Interactive electric shock wood painting device

Abstract

The invention provides an interactive electric shock wood painting device, and belongs to the technical field of electric shock wood painting.Wood texture parameters are obtained through a high-definition camera, a texture parameter matrix is established, wood surface humidity distribution is detected through a humidity sensor array, and a humidity parameter matrix is established; a carbonization key point matrix and a carbonization texture adjoint degree coefficient are calculated based on historical data, humidity parameters are input into a carbonization depth intelligent optimization model based on a UNet architecture for carbonization depth prediction, and the carbonization depth prediction value of each point is calculated through a carbonization prediction algorithm in combination with preset electric shock parameters; the algorithm comprises four stages of current density distribution calculation, temperature field analysis, carbonization reaction kinetic modeling and carbonization depth quantitative prediction, finally, a predicted carbonization image is generated and displayed on a liquid crystal display screen in an overlapped mode, and accurate prediction and real-time visual preview of the carbonization effect in the electric shock wood drawing process are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric shock wood painting, and in particular relates to an interactive electric shock wood painting device. Background Art

[0002] Electric shock wood painting is an artistic technique that uses high-voltage current to produce lightning-like fractal patterns on the surface of wood. By controlling the shock parameters, the technique achieves localized carbonization of the wood surface, creating artworks with unique aesthetic appeal. Traditional electric shock wood painting relies primarily on operator experience and trial and error to control the shock effect. Parameters such as voltage, current, and shock location are adjusted to achieve the desired carbonization pattern. This method is widely used in fields such as handicrafts, art, and decorative crafts. However, traditional electric shock wood painting has significant limitations. Operators cannot accurately predict the carbonization effect before performing the shock operation, and must rely on repeated trials and adjustments to approximate the ideal result. This not only increases material waste and time costs, but also makes it difficult to achieve the precise artistic design intent. Furthermore, factors such as wood grain distribution, moisture variations, and material differences significantly affect the carbonization effect, but traditional methods lack the ability to quantitatively analyze and predict these key parameters. A core issue that hinders traditional techniques is the lack of an effective carbonization prediction mechanism. This inability to provide a visual preview of the effect before the shock operation requires operators to rely blindly on their experience, resulting in a significant lack of controllability and precision in the artistic creation. Summary of the Invention

[0003] In view of this, the present invention provides an interactive electric shock wood painting device, which can solve the technical problem in the prior art that it is impossible to accurately predict the carbonization effect and provide a real-time visual preview during the electric shock wood painting process.

[0004] The present invention is implemented as follows: the present invention provides an interactive electric shock wood painting device, the top of which is fixed with a carbon steel electrode; the wooden board fixing platform is a rectangular steel frame structure, and the table top is provided with a wooden board positioning groove; the camera is installed above the wooden board fixing platform through a universal bracket; an electric shock painting prediction and display module is provided in the control chip, the wooden board surface image is obtained by the camera to extract the wooden board texture parameter matrix, the wooden board humidity parameter matrix is ​​detected by the humidity sensor array, the carbonization key point matrix is ​​calculated according to the wooden board texture parameter matrix and historical electric shock carbonization data, the shortest path algorithm is used to solve the current conduction path optimization problem and calculate the carbonization texture accompanying degree coefficient, the humidity parameter matrix is ​​input into the carbonization depth intelligent optimization model to predict the carbonization depth of each position and calculate the humidity carbonization accompanying degree coefficient, the preset electric shock position coordinates, electric shock angle, and current intensity are substituted into the carbonization prediction algorithm and combined with the texture carbonization accompanying change point matrix and the humidity carbonization accompanying change point matrix to calculate the carbonization depth prediction value, and a predicted carbonization image is generated based on the carbonization depth prediction value and superimposed on the real-time image of the wooden board on the liquid crystal display screen.

[0005] Among them, the current control module is equipped with a high-voltage power converter and a precision current regulator. The high-voltage power converter converts 220V AC power into 2000V~5000V DC high-voltage power, and the precision current regulator accurately adjusts the high-voltage current to the range of 0.1A~10A; the safety protection module is equipped with an overload protection relay, a short-circuit protection switch and a leakage protection circuit breaker.

[0006] Among them, the electric shock painting prediction and display module controls the camera to obtain real-time images of the wooden board surface, grayscales the image and extracts the texture features of the wooden board, and establishes a wooden board texture parameter matrix. The number of matrix rows is the number of image rows, the number of matrix columns is the number of image columns, and the matrix element value represents the texture density of the corresponding position.

[0007] Among them, the electric shock painting prediction and display module controls the humidity sensor array to collect humidity values ​​at 16 measurement points on the surface of the wooden board, establishes a humidity distribution matrix, and expands the humidity distribution matrix into a humidity parameter matrix of the same dimension as the texture parameter matrix through a bilinear interpolation algorithm.

[0008] The carbonization key point matrix is ​​calculated based on the wood board texture parameter matrix and historical electric shock carbonization data. Each element in the matrix represents the probability weight of carbonization reaction occurring at the corresponding position, and the weight value ranges from 0 to 1.

[0009] Among them, the shortest path algorithm converts the wooden board texture parameter matrix into a weighted graph structure, with each pixel point as a node of the graph. The edge weights between adjacent pixels are determined by the texture density difference and humidity difference. The Dijkstra algorithm is used to calculate the shortest current conduction path from the preset electric shock position to each edge point of the wooden board.

[0010] Among them, the carbonization texture accompanying degree coefficient is calculated based on the shortest path result and the wood board texture parameter matrix, which represents the degree of influence of the texture structure at each position on the carbonization process. At the same time, a texture carbonization accompanying change point matrix is ​​generated. The matrix element value of 1 represents a change point, and the value of 0 represents an ordinary point.

[0011] Among them, the carbonization depth intelligent optimization model is a skip connection model based on the UNet architecture, which consists of two main parts: the encoder and the decoder. The encoder is responsible for extracting the multi-scale features of the humidity distribution, and the decoder is responsible for reconstructing the carbonization depth prediction map. The encoder and the decoder transmit detailed information through skip connections.

[0012] Among them, the feature fusion weight parameters in the carbonization depth intelligent optimization model are dynamically adjusted according to the combined value of the current current intensity, wood board thickness, and preset electric shock angle through the weight adjustment function. The weight adjustment function sets the feature fusion weight parameters based on the comprehensive influencing factors using linear, quadratic or exponential interpolation methods.

[0013] Among them, the carbonization prediction algorithm is established based on the current conduction physical model and the wood carbonization chemical reaction mechanism. The core calculation process includes current density distribution calculation, temperature field analysis, carbonization reaction kinetics modeling and carbonization depth quantitative prediction.

[0014] Among them, the control chip is electrically connected to the robotic arm drive motor, carbon steel electrode, camera, LCD display, humidity sensor array, current control module, and safety protection module through a data bus to exchange data; the camera uses a high-definition camera.

[0015] Among them, the humidity sensor array includes 16 humidity sensors, arranged in a 4×4 matrix and embedded under the table top of a wooden board fixing table; the LCD display is installed on the front of the console; the robotic arm base is fixedly installed on one side of the wooden board fixing table; the robotic arm is connected to the robotic arm base through a joint connector.

[0016] Among them, the humidity carbonization accompanying coefficient quantifies the degree of influence of the humidity level of the wood board on the depth of electric shock carbonization. The higher the humidity, the larger the accompanying coefficient. At the same time, a humidity carbonization accompanying change point matrix is ​​generated to identify key positions that may cause sudden changes in carbonization effects due to local humidity differences.

[0017] Among them, the current density distribution calculation stage establishes a group of partial differential equations for the current distribution inside the wooden board based on Ohm's law and Kirchhoff's law. The current density distribution matrix of each point in the electric shock action area is obtained by solving it through the finite element method. The conductivity is calculated through a nonlinear mapping function based on the wooden board texture parameter matrix and the humidity parameter matrix.

[0018] Among them, the temperature field analysis stage establishes a temperature distribution model of the electric shock area based on the principle of Joule heating effect. The temperature is calculated using the heat conduction equation. The heat source intensity per unit volume is obtained by the dot product operation of the current density distribution matrix and the resistivity matrix. The thermal diffusion coefficient is dynamically adjusted through empirical formulas based on the wood type, wood board texture parameter matrix and humidity parameter matrix.

[0019] Among them, in the carbonization reaction kinetics modeling stage, a reaction rate model of wood pyrolysis and carbonization is established based on the Arrhenius equation. The frequency factor and activation energy are obtained through experimental calibration according to the wood species. The reaction mechanism function adopts a three-dimensional diffusion model. The carbonization conversion rate is calculated and updated within each time step through the numerical integration method.

