Method for manufacturing image handicraft by using waste logs

Abstract

The invention provides a method for manufacturing an image handicraft by using waste logs. The method comprises the following steps: (1) collecting the waste logs and transversely cutting the waste logs to obtain log slices; (2) the round wood slices are sequentially subjected to coarse grinding, fine grinding and polishing treatment; (3) the polished round wood slices are subjected to resin dipping treatment; (4) inputting the original image data into an artificial intelligence image generation system, and generating an adaptive image picture based on the surface feature data of the log slices; and (5) transferring the generated image picture to any surface of the log slice impregnated with the resin, and carrying out dust-free airing treatment to obtain the image handicraft. A large amount of collected existing work data are analyzed through the fruit tree pruning waste logs and the artificial intelligence using algorithm, and the problem that the fruit tree pruning waste logs are single in firewood processing is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of resource utilization of waste logs from fruit tree pruning, and particularly relates to a method for making image handicrafts from waste logs. Background Art

[0002] During the process of fruit tree planting, in order to optimize the tree structure, improve the lighting conditions, enhance the photosynthetic efficiency of leaves and increase the nutrient accumulation, it is necessary to regularly prune the growth-crossed, overlapping, competing, overgrown, dead or pest-diseased dry branches, resulting in a large amount of waste logs. If these waste logs are not properly processed, they will not only hinder orchard management (such as affecting operations like weeding and fertilization), but also cause the breeding of pests and diseases due to long-term accumulation, affecting the orchard ecological environment.

[0003] Currently, there are many limitations in the treatment methods for such waste logs: The method of pulverizing and returning to the field is difficult to promote due to the dispersity of mountain orchards and the limitation of mechanization level; The method of burning as firewood has a sharp reduction in demand due to the upgrading of the rural energy structure; Regional fuel sales are less economical due to transportation costs and geographical limitations.

[0004] In the prior art, the reuse of waste wood mostly focuses on low-value-added fields (such as fuel, compost), and its artistic value has not been effectively explored. Although wood handicrafts (such as carving, ornaments) have been applied, traditional processes rely on manual design, are difficult to adapt to the natural texture characteristics of fruit tree log slices, and have high costs. Summary of the Invention

[0005] The present invention provides a method for making image handicrafts from waste logs, which is used to solve the problems in the prior art that waste logs are difficult to be resourcefully utilized and lack ornamental and collection values.

[0006] The present invention provides a method for making image handicrafts from waste logs, which is characterized by including the following steps: (1) Collect waste logs and cut them horizontally to obtain log slices; (2) Conduct rough grinding, fine grinding and polishing treatments on the log slices in sequence; (3) Conduct impregnation resin treatment on the polished log slices; (4) Input the original image data into an artificial intelligence image generation system, and generate an adapted image picture based on the surface feature data of the log slices; (5) Transfer the generated image picture to any surface of the log slices after impregnation resin treatment, and obtain the image handicraft after dust-free drying treatment.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: It solves the single treatment and disposal method of turning the waste round logs from fruit tree pruning into firewood, and adds a new way for comprehensive waste treatment with an innovative concept; by superimposing artificial intelligence on the waste round logs from fruit tree pruning and using algorithms to analyze a large amount of existing work data collected, new works are created on one side of the round log slices, and the aesthetic art left by nature is retained on the other side of the transverse slices of the round logs.

[0008] Further, in step (4), the artificial intelligence image generation system uses a trained generation model, and the generation model is obtained through the following training method: (4.1) Construct a training set containing the surface feature data of the round log slices and a data set of corresponding adapted images; (4.2) Adopt a neural network model including an encoder and a decoder; (4.3) Optimize the parameters of the generation model through a loss function, so that the generation model can generate adapted image pictures according to the input features.

[0009] Further, the surface feature data includes curvature data, texture distribution, and color features obtained through 3D scanning.

[0010] Further, the neural network model adopts a generative adversarial network or a diffusion model.

[0011] Further, the corresponding adapted images are obtained through the following method: Place the test round log slices in a standard lighting environment and take a reference image; The corresponding adapted images are obtained by at least 3 designers adjusting the reference image using image processing software.

[0012] Further, the original image data includes at least one of photos, texts, or graphics.

[0013] Further, the diameter of the waste round logs is 100 - 500 mm, and the thickness of the slices is 10 - 50 mm.

[0014] Further, in step (5), UV inkjet printing technology is adopted, and the ink composition is acrylate materials.

[0015] Further, the resin in step (3) is epoxy resin, and the impregnation time is 30 - 60 min.

