information, and then the distance between the intersection of the normal of the crack boundary pixel and the crack boundary is used as the crack width.

[0065] In some embodiments of the present invention, a crack morphology learning model is constructed in step (5), and an integrated CNN model is established for crack depth prediction when predicting crack depth. Here, the integrated CNN includes a convolutional layer feature extraction module and an extreme gradient boosting (XGBoost) regression module.

[0066] In some embodiments of the present invention, when establishing a three-dimensional model of the crack in step (6), a crack morphology learning model is constructed in advance based on the obtained crack image and the morphological information such as the edge shape of the crack, wood texture, moisture content and depth direction in the image, so that the internal morphology of the crack can be predicted by the crack image on the surface of the component.

[0067] Furthermore, the wood flour / lignin modified polylactic acid material uses polylactic acid as the main material, modified wood flour or lignin fiber to provide reinforcement and toughening properties, KH550 silane coupling agent to improve the tensile strength, melt flow rate and compatibility of the material, starch glue as a binder, and silicone oil as an auxiliary agent.

[0068] Furthermore, the wood flour / lignin-modified polylactic acid material here comprises the following raw materials in parts by weight:

[0069] Silane-modified wood flour: 1-10 parts;

[0070] Polylactic acid masterbatch: 50-100 parts;

[0071] Silane coupling agent ethanol solution: 2~20 parts;

[0072] Silicone oil: 1-5 parts.

[0073] Further, step (6) establishes a three-dimensional model of the crack based on the internal and external morphology of the crack obtained in step (5), and uses the three-dimensional model to form corresponding printing control parameters. In order to obtain better printing effect, it is necessary to adjust and optimize the printing control parameters appropriately for the three-dimensional model of the crack, and use wood flour / lignin modified polylactic acid material as printing material to print crack filling strips on a 3D printer.

[0074] In some embodiments of the present invention, in step (7), a mixture of white latex and wood powder is scraped onto the repair site of the component. In order to reduce the color difference between the repair material and the wooden component and increase the interface bonding effect, wood powder with a fineness modulus of 200 mesh and white latex are mixed in a volume ratio of 1:1.

[0075] In some embodiments of the present invention, after applying a layer of white latex, a tinted water-based paint is applied to the repaired area of ​​the component. This not only improves the durability of the wooden component but also compensates for color differences between the repair material and the wooden component. Furthermore, for environmental and health reasons, the selected tinted water-based paint must meet national environmental protection requirements.

[0076] The following application examples further illustrate the implementation process of the 3D printing-based method for repairing surface cracks in wooden components provided by this invention.

[0077] Combination Figure 1 As shown, the repair process for surface cracks in wooden structures using the 3D printing-based method for repairing surface cracks in wooden components, as presented in this invention, is as follows:

[0078] S1: First, conduct on-site inspection and testing of the wooden structure to be repaired.

[0079] (a) On-site inspection of timber structures: When inspecting and reinforcing timber structures, an on-site inspection of the timber structure should be conducted first to determine the damage status of each timber component.

[0080] Specifically, it is necessary to test and assess the stress patterns, wood species, dimensions, quantity, and residual mechanical properties of the components. When it is determined that the structure and components have no safety issues, meet the current stress conditions, and only require localized repairs to improve their durability, the repair method provided in this invention can be adopted.

[0081] Furthermore, before any repair work is carried out, the moisture content of the wooden components should be measured. Repairs can only be performed if the moisture content is between 9% and 15% to prevent the internal moisture from not evaporating easily and causing internal rot.

[0082] (b) When the surface crack width of the load-bearing wooden column is 3 mm to 30 mm, or when the surface crack width of the load-bearing wooden beam is 3 mm to 30 mm, it is determined that the surface of the load-bearing component is damaged and affects safety, and the method of the present invention is used for reinforcement treatment.

[0083] S2: After completing the on-site survey and testing, use a professional camera or a mobile phone with a fixed bracket to accurately sample the surface of the wooden components.

[0084] Specifically, the shooting plane is required to be parallel to the surface of the component, and precise sampling is required to collect 5 images from different perspectives of cracks on the surface of the wooden structure, including a full overhead view, a north-south overhead view, an east-south overhead view, a south-south overhead view, and a west-south overhead view.