[0020] Among them, the carbonization depth quantitative prediction stage converts the carbonization conversion rate matrix into a visual carbonization depth value. During the conversion, the theoretical maximum carbonization depth is calculated using an empirical formula based on the wood thickness and density. The texture correction coefficient is obtained by weighted calculation of the carbonization texture accompanying degree coefficient and the texture carbonization accompanying change point matrix. The humidity correction coefficient is obtained by weighted calculation of the humidity carbonization accompanying degree coefficient and the humidity carbonization accompanying change point matrix.

[0021] Among them, the predicted carbonization image generation algorithm includes four main steps: depth value mapping, texture rendering, light and shadow effect processing, and image synthesis. The depth value mapping uses a piecewise linear mapping function to convert the carbonization depth prediction value into a grayscale value. Texture rendering generates a realistic carbonization texture effect based on the wood board texture parameter matrix and the carbonization depth prediction value. The light and shadow effect processing simulates the lighting characteristics and shadow effects of the carbonized area based on the Phong lighting model. The image synthesis uses the Alpha blending algorithm to superimpose and fuse the rendered carbonization texture image with the real-time image of the wood board.

[0022] This invention integrates technologies such as wood texture recognition, moisture distribution detection, and carbonization physical modeling to achieve accurate prediction and real-time visualization of the carbonization effect during electric shock wood painting. This method establishes a complete mapping relationship from wood surface features to carbonization depth distribution, capable of generating high-fidelity preview images of the carbonization effect based on preset electric shock parameters. By constructing a carbonization prediction algorithm based on physical mechanisms, this invention effectively addresses the problem of blind parameter adjustment in traditional techniques. This algorithm comprehensively considers multiple physical processes such as current conduction, temperature field distribution, and carbonization reaction kinetics, accurately simulating the carbonization response characteristics of different wood species under various conditions. Furthermore, a deep learning optimization model is used to intelligently process key parameters such as moisture, significantly improving prediction accuracy and adaptability, enabling the operator to accurately determine the final effect before performing the electric shock. This invention fundamentally addresses the core technical issue of unpredictable carbonization effects in electric shock wood painting. By establishing a complete prediction and visualization system, it transforms traditional empirical operations into precise scientific control, achieving intelligent and precise artistic creation, and providing reliable technical support for the large-scale application of electric shock wood painting technology and the stable control of artistic quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 Schematic diagram of the structure of the device of the present invention.

[0024] Figure 2 Flowchart of the steps executed by the electric shock drawing prediction display module of the present invention.

[0025] Figure 3 This is a schematic structural diagram of the device in Example 2.

[0026] Figure 4 This is a comparison diagram of the current image and the predicted result image in Example 2, where sub-image (a) is the current image and (b) is the predicted result image.

[0027] In the accompanying drawings, the reference numerals are explained as follows: 1. Wooden board fixing platform; 2. Electrode; 3. Wooden board fixing clamp; 4. Wooden board; 5. Touch screen; 6. Robotic arm control switch; 7. Up and down control joystick; 8. Left and right control joystick; 9. UV protection screen; 10. Robotic arm; 11. Equipment status indicator light; 12. Equipment power switch; 13. Camera; 14. Drawing cover. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0029] like Figure 1FIG. 1 is a schematic diagram of the structure of an interactive electric shock wood painting device provided by the present invention, comprising: a robotic arm 10, a control console, a high-definition camera 13, a liquid crystal display 51, a wooden board fixing platform 1, a humidity sensor array 15, a current control module, a safety protection module, and a control chip. A carbon steel electrode is fixedly provided on the top of the robotic arm, and the robotic arm is connected to the robotic arm base via a joint connector. The robotic arm base is fixedly mounted on one side of the wooden board fixing platform. The wooden board fixing platform is a rectangular steel frame structure, and the table top is provided with wooden board positioning grooves, which are used to fix the position of the wooden board to be processed. The humidity sensor array includes 16 humidity sensors arranged in a 4×4 matrix and embedded under the wooden table, with a spacing of 50 mm between each humidity sensor; the high-definition camera is installed 300 mm above the wooden table through a universal bracket, with the camera lens facing the wooden table; the LCD screen is installed on the front of the console, and the screen size is 15.6 inches; an electronic control cabin is provided inside the console, and the control chip is installed in the electronic control cabin, and the control chip is connected to the robotic arm drive motor, the carbon steel electrode, the high-definition camera through a data bus. The camera, the LCD screen, the humidity sensor array, the current control module, and the safety protection module are electrically connected and exchange data; the current control module is provided with a high-voltage power converter and a precision current regulator, the high-voltage power converter converts 220V AC power into 2000V~5000V DC high-voltage power, and the precision current regulator accurately adjusts the high-voltage current to a range of 0.1A~10A; the safety protection module is provided with an overload protection relay, a short-circuit protection switch and a leakage protection circuit breaker. When the current exceeds the set threshold of 12A, the overload protection relay automatically disconnects the circuit When a short circuit fault is detected, the short circuit protection switch cuts off the power supply within 10ms. When the leakage current is detected to be greater than 30mA, the leakage protection circuit breaker immediately disconnects the main circuit. The control chip is provided with an electric shock painting prediction and display module, which is used to generate a predicted carbonization image of the next electric shock based on the wood board texture parameters extracted from the wood board surface image acquired by the high-definition camera, the wood board grid humidity parameters detected in real time by the humidity sensor array, the current wood board electric shock carbonization image, the next preset electric shock position coordinates, electric shock angle parameters, and current intensity parameters, and superimpose the predicted carbonization image of the next electric shock on the real-time image of the wood board through the LCD screen.

[0030] like Figure 2 As shown, the electric shock painting prediction and display module is used to perform the following steps:

[0031] S01, controlling the high-definition camera to obtain a real-time image of the current wooden board surface, gray-scaling the image and extracting the wooden board texture features, and establishing a wooden board texture parameter matrix, wherein the number of matrix rows is the number of image rows, the number of matrix columns is the number of image columns, and the matrix element value represents the texture density at the corresponding position;

[0032] S02, controlling the humidity sensor array to collect humidity values ​​at 16 measurement points on the surface of the wooden board, establishing a humidity distribution matrix, and expanding the humidity distribution matrix into a humidity parameter matrix of the same dimension as the texture parameter matrix through a bilinear interpolation algorithm;

[0033] S03. Calculate a carbonization key point matrix based on the wood board texture parameter matrix and historical electric shock carbonization data, where each element in the matrix represents a probability weight of carbonization reaction occurring at the corresponding position, with the weight value ranging from 0 to 1;

[0034] S04. Using the shortest path algorithm to solve the problem of optimizing the conduction path of current from the electric shock point to the edge of the wooden board, the wooden board texture parameter matrix is ​​converted into a weighted graph structure, with each pixel point as a node of the graph. The edge weights between adjacent pixels are determined by the texture density difference and the humidity difference. The shortest current conduction path from the preset electric shock position to each edge point of the wooden board is calculated using the Dijkstra algorithm. Based on the shortest path result and the wooden board texture parameter matrix, the carbonization texture accompanying degree coefficient is calculated. The coefficient represents the degree of influence of the texture structure at each position on the carbonization process. At the same time, a texture carbonization accompanying change point matrix is ​​generated.

[0035] S05. Inputting the humidity parameter matrix into a carbonization depth intelligent optimization model, the model predicts the carbonization depth at each location based on humidity distribution and current intensity, wherein the feature fusion weight parameter of the model is dynamically adjusted according to the current current intensity, the thickness of the wood board, and the preset electric shock angle, and combining the carbonization key point matrix to calculate the humidity carbonization accompanying degree coefficient, which represents the regulating effect of the humidity level at each location on the carbonization depth, and simultaneously generates a humidity carbonization accompanying change point matrix;

[0036] S06, substituting the preset coordinates of the next electric shock position, electric shock angle, and current intensity into the carbonization prediction algorithm, combining the texture carbonization accompanying change point matrix and the humidity carbonization accompanying change point matrix, and calculating the carbonization depth prediction value of each point within the electric shock action radius;

[0037] S07. Generate a predicted carbonization image for the next electric shock based on the predicted carbonization depth, and display the predicted image in a semi-transparent overlay on the real-time image of the wooden board on the LCD screen. The overlay transparency is set to 60% to facilitate the operator to preview the electric shock effect.

[0038] Among them, the carbonization key point matrix is ​​a probability weight matrix established according to the texture density distribution of the wooden board and the material characteristics of the wood, which is used to identify the sensitivity of the electric shock carbonization reaction at each position on the surface of the wooden board.

[0039] Among them, the carbonization texture accompaniment degree is a quantitative parameter that describes the interaction strength between the texture direction of the wood board and the current conduction path. The larger the value, the stronger the guiding effect of the texture at that position on the carbonization process.

[0040] Among them, the texture carbonization accompanying change point matrix records the spatial position coordinates where the carbonization effect changes significantly due to differences in texture structure. The matrix element value of 1 represents a change point, and the value of 0 represents an ordinary point.