[0016] Further, the surface roughness Ra of the polished round log slices in step (2) is not greater than 1.6 μm. Specific Embodiments

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will, in conjunction with the embodiments of the present invention, clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0018] The present invention provides a method for making image handicrafts from waste logs, comprising the following steps: (1) Collect waste logs and cut them horizontally to obtain log slices; (2) Successively perform rough grinding, fine grinding, and polishing on the log slices; (3) Perform impregnation resin treatment on the polished log slices; (4) Input the original image data into an artificial intelligence image generation system, and generate an adapted image picture based on the surface feature data of the log slices; (5) Transfer the generated image picture to any surface of the log slices after impregnation resin treatment, and obtain the image handicraft after dust-free drying treatment.

[0019] Specifically, waste logs can be collected by means of on-site collection and modern networks, etc., either proximally or remotely through logistics. The waste logs are preferably dry branches with bark and no hollow, and there is no length limit; use machinery such as band saws and circular saws to cut perpendicular to the tree trunk cross-section to obtain log slices. The horizontally cut log slices cut by the cutting machine can be sorted according to technical requirements and immediately placed in a moisture balance box to avoid the slices from being exposed to the sun and cracked due to wind.

[0020] Optionally, for the rough grinding treatment, use a angle grinder to grind once to remove saw marks and burrs, and then use a rough sandpaper grinding machine for rough grinding; the horizontally cut log slices after rough grinding are successively subjected to fine grinding treatment by a disc-type and belt-type grinder from a small grit number to a large grit number of sandpaper. For example, successively use sandpaper with 80 meshes, 180 meshes, and 400 meshes for grinding, and finally use alumina polishing paste and a wool wheel for fine polishing to achieve the best polishing effect. Performing impregnation resin treatment on the polished log slices can enable the resin to uniformly penetrate into the slice cavity tissue and the surface of the slices, generating polymers insoluble in water, which is beneficial to making the slice size stable, moisture-proof and anti-corrosion, and significantly improving the mechanical strength.

[0021] It can be understood that the prepared image handicrafts can also be packaged as finished products using anti-abrasion, impact-resistant, and compression-resistant packaging boxes.

[0022] The method for making video handicrafts from discarded logs provided by the present invention solves the single treatment and disposal method of turning discarded logs from fruit tree pruning into firewood, and adds a new way for comprehensive waste treatment with an innovative concept; by superimposing artificial intelligence on discarded logs from fruit tree pruning and using algorithms to analyze a large amount of existing work data collected, new works are created on one side of the log slices, and the other side of the horizontal log slices retains the aesthetic art left by nature.

[0023] Further, in step (4), the artificial intelligence image generation system adopts a trained generation model, and the generation model is trained in the following manner: (4.1) Construct a training set containing the surface feature data of log slices and a data set of corresponding adapted images; (4.2) Adopt a neural network model including an encoder and a decoder; (4.3) Optimize the parameters of the generation model through a loss function, so that the generation model can generate adapted video pictures according to the input features.

[0024] In a specific embodiment, the encoder adopts the ResNet50 architecture, freezes the first three convolutional layers after loading the ImageNet pre-trained weights. The decoder adopts the U-Net structure, the upsampling multiple is set to 4 times, and the skip connection is set between the third and fourth levels. The loss function adopts a weighted combination method, where the mean square error weight is 0.3, the perceptual loss weight based on VGG-19 is 0.7, and the Adam optimizer is used for training. The initial learning rate is set to 0.0002, and it decays by 10% every 20 epochs.

[0025] During the training process, a perfect monitoring mechanism needs to be set. After each round of training, use the validation set to calculate key metrics: including structural similarity (SSIM), peak signal-to-noise ratio (PSNR), and color difference (ΔE). When the SSIM improvement amplitude in 5 consecutive epochs is less than 0.5%, the early stopping mechanism is triggered. After training is completed, use the TensorRT tool to convert the model into the ONNX format for easy deployment to the production environment.

[0026] Among them, the surface feature data includes curvature data, texture distribution, and color features obtained through 3D scanning.

[0027] The curvature data reflects the surface geometric shape of the log slices, the texture distribution records the unique texture characteristics of the wood, and the color features contain the natural color information of the wood. The training set based on these surface feature data provides a comprehensive reference basis for subsequent image generation.

[0028] In a specific embodiment, a 3D scanning device of the EinScan Pro 2X model is used to scan the log slice sample. This device is equipped with a structured light module with an accuracy of 0.1 mm to accurately obtain the three-dimensional point cloud data of the wood surface. During scanning, the slice sample needs to be fixed on a rotating platform, and continuous scanning is carried out at 24 angles at intervals of 15°. Finally, a complete surface model is synthesized through software. At the same time, a Canon EOS 5D Mark IV single-lens reflex camera is used to take two-dimensional texture images of the slices under a D65 standard light source. The CloudCompare software is used to denoise and streamline the point cloud data, and a curvature matrix is generated through the Poisson reconstruction algorithm. A custom script is run in the MATLAB environment to calculate the texture gradient vectors in 8 directions, specifically using the Sobel operator for convolution operations with a kernel size of 5×5 pixels. The extraction of color features uses the OpenCV library in Python to calculate the 256-level histogram distributions of the RGB three channels respectively.