[0085] S3: Bilateral Filters are used to filter the five images acquired from different viewpoints. Specifically, the intensity of the crack edge pixels is represented by a weighted average of the brightness values ​​of surrounding pixels based on a Gaussian distribution, thus preserving the crack edges in the image while smoothing and reducing noise.

[0086] S4: The image is processed by three-segment linear grayscale transformation to improve the contrast of the crack image. After the photo is grayscaled, the image is corrected to the ideal viewing angle. Then, the five images are superimposed together according to different weights using an image overlay algorithm.

[0087] S5: Build an integrated CNN model for crack depth prediction.

[0088] S6: Based on the crack morphology obtained from image recognition and deep learning, a three-dimensional model of the crack is built in the computer, and the .obj or .stl file is exported. The crack filling strip is then printed on a 3D printer using wood flour / lignin modified polylactic acid material.

[0089] S7: Fill the cracks on the surface of the wooden component with the printed crack filler strip, then apply a layer of white glue to the repaired area to ensure a tight bond between the repair material and the wooden component. Apply a tinted water-based paint to the repaired area to improve the durability of the wooden component and to compensate for any color difference between the repair material and the wooden component.

[0090] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for repairing surface cracks in wooden components based on 3D printing, characterized in that, Includes the following steps: Step (1) Determine the moisture content of the wooden component to be repaired; Step (2) Obtain multi-view images of the cracks on the surface of the wooden component to be repaired; Step (3) Filter the image obtained in step (2) to remove small details at the boundary and connect small discontinuities in the image to obtain a smooth image at the boundary; Step (4) Obtain the average image from the image filtered in step (3) and form the final crack image and the corresponding crack width information; When obtaining the average image in step (4), the following steps are included: (4.1) Perform three-segment linear grayscale transformation on the obtained multi-view images respectively; (4.2) Perform Fourier transform on the multi-view images after grayscale processing to correct each image to a completely top-down view; (4.3) The multiple images processed in step (4.2) are superimposed together according to different weights using a superposition algorithm; (4.4) Based on deep learning methods, grayscale recognition is performed on the cracks in the superimposed image to obtain the crack image; (4.5) Perform edge detection processing on the crack image obtained in step (4.4) to obtain the crack image edge gradient information, and determine the distance between the intersection of the normal of the crack boundary pixel and the crack boundary as the crack width. Step (5) Based on the crack image obtained in step (4) and the crack width information in the image, construct a crack morphology learning model to predict the crack depth. Step (6) Based on the crack morphology obtained in step (5), a three-dimensional model of the crack is established, and the corresponding printing control parameters are formed using the three-dimensional model. Wood flour / lignin modified polylactic acid material is used as the printing material to print crack filling strips on a 3D printer. Step (7) Fill the cracks on the surface of the wooden component with the crack filler strip formed in step (6) and apply a layer of white glue to the repaired area.

2. The method for repairing surface cracks in wooden components based on 3D printing according to claim 1, characterized in that, In step (2), when acquiring multi-view images of the cracks on the surface of the wooden component to be repaired, the images are acquired from five perspectives: a top-down view, a north-south top-down view, a east-south top-down view, a south-south top-down view, and a west-south top-down view, with the shooting plane parallel to the surface of the component.

3. The method for repairing surface cracks in wooden components based on 3D printing according to claim 1, characterized in that, In step (3), when the image is filtered, the intensity of the crack edge pixels is represented by the weighted average of the brightness values ​​of the surrounding pixels based on the Gaussian distribution, while the weights of the Euclidean distance of the pixels and the radiation difference in the pixel range are introduced.

4. The method for repairing surface cracks in wooden components based on 3D printing according to claim 1, characterized in that, In step (5), the crack image obtained in step (4) and the edge shape, wood texture, moisture content and depth direction morphology information of the crack in the image are extracted to construct a crack morphology learning model, and the internal morphology of the crack is predicted by the crack image on the surface of the component.

5. The method for repairing surface cracks in wooden components based on 3D printing according to claim 1, characterized in that, The repair method includes applying a layer of white glue and wood powder mixture to the repair area to ensure a tight bond between the repair material and the wooden component, and applying a tinted water-based paint to the repair area.

Images