[0041] Among them, the humidity carbonization accompaniment degree is a parameter that quantifies the degree of influence of the humidity level of the wood board on the depth of electric shock carbonization. The higher the humidity, the larger the accompaniment coefficient, which means that deep carbonization is more likely to occur at that location.

[0042] Among them, the humidity carbonization change point matrix identifies the key locations where carbonization effects may change suddenly due to local humidity differences, and is used to predict uneven carbonization during the electric shock process.

[0043] The specific structure of the carbonization depth intelligent optimization model is a jump connection model based on the UNet architecture. The model consists of two main parts: an encoder and a decoder. The encoder is responsible for extracting multi-scale features of humidity distribution, and the decoder is responsible for reconstructing the carbonization depth prediction map. Detailed information is transmitted between the encoder and the decoder through jump connections. The jump connection layer uses feature fusion weight parameters to control the fusion ratio of features at different levels. The model input is a 16-dimensional humidity vector and a current parameter vector, and the output is a carbonization depth prediction matrix corresponding to the size of the wooden board. The feature fusion weight parameters in the model are dynamically adjusted according to the combined value of the current current intensity, the thickness of the wooden board, and the preset electric shock angle through the weight adjustment function.

[0044] The steps for establishing the training data set of the carbonization depth intelligent optimization model specifically include collecting electric shock experimental data of different wood types. The experimental data include the electric shock carbonization results of 10 common wood types such as pine, oak, and walnut under different humidity conditions. Each type of wood is subjected to 100 electric shock experiments under 5 different humidity levels. The electric shock parameters include 16 levels of current intensity ranging from 0.5A to 8A, and 8 levels of electric shock angles ranging from 15 degrees to 90 degrees. The humidity distribution data, electric shock parameters and final carbonization depth distribution image of each experiment are recorded. The humidity distribution data and electric shock parameters are used as input features, and the carbonization depth distribution image is used as a label. A total of 8,000 groups of training samples are collected, and the data set is divided into training set, validation set and test set in a ratio of 7:2:1.

[0045] The steps of training the carbonization deep intelligent optimization model specifically include using the Adam optimizer to optimize the model parameters, setting the learning rate to 0.001, the batch size to 32, the total number of training rounds to 200 rounds, and the loss function using a weighted combination of mean square error loss and structural similarity loss with a weight ratio of 0.7:0.3. During the training process, a learning rate decay strategy is adopted, multiplying the learning rate by 0.8 every 50 rounds, and using an early stopping mechanism to prevent overfitting. When the validation set loss does not decrease for 10 consecutive rounds, training is stopped. After the model training is completed, performance evaluation is performed on the test set, and the evaluation indicators include mean absolute error, peak signal-to-noise ratio, and structural similarity index.

[0046] The weight adjustment function is used to adjust the feature fusion weight parameters of the carbonization depth intelligent optimization model. The function calculates the comprehensive influence factor based on the three parameters of current current intensity, wood board thickness, and preset electric shock angle. The comprehensive influence factor is calculated by dividing the current intensity value by 10 and multiplying it by the wood board thickness value, and then multiplying it by the sine value of the electric shock angle. When the comprehensive influence factor is less than 0.5, the linear weight adjustment function is used to set the feature fusion weight parameter to a linear interpolation result between 0.3 and 0.5. When the comprehensive influence factor is between 0.5 and 1.5, the quadratic weight adjustment function is used to set the feature fusion weight parameter to a quadratic interpolation result between 0.5 and 0.8. When the comprehensive influence factor is greater than 1.5, the exponential weight adjustment function is used to set the feature fusion weight parameter to an exponential interpolation result between 0.8 and 1.0.

[0047] The carbonization prediction algorithm is a comprehensive prediction algorithm established based on the current conduction physical model and the wood carbonization chemical reaction mechanism. The carbonization prediction algorithm realizes accurate prediction of the electric shock carbonization effect through multi-parameter coupling calculation. The core calculation process of the carbonization prediction algorithm includes four main stages: current density distribution calculation, temperature field analysis, carbonization reaction kinetics modeling and carbonization depth quantitative prediction. Among them, the current density distribution calculation stage establishes a group of partial differential equations for the current distribution inside the wooden board based on Ohm's law and Kirchhoff's law, and solves the current density distribution matrix of each point in the electric shock action area by the finite element method. The current density calculation formula is J(x, y) = σ(x, y) × E(x, y), where J(x, y) represents the current density at the position (x, y), σ(x, y) represents the conductivity at the position, and E(x, y) represents the electric field strength at the position. The conductivity is calculated by a nonlinear mapping function based on the wooden board texture parameter matrix and the humidity parameter matrix. The nonlinear mapping function takes into account the influence of wood fiber directionality, moisture content change and temperature effect on the conductivity.

[0048] The temperature field analysis stage establishes a temperature distribution model of the electric shock area based on the Joule thermal effect principle, and the temperature is calculated using the heat conduction equation. Wherein T represents temperature, α represents thermal diffusion coefficient, Q represents heat source intensity per unit volume, ρ represents wood density, and c represents specific heat capacity. The heat source intensity per unit volume Q is obtained by the dot product operation of the current density distribution matrix and the resistivity matrix. The thermal diffusion coefficient α is dynamically adjusted according to the wood species, the wood board texture parameter matrix, and the humidity parameter matrix through an empirical formula. The temperature field calculation is numerically solved using an explicit difference format, the time step is set to 0.1 ms, the spatial grid size is set to 0.5 mm, and the boundary condition is set to the third type boundary condition to consider convective heat transfer with the environment.

[0049] In the carbonization reaction kinetic modeling stage, a reaction rate model of wood pyrolysis carbonization is established based on the Arrhenius equation. The carbonization reaction rate equation is dα / dt=A×exp(-E / (RT))×f(α), where α represents the carbonization conversion rate, A represents the frequency factor, E represents the activation energy, R represents the gas constant, T represents the absolute temperature, and f(α) represents the reaction mechanism function. The frequency factor A and the activation energy E are obtained through experimental calibration according to the wood species. The frequency factor of pine wood is 1.2×10 13 s -1 , the activation energy is 180 kJ / mol, and the frequency factor of oak is 8.5×10 12 s -1 , the activation energy is 195 kJ / mol, and the frequency factor of walnut is 1.8×10 13 s -1 , the activation energy is 175 kJ / mol, and the reaction mechanism function f(α) adopts a three-dimensional diffusion model f(α)=1.5(1-α)[-ln(1-α)] 2 / 3 , the calculation of the carbonization conversion rate α is updated in each time step by a numerical integration method.

[0050] The carbonization depth quantitative prediction stage converts the carbonization conversion rate matrix into a visualized carbonization depth value. The conversion formula is D(x, y) = α(x, y) × D max ×K texture (x, y) × K humidity (x, y), where d(x, y) represents the carbonization depth at the position (x, y), α(x, y) represents the carbonization conversion rate at the position, and D max Indicates the theoretical maximum carbonization depth, K texture (x, y) represents the texture correction coefficient, K humidity (x, y) represents the humidity correction coefficient, the theoretical maximum carbonization depth D max The texture correction coefficient K is calculated by an empirical formula based on the thickness and density of the wood. textureThe humidity correction coefficient K is obtained by weighted calculation of the carbonization texture accompanying degree coefficient and the texture carbonization accompanying change point matrix. humidity The final carbonization depth prediction value is obtained by weighted calculation of the humidity carbonization accompanying degree coefficient and the humidity carbonization accompanying change point matrix, and the range of the final carbonization depth prediction value is limited to between 0 and the thickness of the wood board, and the values ​​outside the range are truncated.

[0051] The algorithm for generating a predicted carbonization image for the next electric shock based on the carbonization depth prediction value includes four main steps: depth value mapping, texture rendering, light and shadow effect processing, and image synthesis. The algorithm converts the numerical carbonization depth prediction matrix into an intuitive visual effect image, which facilitates the operator to predict the effect of the electric shock operation. The depth value mapping step uses a piecewise linear mapping function to convert the carbonization depth prediction value into a grayscale value. The piecewise linear mapping function is defined as a grayscale value of 255-400× when the carbonization depth prediction value is less than 0.5 mm. The carbonization depth prediction value, when the carbonization depth prediction value is between 0.5 mm and 2.0 mm, the grayscale value is 55-30×(the carbonization depth prediction value-0.5), when the carbonization depth prediction value is greater than 2.0 mm, the grayscale value is 10, and the grayscale value range is limited to between 0 and 255. The smaller the grayscale value, the deeper the carbonization degree. During the mapping process, Gaussian blur processing is performed on the boundary area, the blur radius is set to 2 pixels, and the standard deviation is set to 0.8 to eliminate the visual discontinuity caused by the sudden change of the carbonization depth prediction value.