[0029] Optionally, the neural network model adopts a generative adversarial network (GAN) or a diffusion model (DiffusionModel).

[0030] Using the neural network model defined above as a generative model can better handle the complex features of the wood surface, and further make the output image adapt to the natural surface features of the log slice. In a specific embodiment, the generative model selects the Stable Diffusion 2.1 version, and when using it, the controlnet_condition intensity parameter is set to about 0.75.

[0031] Furthermore, the corresponding adapted image is obtained in the following way: placing the test log slice in a standard lighting environment and taking a reference image; and obtaining the corresponding adapted image by at least 3 designers adjusting based on the reference image using image processing software.

[0032] The image processing software uses Adobe Photoshop for post-adjustment, and the adjustment includes fine adjustment of three aspects: surface deformation, texture fusion, and color balance of the image. The standard lighting environment is a D65 light source with an illuminance of 1000±50 lux.

[0033] In a specific embodiment, the "Inflate" mode in the "Warp" tool in Adobe Photoshop software is used to correct the surface deformation, and the deformation amount is controlled within the range of 3-12% according to the curvature matrix data; texture fusion is achieved through the "Neutralize" option in the "Match Color" function, and the opacity is set to 70-90%; finally, the "Color Balance" tool is used to adjust the middle tones, and the color level offset is controlled within ±15.

[0034] The present invention does not limit the specific selection of the original image data, which may include at least one of photos, texts or graphics, and further enables the finally generated image handicrafts to meet the personalized needs of different users.

[0035] In a specific embodiment, the log diameter of the waste log is 100 - 500 mm, and the thickness of the sliced pieces is 10 - 50 mm. Step (5) adopts UV inkjet printing technology, and the ink composition is acrylate material. The resin described in step (3) is epoxy resin, and the impregnation time is 30 - 60 min. The surface roughness Ra of the polished log slices in step (2) is not greater than 1.6 μm.

[0036] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; if these modifications and variations fall within the scope of the claims of the present invention and their equivalent technologies, they should all be regarded as the protection scope of the present invention.

Claims

1. A method for making video handicrafts from waste logs, characterized in that, It includes the following steps: (1) Collect waste logs and cut them transversely to obtain log slices; (2) Conduct rough grinding, fine grinding, and polishing treatments on the log slices in sequence; (3) Conduct impregnation resin treatment on the polished log slices; (4) Input the original image data into an artificial intelligence image generation system, and generate an adapted image based on the surface feature data of the log slices; (5) Transfer the generated image to any surface of the log slice after impregnation resin, and obtain the image handicraft after dust-free drying treatment.

2. The method for manufacturing an imaging handicraft from waste logs according to claim 1, wherein In step (4), the artificial intelligence image generation system uses a trained generation model, and the generation model is trained through the following method: (4.1) Construct a training set containing the surface feature data of log slices and a data set of corresponding adapted images; (4.2) Use a neural network model including an encoder and a decoder; (4.3) Optimize the generation model parameters through a loss function, so that the generation model can generate an adapted image according to the input features.

3. The method for making an imaging handicraft from waste logs according to claim 2, characterized in that, The surface feature data includes curvature data, texture distribution, and color features obtained through 3D scanning.

4. The method for manufacturing an imaging handicraft from waste logs according to claim 2, wherein, The neural network model uses a generative adversarial network or a diffusion model.

5. The method for manufacturing an imaging handicraft from waste logs according to claim 2, characterized in that, The corresponding adapted image is obtained through the following method: Place the test log slice in a standard lighting environment and take a reference image; At least 3 designers adjust the reference image using image processing software to obtain the corresponding adapted image.

6. The method for making an imaging handicraft from waste logs according to claim 2, characterized in that, The original image data includes at least one of photos, texts, or graphics.

7. The method for making an image handicraft from waste logs according to any one of claims 1 to 6, characterized in that, The wood diameter of the waste log is 100 - 500 mm, and the thickness of the slice is 10 - 50 mm.

8. The method for manufacturing an imaging handicraft from waste logs according to any one of claims 1 to 6, characterized in that, Step (5) uses UV inkjet printing technology, and the ink composition is acrylate material.

9. The method for making an imaging handicraft from waste logs according to any one of claims 1 to 6, characterized in that, In step (3), the resin is epoxy resin, and the impregnation time is 30 - 60 min.

10. The method for manufacturing an image handicraft from waste logs according to any one of claims 1 to 6, characterized in that, In step (2), the surface roughness Ra of the polished log slice is not greater than 1.6 μm.