[0052] The texture rendering step generates a realistic carbonized texture effect based on the wood board texture parameter matrix and the carbonization depth prediction value. The texture rendering step first extracts a texture direction vector field from the wood board texture parameter matrix, the texture direction vector field is calculated by a structure tensor method, and the eigenvector of the structure tensor represents the main texture direction. Then, the texture contrast and clarity are adjusted according to the carbonization depth prediction value. The adjustment formula is new texture value = original texture value × (1-0.6 × normalized carbonization depth prediction value) + carbonization base color value × normalized carbonization depth prediction value. The carbonization base color value is set according to the wood type, pine is RGB (101, 67, 33), oak is RGB (85, 53, 26), and walnut is RGB (71, 45, 22). The texture rendering step maintains the continuity and directionality of the natural texture of the wood to avoid unnatural fracture effects.

[0053] The light and shadow effect processing step simulates the lighting characteristics and shadow effects of the carbonized area to enhance the three-dimensional sense and realism of the predicted carbonized image. The light and shadow calculation is based on a simplified Phong lighting model, the ambient light component is set to 0.3, the diffuse reflection coefficient is dynamically adjusted according to the carbonization depth prediction value, and the calculation formula is diffuse reflection coefficient = 0.8-0.4×normalized carbonization depth prediction value, the specular reflection coefficient is set to 0.1, the light source position is set at an angle of 45 degrees directly above the wooden board, the light source intensity is 1.0, the normal vector is obtained according to the gradient calculation of the carbonization depth prediction value, and the gradient calculation of the carbonization depth prediction value adopts the Sobel operator. The shadow effect is calculated by the second-order derivative of the carbonization depth prediction value. Shadows are added to the concave area and highlights are added to the convex area. The shadow intensity is proportional to the second-order derivative of the carbonization depth prediction value, and the highlight intensity is proportional to the negative of the second-order derivative.

[0054] The image synthesis step superimposes and fuses the rendered carbonized texture image with the real-time image of the wooden board to generate the final predicted carbonized image. The image synthesis step adopts an Alpha blending algorithm, and the blending formula is output pixel value = pixel value of the real-time image of the wooden board × (1-transparency) + pixel value of the carbonized texture image × transparency. The transparency parameter is set to 0.6. Feathering processing is used at the edge of the carbonized area to reduce boundary mutations. The feathering radius is set to 3 pixels. Gamma correction is performed on the color during the synthesis process, and the gamma value is set to 1.2 to ensure that the display effect matches the actual carbonized color of the wood. The resolution of the predicted carbonized image finally generated is consistent with the image captured by the high-definition camera, and the frame rate is maintained at above 30fps, meeting the smoothness requirements of the real-time preview. The output format of the predicted carbonized image is a 24-bit RGB bitmap, which is transmitted and displayed in real time through the video interface of the LCD screen.

[0055] The carbonization conversion rate refers to the percentage of the mass of wood that has undergone carbonization reaction during the electric shock pyrolysis process to the mass of the original wood, and the numerical range is 0 to 1, where 0 means completely uncarbonized and 1 means completely carbonized. The theoretical maximum carbonization depth refers to the maximum carbonization penetration depth that wood can theoretically reach under given electric shock parameters, which is determined by the wood species, density and thickness. The texture correction coefficient is a correction parameter used to adjust the carbonization depth calculation result to adapt to the wood texture characteristics, taking into account the influence of texture density and direction on the carbonization process. The humidity correction coefficient is a correction parameter used to adjust the carbonization depth calculation result to adapt to the wood humidity distribution, reflecting the regulatory effect of humidity level on carbonization depth.

[0056] The piecewise linear mapping function is a piecewise definition function that maps a continuous numerical domain to a target numerical domain, and adopts different linear transformation formulas in different numerical intervals.

[0057] The texture direction vector field is a two-dimensional vector field that describes the texture orientation at each point on the wood surface. Each vector represents the primary texture direction at the corresponding location. The structure tensor is a second-order tensor used to calculate the local structural features of the image. Texture direction and intensity information is extracted by analyzing image gradient information. The normalized carbonization depth prediction value is a dimensionless parameter obtained by dividing the carbonization depth prediction value by the maximum possible carbonization depth, and has a value range of 0 to 1.

[0058] The Phong lighting model is a classic model used in computer graphics to simulate surface lighting effects. It includes three components: ambient light, diffuse light, and specular light. The Sobel operator is a linear operator used in digital image processing for edge detection and gradient calculation. It calculates the first-order derivative of an image through a convolution operation. The Alpha blending algorithm is a standard algorithm for image transparency blending in computer graphics. It controls the blending ratio of foreground and background images through a transparency parameter. Gamma correction is a nonlinear transformation used to adjust image brightness and contrast, compensating for the nonlinear response characteristics of display devices.

[0059] The specific implementation methods of the above steps are described in detail below. The interactive electric shock wood painting device of the present invention adopts an integrated design and is mainly composed of four modules: a mechanical execution system, an intelligent control system, a safety protection system, and a human-computer interaction system. The entire device occupies an area of ​​approximately 2 meters x 1.5 meters and is 1.8 meters high. It uses a steel frame structure to ensure the stability and operational safety of the equipment. The robotic arm adopts a six-degree-of-freedom articulated design with an arm length of 800mm and an effective working radius of 600mm, capable of precise positioning in three-dimensional space. The base of the robotic arm is made of cast iron and weighs 80kg. It is fixed to the left side of the wooden fixing platform with four M16 bolts to ensure the stability of the equipment during operation. Each joint of the robotic arm is driven by a servo motor and equipped with a high-precision encoder for position feedback. The repeatability accuracy can reach ±0.05mm. The carbon steel electrode is a cylindrical design with a diameter of 8mm. The electrode tip is precisely machined to a 60-degree cone and the surface is chrome-plated to improve conductivity and corrosion resistance. The wood panel fixing platform consists of a 1200mm x 800mm rectangular steel frame with a 20mm-thick top made of rust-resistant Q235 steel. A 1000mm x 600mm wood panel positioning slot, 15mm deep, is located in the center of the platform. Adjustable clamps are located around the edges to accommodate wood panels of varying sizes. An insulating rubber mat is placed at the bottom of the slot to ensure safe and controllable current conduction.

[0060] The humidity sensor array consists of 16 SHT30 digital humidity and temperature sensors, evenly distributed in a 4×4 pattern. Each sensor is encapsulated in a waterproof housing and connected to the data acquisition module through a sealed connector. The sensor measurement accuracy is

[0061] ±2% RH, with a response time of less than 8 seconds. Data acquisition is achieved via the I2C bus with a sampling frequency of 1Hz. The HD camera uses a 5-megapixel CMOS sensor and an autofocus lens. The gimbal allows for ±45-degree pitch adjustment and 360-degree rotation. The aluminum alloy bracket provides excellent rigidity and stability.

[0062] The console shell is made of 1.5mm thick cold rolled steel plate, with electrostatic spraying treatment on the surface.

[0063] 600mm × 400mm × 1200mm. The internal electronic control compartment is divided into high-voltage and low-voltage areas, physically isolated by metal partitions. The control chip uses an ARM Cortex-A7 quad-core processor with a main frequency of 1.2GHz, equipped with 1GB of DDR3 memory and 8GB of eMMC storage, and runs an embedded Linux operating system.

[0064] The current control module integrates a high-frequency switching power converter with an input voltage range of 180V-260V and continuously adjustable output voltage from 2000V to 5000V, with a ripple factor of less than 1%. The precision current regulator utilizes PWM modulation technology, achieving a current regulation accuracy of 0.01A and a response time of less than 10ms. The safety protection module features a triple protection mechanism. The overload protection relay utilizes a thermal-magnetic design with an operating current of 12A±5%, the short-circuit protection switch has a response time of less than 10ms, and the leakage protection circuit breaker has a sensitivity of 30mA, ensuring operator and equipment safety.

[0065] The LCD screen uses a 15.6-inch IPS panel with a resolution of 1920×1080 and a brightness of 300cd / m 2 , supports touch operation, connects to the control chip through the HDMI interface, and realizes real-time image display and human-computer interaction functions.

[0066] The specific implementation methods of the electric shock painting prediction and display module execution steps are described in detail below.

[0067] The specific implementation method of step S01 is to control the high-definition camera to collect real-time color images of the wooden board surface. The image resolution is set to 1920×1080 pixels and the frame rate is 30fps. After the acquisition is completed, the image is immediately grayscaled. The grayscale processing adopts the weighted average method, and the pixel values ​​of the three color channels of red, green and blue are linearly combined according to the weight coefficients of 0.299, 0.587 and 0.114 to obtain a single-channel grayscale image. Next, the grayscale image is subjected to texture feature extraction. The local binary pattern algorithm is used to analyze the grayscale relationship between each pixel and its 8 neighboring pixels, and the texture density parameter is calculated. The texture density parameter reflects the roughness and directional characteristics of the wood surface texture. The value range is 0 to 255. The larger the value, the more obvious the texture. Finally, a texture parameter matrix corresponding to the image size is established, and each element in the matrix stores the texture density value of the corresponding pixel position.

[0068] The specific implementation of step S02 involves controlling the 16 sensors in the humidity sensor array to synchronously collect humidity values ​​at different locations on the wooden surface. The acquisition frequency is set to 1 Hz, and the humidity measurement accuracy is ±2% relative humidity. The sensors are arranged in a 4×4 matrix, with 50 mm spacing between adjacent sensors, covering a 200 mm × 200 mm wooden surface area. The 16 humidity measurements are arranged according to the spatial positions of the sensors to form a 4×4 humidity distribution matrix. To maintain the same spatial resolution as the texture parameter matrix, the humidity distribution matrix is ​​expanded using a bilinear interpolation algorithm. This algorithm uses distance weighting based on the four nearest known humidity values ​​to calculate a humidity estimate at the intermediate location. The weighting calculation takes into account the inverse square relationship of spatial distance. The expanded humidity parameter matrix has the same row and column dimensions as the texture parameter matrix, with each matrix element representing the humidity percentage at the corresponding location.

[0069] The specific implementation method of step S03 is to calculate the carbonization key point matrix based on the wood board texture parameter matrix and the historical electric shock carbonization database. The historical carbonization database contains the electric shock carbonization test results of different wood species under various texture conditions, and the data volume reaches 5,000 groups of samples. The Bayesian statistical method is used to analyze the correlation between texture density and carbonization probability, and a carbonization probability prediction model is established. For areas with texture density values ​​below 50, the carbonization probability weight is set to between 0.2 and 0.4; for areas with texture density values ​​between 50 and 150, the carbonization probability weight is set to between 0.4 and 0.7; for areas with texture density values ​​above 150, the carbonization probability weight is set to between 0.7 and 1.0. The weight value of each element in the carbonization key point matrix is ​​calculated by a nonlinear mapping function of texture density. The mapping function takes into account the differences in material properties of wood species. The mapping coefficient of pine is 1.2, oak is 0.9, and walnut is 1.1.

[0070] The specific implementation of step S04 is to convert the wood board texture parameter matrix into a weighted undirected graph structure, with each pixel as a node of the graph and edges connecting adjacent pixels. The calculation of edge weights takes into account the two factors of texture density difference and humidity difference. The calculation formula is that the edge weight is equal to the absolute value of the texture density difference multiplied by 0.6 plus the absolute value of the humidity difference multiplied by 0.4. The Dijkstra shortest path algorithm is used to calculate the shortest conduction path from the preset electric shock position to the four edges of the wood board. The time complexity of the algorithm is O(n 2 ), where n is the total number of nodes. The carbonization texture adjacency coefficient is calculated based on the shortest path results. This coefficient reflects the strength of the texture structure's guiding effect on the current conduction path. The texture adjacency coefficient is determined by the ratio of path length to straight-line distance; a larger ratio indicates a more significant texture influence on current conduction. A texture carbonization adjacency change point matrix is ​​also generated to identify spatial locations where significant changes in the texture structure occur. The threshold is set at pixels where the texture density gradient exceeds 30.

[0071] Step S05 involves inputting the expanded humidity parameter matrix into a deep learning-based intelligent optimization model for carbonization depth prediction. The model employs an encoder-decoder architecture. The encoder consists of four downsampling layers, each using a 3×3 convolution kernel and a 2×2 max pooling operation, with a Reinforced Lu (ReLU) activation function. The decoder consists of four upsampling layers, using transposed convolutions to restore the feature map size. Skip connections fuse the features of the corresponding encoder and decoder layers, with the fusion weight parameter dynamically adjusted based on the current intensity, board thickness, and the preset shock angle. When the current intensity is below 3 amps, the fusion weight is set to 0.3; when the current intensity is between 3 and 6 amps, the fusion weight is set to 0.5; and when the current intensity is above 6 amps, the fusion weight is set to 0.8. The model output is a carbonization depth prediction matrix corresponding to the board size, with values ​​in millimeters. The humidity-carbonization dependency coefficient is calculated in conjunction with the carbonization key point matrix. This coefficient represents the regulatory effect of humidity level on carbonization depth, with higher humidity indicating a greater dependency coefficient.

[0072] The specific implementation of step S06 involves inputting the three parameters (preset coordinates of the next electric shock location, electric shock angle, and current intensity) into a carbonization prediction algorithm for comprehensive calculation. The carbonization prediction algorithm is based on a physical model of current conduction and the chemical reaction mechanism of wood pyrolysis. It comprises four computational stages: current density distribution calculation, temperature field analysis, carbonization reaction kinetics modeling, and quantitative prediction of carbonization depth. The current density distribution calculation uses the finite element method to solve the Poisson equation, with a meshing accuracy of 0.5 mm and first-class boundary conditions. The temperature field analysis is based on the Joule heating effect and the heat conduction equation, with a time step of 0.1 milliseconds and a central difference scheme for spatial discretization. The carbonization reaction kinetics modeling uses the Arrhenius equation to describe the pyrolysis reaction rate. The reaction activation energy is determined by the wood species: 180 kilojoules per mole for pine and 195 kilojoules per mole for oak. Combining the matrix of texture carbonization-related changes and the matrix of humidity carbonization-related changes, the predicted carbonization depth is calculated for each point within the electric shock's action radius. The action radius is determined by the current intensity and ranges from 20 to 80 mm.

[0073] The specific implementation of step S07 involves generating a visual predicted carbonization image based on the predicted carbonization depth value and displaying it in real time. The image generation process includes four stages: depth value mapping, texture rendering, lighting and shadow processing, and image synthesis. Depth value mapping uses a piecewise linear function to convert the carbonization depth value into a grayscale value, with shallow carbonization areas mapped to light gray and deep carbonization areas mapped to dark gray. Texture rendering generates a realistic carbonization effect based on the original wood board texture and carbonization depth values, using a structured tensor method to extract texture directional information and maintain the continuity of the carbonized texture. Light and shadow processing uses a simplified Phong lighting model to simulate the light and dark variations in the carbonized area. The light source is positioned at a 45-degree angle directly above the wood board, and the ambient light coefficient is 0.3. Image synthesis uses an alpha blending algorithm to overlay the predicted carbonization image with the real-time wood board image, with a transparency setting of 60%. The synthesized result is displayed in real time on an LCD screen. The display frame rate is maintained at above 30 fps, and the image resolution is consistent with the image captured by the camera, ensuring smoothness and clarity of the preview effect.

[0074] The carbonization depth intelligent optimization model utilizes a deep convolutional neural network based on the UNet architecture. Its overall structure comprises three main components: an encoder, a decoder, and skip connections. The encoder consists of four downsampling modules, each consisting of two 3×3 convolutional layers, a batch normalization layer, a ReLU activation function, and a 2×2 max pooling layer. The first downsampling module has 16 input channels, corresponding to the data from 16 humidity sensors, and outputs 32 channels; the second module has 64 output channels; the third module has 128 output channels; and the fourth module has 256 output channels. The encoder extracts multi-scale feature representations of the humidity distribution layer by layer, from local details to global semantic features. The decoder consists of four upsampling modules, each consisting of a 2×2 transposed convolutional layer, two 3×3 convolutional layers, a batch normalization layer, and a ReLU activation function. The number of channels is 128, 64, 32, and 1, respectively. The final output is a single-channel carbonization depth prediction map. Skip connections concatenate and fuse the feature maps of the corresponding encoder and decoder layers, preserving detailed information and preventing information loss. The fusion weight parameters are implemented through a learnable 1×1 convolutional layer, and the weight values ​​are dynamically adjusted according to the current intensity, board thickness, and electric shock angle.

[0075] The training dataset for the intelligent carbonization depth optimization model consists of four steps: data collection, preprocessing, annotation, and partitioning. During the data collection phase, electric shock carbonization experiments were conducted on 10 common wood species, including pine, oak, and walnut. Five different humidity levels were selected for each species: 15%, 25%, 35%, 45%, and 55% relative humidity. The electric shock parameter settings included 16 current intensity levels, evenly distributed from 0.5 amps to 8 amps, and eight shock angle levels, evenly distributed from 15 degrees to 90 degrees. Each parameter combination was repeated 10 times, resulting in a total of 8,000 experimental samples. During the data preprocessing phase, the humidity sensor data was filtered and denoised using a three-point moving average filter with a filter window length of three sampling points to remove measurement noise. After the experiments, the carbonization depth distribution of the wood panels was measured using a high-precision 3D scanner with a scanning accuracy of 0.1 mm, generating corresponding annotated images of the carbonization depth. The dataset is divided into training set, validation set, and test set in a ratio of 7:2:1. The training set contains 5,600 samples, the validation set contains 1,600 samples, and the test set contains 800 samples, ensuring that different wood species and parameter conditions are evenly distributed in each dataset.

[0076] The core technical ideas of this invention mainly include three key innovations: real-time monitoring technology of wood board status based on humidity sensor array, multi-parameter coupling carbonization prediction algorithm technology, and real-time prediction carbonization image overlay display technology.

[0077] The real-time monitoring technology of the wooden board status based on the humidity sensor array realizes the grid-based precise detection of the humidity distribution on the wooden board surface through the arrangement of 16 humidity sensors in a 4×4 matrix. Compared with the traditional electric wood painting device that only relies on experience to judge the state of wood, this technology can obtain the humidity parameter matrix of each area of ​​the wooden board in real time, providing accurate basic data support for the subsequent carbonization effect prediction. It fundamentally solves the problem of uncontrollable carbonization effect caused by the traditional method of inability to accurately grasp the state of wood, and significantly improves the predictability and precision control ability of the electric painting process.

[0078] The multi-parameter coupled carbonization prediction algorithm technology deeply integrates and calculates multi-dimensional information such as the wood board texture parameter matrix, humidity parameter matrix, current intensity parameters, and electric shock angle parameters, and establishes a comprehensive prediction system based on physical conduction models and chemical reaction kinetics. Compared with the traditional method of making rough estimates based on operator experience, this algorithm can accurately calculate the carbonization conversion rate and depth distribution based on scientific principles such as the Arrhenius equation and the heat conduction equation, thereby realizing quantitative prediction of the electric shock carbonization effect, which has completely changed the technical status quo in the field of electric shock wood painting that has long relied on subjective judgment.

[0079] The real-time predicted carbonization image overlay display technology converts the carbonization depth prediction value calculated by the prediction algorithm into an intuitive visual effect image through depth value mapping, texture rendering, light and shadow effect processing and other steps, and displays it superimposed on the real-time image of the wooden board with 60% transparency. Compared with the traditional method of blind operation where the electric shock effect cannot be foreseen at all, this technology enables the operator to intuitively preview the carbonization effect before implementing the electric shock, greatly reducing the risk of operational errors and improving painting accuracy and success rate.

[0080] The synergistic effect of the three key technologies forms a complete technical closed loop from data collection to intelligent prediction to visual presentation. The precise status data provided by the humidity sensor array provides reliable input for the multi-parameter coupling algorithm. The algorithm prediction results are intuitively presented to the operator through image overlay technology, realizing a fundamental transformation of electric wood painting from the traditional experience-driven mode to the data-driven intelligent mode. Compared with the blind operation mode of the existing technology, the technical solution of the present invention can achieve precise control and real-time preview of the carbonization effect, and completely solves the core technical problems of unpredictable effects and high operational risks that have long existed in the field of electric wood painting from a technical principle.

[0081] Specifically, the principle of the present invention is: the core of the present invention's ability to solve the problem of carbonization effect prediction lies in the construction of a multi-level physical-chemical coupling prediction model, which accurately describes the complete physical process from electric shock parameters to the final carbonization effect. First, the texture information of the wood surface is obtained through a high-definition camera and a texture parameter matrix is ​​established. At the same time, the humidity distribution on the wood surface is detected using a humidity sensor array. These two key parameter matrices provide accurate initial conditions for subsequent carbonization predictions. The texture parameters directly affect the conduction path of the current inside the wood, while the humidity distribution determines the difference in conductivity in different areas. These factors together determine the spatial distribution characteristics of the carbonization reaction.

[0082] The physical foundation of the prediction algorithm is based on Ohm's law and the Joule heating effect. By solving a set of partial differential equations for the current density distribution within the wood, the current density and temperature distribution at each point within the electric shock zone are accurately calculated. The temperature field is calculated using the heat conduction equation and takes into account the anisotropic properties of wood, ensuring the accuracy of the temperature prediction. Furthermore, a carbonization reaction kinetic model based on the Arrhenius equation is introduced to convert the temperature field into a carbonization conversion rate distribution. This model fully accounts for the differences in the pyrolysis characteristics of different wood species, and the reliability of the prediction results is guaranteed by experimentally calibrated kinetic parameters.

[0083] The logical rationality of the technical solution of the present invention is reflected in the organic combination of its multi-scale modeling strategy and intelligent optimization mechanism. The deep learning model of the UNet architecture performs intelligent processing of complex parameters such as humidity, effectively capturing the nonlinear influence of humidity distribution on carbonization depth. The jump connection mechanism in the model ensures the effective fusion of multi-scale feature information. At the same time, the weight adjustment function dynamically adjusts the model weight according to the real-time changes of the electric shock parameters, so that the prediction results can adapt to different operating conditions. Finally, through the image rendering algorithm based on the physical illumination model, the numerical carbonization depth prediction results are converted into intuitive visual images, realizing a complete closed loop from physical prediction to visual presentation, and providing the operator with an accurate and reliable effect preview function.

[0084] A specific embodiment 1 of the present invention is provided below. The specific implementation of each step in this embodiment 1 is described in detail as follows.

[0085] The specific implementation of step S01 is the same as above and will not be described in detail here.

[0086] In step S02, the calculation process of the bilinear interpolation algorithm is described in detail as follows:

[0087]

[0088] Where H(x, y) is the humidity interpolation result at the target position (x, y); H 11, H 21 , H 12 , H 22 is the measurement value of the nearest four humidity sensors; (x1, y1), (x2, y1), (x1, y2), (x2, y2) are the spatial coordinate positions of the four humidity sensors. Among them, the humidity sensor measurement value H 11 , H 21 , H 12 , H 22 The humidity is obtained experimentally, including step 1: preheating the sensor for 60 seconds to reach a stable state; step 2: continuously collecting 10 humidity values ​​and calculating the average value as the final measurement result.

[0089] The specific implementation of step S03 is the same as above and will not be described in detail here.

[0090] In step S04, the edge weight calculation formula is specifically expressed as:

[0091] W ij =0.6×|T i -T j |+0.4×|H i -H j |+α;

[0092] Where W ij is the edge weight between pixel i and pixel j; T i , T j is the texture density value of the corresponding pixel point; H i , H j is the humidity parameter value of the corresponding pixel point; α is the weight adjustment factor, which ranges from 0.1 to 0.3. The calculation formula of the carbonization texture accompanying degree coefficient is specifically expressed as:

[0093]

[0094] Where K texture (x, y) is the carbonization texture accompanying coefficient at position (x, y); L path (x, y) is the shortest path length from the shock point to the position (x, y); L direct (x, y) is the straight-line distance from the shock point to the position (x, y); β wood is the wood species correction coefficient, which is 1.2 for pine, 0.9 for oak, and 1.1 for walnut; f(T(x, y)) is the texture density correction function, expressed as f(T) = 1 + 0.002 × T, where T is the texture density value.

[0095] In step S05, the calculation formula of the humidity carbonization accompanying degree coefficient is specifically expressed as:

[0096]

[0097] Where K humidity (x, y) is the humidity carbonization coefficient at the position (x, y), dimensionless; H(x, y) is the humidity parameter value at the position (x, y), unit is %; is the humidity gradient, in % / mm; γ1, γ2, γ3 are empirical coefficients, which are 0.8, 1.3, and 0.2 respectively; ε h is the humidity correction error term, ranging from 0.05 to 0.15. The calculation uses the Sobel operator, which is expressed as:

[0098]

[0099] Where G x and G y The gradient components in the x-direction and y-direction are calculated in % / mm using a 3×3 convolution kernel. The weight adjustment function is used to adjust the feature fusion weight parameters and the comprehensive influencing factor F combine The calculation formula is expressed as:

[0100]

[0101] Where, F combine is the comprehensive impact factor, dimensionless; I current is the current intensity, the unit is A; h board is the thickness of the board, in mm; θ angle is the electric shock angle, in degrees. Feature fusion weight parameter W fusion The calculation uses piecewise functions:

[0102] When F combine <0.5: W fusoin =0.3+0.4×F combine ;

[0103] When 0.5≤F combine ≤1.5: W fusion =0.5+0.3×(F combine -0.5) 2 ;

[0104] When F combine >1.5:

[0105] Where W fusion It is the feature fusion weight parameter, dimensionless, and ranges from 0.3 to 1.0.

[0106] In step S06, the calculation formula of the current density distribution is specifically expressed as:

[0107] J(x,y)=σ(x,y)×E(x,y);

[0108] Where J(x, y) is the current density at position (x, y) in A / m 2 ; σ(x, y) is the conductivity at position (x, y), in S / m; E(x, y) is the electric field intensity at position (x, y), in V / m. The nonlinear mapping function of conductivity σ(x, y) is expressed as:

[0109] σ(x,y)=σ0×[1+δ1×T(x,y)+δ2×H(x,y)+δ3×Θ(x,y)];

[0110] Where σ0 is the reference conductivity, which is 5×10 -4 S / m; δ1, δ2, δ3 are correction coefficients, which are 2×10 -3 , 1.5×10 -2 , 8×10 -3 Θ(x, y) is the temperature effect parameter, dimensionless, ranging from 0.8 to 1.2. The heat conduction equation in temperature field analysis is expressed as:

[0111]

[0112] Where T is temperature in K; t is time in seconds; α is thermal diffusion coefficient in meters. 2 / s; Q is the heat source intensity per unit volume, in W / m 3 ; ρ is the wood density, unit is kg / m 3 ; c is the specific heat capacity, the unit is J / (kg·K). The calculation formula of the heat source intensity per unit volume Q is:

[0113] Q(x, y) = J(x, y) 2 ×R(x,y);

[0114] Where R(x, y) is the resistivity at position (x, y) in Ω·m. The reaction rate equation in carbonization reaction kinetics modeling is expressed as:

[0115]

[0116] Where α is the carbonization conversion rate, dimensionless, ranging from 0 to 1; A is the frequency factor, unit is s -1E is the activation energy, in J / mol; R is the gas constant, 8.314 J / (mol·K); T is the absolute temperature, in K; f(α) is the reaction mechanism function, expressed using a three-dimensional diffusion model as f(α) = 1.5(1-α)[-ln(1-α)] 2 / 3 The conversion formula for quantitative prediction of carbonization depth is expressed as:

[0117] D(x, y) = α(x, y) × D max ×K texture (x, y) × K humidity (x, y);

[0118] Where D(x, y) is the carbonization depth at position (x, y), in mm; D max The theoretical maximum carbonization depth, in mm, is obtained through the empirical formula Calculation, where ζ1, ζ2, ζ3 are empirical parameters, which are 15.8, 0.6, and 0.4 respectively, and h is the thickness of the board in mm.

[0119] In step S07, the calculation formula of the piecewise linear mapping function is specifically expressed as:

[0120] When D(x, y)<0.5: G(x, y)=255−400×D(x, y);

[0121] When 0.5≤D(x, y)≤2.0: G(x, y)=55-30×(D(x, y)-0.5);

[0122] When D(x, y)>2.0: G(x, y)=10;

[0123] Where G(x, y) is the grayscale value at position (x, y); D(x, y) is the predicted carbonization depth, in mm. The calculation formula for the new texture value in texture rendering is expressed as:

[0124] T new (x, y) = T orig (x, y)×[1-0.6×D norm (x,y)]+C base ×D norm (x, y);

[0125] Where, T new (x, y) is the new texture value; T orig (x, y) is the original texture value; D norm (x, y) is the normalized carbonization depth prediction value, calculated as D norm (x, y) = D(x, y) / D max ; C baseis the carbonization base color value. The calculation formula of the diffuse reflection coefficient in light and shadow effect processing is expressed as:

[0126] k d (x, y) = 0.8 - 0.4 × D norm (x, y);

[0127] Where k d (x, y) is the diffuse reflection coefficient at the position (x, y). The Alpha blending algorithm formula for image synthesis is expressed as:

[0128] I output (x, y) = I board (x, y)×(1-τ)+I carbon (x, y) × τ;

[0129] Where, I output (x, y) is the output pixel value; I board (x, y) is the real-time image pixel value of the wooden board;

[0130] I carbon (x, y) is the pixel value of the carbonized texture image; τ is the transparency parameter, which is 0.6.

[0131] To better understand and implement the present invention, Example 2, a specific application scenario, is provided below: Researchers constructed an interactive electric shock wood painting device in a woodworking laboratory to study the application of electric shock carbonization technology in artistic creation. This device was primarily used to create artistic carbonized texture patterns on oak wood surfaces.

[0132] like Figure 3 As shown, the core components of the device in this embodiment include a wood panel fixing platform 1, made of 304 stainless steel, with a surface measuring 500×400 mm and a thickness of 10 mm. The platform is equipped with 5 mm deep wood panel positioning grooves for securing the oak panels to be processed. The wood panel fixing clamps 3, made of aluminum alloy with an adjustable clamping force range of 50 to 200 N, are bolted to the four corners of the panel fixing platform 1. The oak panels 4 to be processed measure 300×250×20 mm, have a moisture content of 25%, and are surface-polished with 220-grit sandpaper.

[0133] Robotic arm 10 is a six-axis industrial robot with an 800mm working radius, a repeatability of ±0.05mm, and a maximum load capacity of 10kg. A carbon steel electrode 2 is mounted at the end of robot arm 10. Made of high-carbon steel, the electrode has an 8mm diameter and a length of 50mm. The tip of the electrode is tapered at a 45° angle. Robotic arm 10 is controlled via a robot control switch 6, which integrates an up / down control joystick 7 and a left / right control joystick 8. The joysticks have a travel range of ±15mm and a control accuracy of 0.1mm.

[0134] Camera 13 is a high-definition industrial camera with a resolution of 1920×1080 pixels and a frame rate of 30 fps. It is equipped with a zoom lens with a focal length range of 12 to 36 mm. Camera 13 is mounted 300 mm above the wooden table 1 using a universal bracket, with the lens facing the tabletop. A UV protection screen 9, made of UV-resistant glass measuring 400×300×5 mm and boasting a light transmittance of 92%, is installed in front of camera 13 to protect it from UV damage generated during the electric shock process.

[0135] Touchscreen 5 is a 15.6-inch capacitive touchscreen with a resolution of 1920×1080 pixels, a response time of 10ms, and support for 10 simultaneous touch points. Touchscreen 5 is mounted on the front of the console at a 30° tilt angle, making it easier for the operator to observe and operate. Drawing hood 14 is made of transparent acrylic, measuring 600×500×400mm with an 8mm thickness. A 20mm diameter exhaust hole is located on the top to remove smoke generated during the electric shock process.

[0136] As shown in Table 1, the detailed parameter configuration of the humidity sensor in the device is:

[0137] Table 1 Humidity sensor array parameter configuration table

[0138] Sensor number Position coordinates (mm) Measuring range (%) Accuracy (%) Response time (s) H01 (50,50) 0～100 ±2 8 H02 (50,100) 0～100 ±2 8 H03 (50,150) 0～100 ±2 8 H04 (50,200) 0～100 ±2 8 H05 (100,50) 0～100 ±2 8 H06 (100,100) 0～100 ±2 8 H07 (100,150) 0～100 ±2 8 H08 (100,200) 0～100 ±2 8 H09 (150,50) 0～100 ±2 8 H10 (150,100) 0～100 ±2 8 H11 (150,150) 0～100 ±2 8 H12 (150,200) 0～100 ±2 8 H13 (200,50) 0～100 ±2 8 H14 (200,100) 0～100 ±2 8 H15 (200,150) 0～100 ±2 8 H16 (200,200) 0～100 ±2 8

[0139] The 16 humidity sensors are arranged in a 4×4 matrix and embedded under the wooden table 1. The sensor spacing is 50mm and the coverage area is 200×200mm. Each sensor is connected to the control chip through a data line, and the data transmission adopts I 2 C bus protocol with a transmission rate of 400kHz.

[0140] The device status indicator 11 is a three-color LED with a diameter of 10mm. Green indicates normal operation, yellow indicates standby mode, and red indicates a fault or emergency shutdown. Mounted on the top of the console, the indicator has a brightness of 1000cd and a visible distance of up to 10m. The device power switch 12 is a rotary switch rated at 25A and 220V. It has an integrated indicator light and fuse.

[0141] The control chip uses an ARM Cortex-A7 processor with a main frequency of 1.2GHz, a memory capacity of 1GB, and a storage capacity of 16GB. The control chip is connected to the drive motor of the robotic arm 10 via a CAN bus with a communication rate of 1Mbps. It is connected to the carbon steel electrode 2 via an analog input interface with 12-bit analog input accuracy and a sampling frequency of 1kHz. It is connected to the camera 13 via a USB interface that supports the USB 3.0 standard and a transmission rate of 5Gbps. It is connected to the touch screen 5 via an HDMI interface that supports 1080p resolution output.

[0142] The current control module, integrated within the control console, consists of a high-voltage power converter and a precision current regulator. The high-voltage power converter accepts a 220V AC input voltage and has an adjustable output voltage range of 2000-5000V DC. It delivers 2kW of power and achieves 90% efficiency. The precision current regulator utilizes IGBT power devices, offering a current regulation range of 0.1-10A, an accuracy of 0.01A, and a response time of 1ms.

[0143] As shown in Table 2, the detailed configuration parameters of the security protection module are:

[0144] Table 2 Security protection module parameter configuration table

[0145] protective device Operating current (A) Action time (ms) Reset method Operating voltage (V) Overload protection relay 12 100 Manual 220 Short-circuit protection switch 50 10 automatic 220 Leakage protection circuit breaker 0.03 30 Manual 220

[0146] The safety protection module ensures that the device can promptly cut off power in abnormal situations, protecting operators and equipment. The overload protection relay uses a thermal magnetic structure and automatically disconnects the circuit when the current exceeds the set threshold of 12A. The short-circuit protection switch uses an electromagnetic structure with instantaneous operation characteristics, and cuts off power within 10ms when a short circuit fault is detected. The leakage protection circuit breaker uses an electronic structure and immediately disconnects the main circuit when it detects a leakage current exceeding 30mA.

[0147] During operation, the researchers first placed an oak board 4 in the positioning slot of the board fixture 1 and secured it with the board clamp 3. The device was activated using the power switch 12. The device status indicator 11 turned green, indicating normal system operation. The operator then used the touch screen 5 to set the shock parameters, including the shock position coordinates (150, 125), a shock angle of 60°, a current intensity of 3.5A, and a shock duration of 2 seconds.

[0148] Camera 13 begins capturing real-time images of the wood board surface. After grayscale processing, texture features are extracted and a 300×250 texture parameter matrix is ​​established. The humidity sensor array simultaneously collects humidity values ​​at 16 measurement points on the wood board surface. The measured humidity values ​​are: H01 = 28%, H02 = 26%, H03 = 24%, H04 = 22%, H05 = 30%, H06 = 28%, H07 = 26%, H08 = 24%, H09 = 32%, H10 = 30%, H11 = 28%, H12 = 26%, H13 = 34%, H14 = 32%, H15 = 30%, and H16 = 28%.

[0149] Based on the collected texture and humidity parameters, the control chip uses an intelligent carbonization depth optimization model to predict the carbonization effect. The model inputs are a 16-dimensional humidity vector and a current parameter vector, and outputs a 300×250 carbonization depth prediction matrix. The prediction results show a carbonization depth of 1.8mm at the shock location (150, 125), with an impact radius of 35mm.

[0150] The robotic arm 10 moves to the designated position according to preset electric shock parameters. The carbon steel electrode 2 contacts the surface of the wooden board, and the current control module outputs a 3.5A current for electric shock. During the electric shock process, the high-voltage power converter converts 220V AC power to 3500V DC power, and the precision current regulator precisely adjusts the current to 3.5A. After the electric shock lasts for 2 seconds, the robotic arm 10 automatically raises the electrode, completing the electric shock painting operation.

[0151] As shown in Table 3, key parameters monitored and recorded during the electric shock process:

[0152] Table 3 Key parameter monitoring record table of electric shock process

[0153]

[0154]

[0155] After the electric shock is completed, the camera 13 recaptures the surface image of the wooden board and compares and analyzes it with the predicted result, such as Figure 4As shown, an arrow indicates the predicted result. The actual carbonization depth is 1.8mm, exactly as predicted, and the carbonized area exhibits a natural wood grain effect. Touch screen 5 allows for real-time observation of the before-and-after effects, with a 94% similarity between the predicted image and the actual result.

[0156] Traditional wood carbonization processes usually use flame spray guns or hot irons for surface treatment. These methods have technical problems such as uneven carbonization depth, unnatural texture effects, and high operational risks. The standard deviation of carbonization depth of the flame spray gun method is 0.8mm, and the standard deviation of carbonization depth of the hot iron method is 0.6mm. In addition, traditional methods cannot accurately predict the carbonization effect, and often require multiple experiments to obtain the ideal texture effect. The present invention adopts precise current control and intelligent prediction technology, and the standard deviation of carbonization depth is reduced to 0.2mm, and the prediction accuracy reaches 94%, which is 15% higher than the traditional method in terms of precision control. At the same time, the carbonization texture produced by the electric shock method is more natural, and can be carbonized along the natural texture direction of the wood, maintaining the aesthetic characteristics of the wood itself. The electric shock process can be precisely controlled by preset parameters, avoiding the human operation errors in the traditional method and improving the consistency and reproducibility of the wood carbonization process.

[0157] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. An interactive electric shock wood painting device, comprising a robotic arm, a control console, a camera, a liquid crystal display, a wooden board fixing platform, a humidity sensor array, a current control module, a safety protection module, and a control chip, characterized in that: A carbon steel electrode is fixed on the top of the robotic arm; the wooden board fixing platform is a rectangular steel frame structure, and the table top is provided with a wooden board positioning slot; the camera is installed above the wooden board fixing platform through a universal bracket; an electric shock painting prediction and display module is provided in the control chip, the wooden board surface image is obtained by the camera to extract the wooden board texture parameter matrix, the wooden board humidity parameter matrix is ​​detected by the humidity sensor array, the carbonization key point matrix is ​​calculated according to the wooden board texture parameter matrix and the historical electric shock carbonization data, the shortest path algorithm is used to solve the current conduction path optimization problem and calculate the carbonization texture accompanying degree coefficient, the humidity parameter matrix is ​​input into the carbonization depth intelligent optimization model to predict the carbonization depth of each position and calculate the humidity carbonization accompanying degree coefficient, the preset electric shock position coordinates, electric shock angle, and current intensity are substituted into the carbonization prediction algorithm and combined with the texture carbonization accompanying change point matrix and the humidity carbonization accompanying change point matrix to calculate the carbonization depth prediction value, the predicted carbonization image is generated based on the carbonization depth prediction value and superimposed on the real-time image of the wooden board on the LCD screen.

2. The interactive electric shock wood painting device according to claim 1, characterized in that: The current control module is equipped with a high-voltage power converter and a precision current regulator. The high-voltage power converter converts 220V AC power into 2000V~5000V DC high-voltage power, and the precision current regulator accurately adjusts the high-voltage current to the range of 0.1A~10A; the safety protection module is equipped with an overload protection relay, a short-circuit protection switch and a leakage protection circuit breaker.

3. The interactive electric shock wood painting device according to claim 2, characterized in that: The electric shock painting prediction and display module controls the camera to obtain real-time images of the wooden board surface, grayscales the images and extracts the texture features of the wooden board, and establishes a wooden board texture parameter matrix. The number of matrix rows is the number of image rows, the number of matrix columns is the number of image columns, and the matrix element value represents the texture density at the corresponding position.

4. The interactive electric shock wood painting device according to claim 3, characterized in that: The electric shock painting prediction and display module controls the humidity sensor array to collect humidity values ​​at 16 measurement points on the surface of the wooden board, establishes a humidity distribution matrix, and expands the humidity distribution matrix into a humidity parameter matrix of the same dimension as the texture parameter matrix through a bilinear interpolation algorithm.

5. The interactive electric shock wood painting device according to claim 4, characterized in that: The carbonization key point matrix is ​​calculated based on the wood board texture parameter matrix and historical electric shock carbonization data. Each element in the matrix represents the probability weight of carbonization reaction occurring at the corresponding position, and the weight value ranges from 0 to 1.

6. The interactive electric shock wood painting device according to claim 5, characterized in that: The shortest path algorithm converts the wooden board texture parameter matrix into a weighted graph structure, with each pixel as a node of the graph. The edge weights between adjacent pixels are determined by the texture density difference and humidity difference. The Dijkstra algorithm is used to calculate the shortest current conduction path from the preset electric shock position to each edge point of the wooden board.

7. The interactive electric shock wood painting device according to claim 6, characterized in that: The carbonization texture accompanying degree coefficient is calculated based on the shortest path result and the wood board texture parameter matrix, which indicates the degree of influence of the texture structure at each position on the carbonization process. At the same time, a texture carbonization accompanying change point matrix is ​​generated. The matrix element value of 1 represents a change point, and the value of 0 represents an ordinary point.

8. The interactive electric shock wood painting device according to claim 7, characterized in that: The carbonization depth intelligent optimization model is a skip connection model based on the UNet architecture, which consists of two main parts: an encoder and a decoder. The encoder is responsible for extracting the multi-scale features of the humidity distribution, and the decoder is responsible for reconstructing the carbonization depth prediction map. Detailed information is transmitted between the encoder and decoder through skip connections.

9. The interactive electric shock wood painting device according to claim 8, characterized in that: The feature fusion weight parameters in the carbonization depth intelligent optimization model are dynamically adjusted according to the combined value of the current current intensity, wood board thickness, and preset electric shock angle through the weight adjustment function. The weight adjustment function sets the feature fusion weight parameters based on the comprehensive influencing factors using linear, quadratic, or exponential interpolation methods.

10. The interactive electric shock wood painting device according to claim 9, characterized in that: The carbonization prediction algorithm is based on the current conduction physical model and the chemical reaction mechanism of wood carbonization. The core calculation process includes current density distribution calculation, temperature field analysis, carbonization reaction kinetics modeling and quantitative prediction of carbonization depth.

Images