Additive manufacturing deposition layer morphology measuring device and method based on intelligent image reconstruction

Abstract

The invention discloses an additive manufacturing deposition layer morphology measuring device and method based on intelligent image reconstruction. The method comprises the steps that a molten pool area image and vibration data are collected; the collected molten pool area image and vibration data are input into a depth self-encoder to be processed, and a reconstructed image is generated through the depth self-encoder; the quality and the stability of the reconstructed image are evaluated by using the PSNR and the SSIM; if the image quality does not reach the preset standard, the system calculates an optimal illumination angle and a camera position required for improving the image quality; the angles and the positions of the LED light and the camera are adjusted through the hollow shaft motor, it is ensured that a light path evenly covers the surface of the molten pool, and all possible illumination dark areas are captured; and if the image quality reaches a preset standard, outputting a reconstructed image. By combining the improved monitoring equipment and the deep learning technology, the challenge faced by the deposition layer morphology test in the laser additive manufacturing process is effectively solved, and the test quality and reliability are improved.

Description

Technical Field

[0001] The invention relates to a high-precision measurement device and method for the morphology of a deposited layer in laser additive manufacturing based on intelligent image reconstruction, and belongs to the technical field of laser additive manufacturing. Background Art

[0002] In the laser additive manufacturing process, accurate testing of the surface morphology of the deposited layer is crucial to ensure the manufacturing quality. However, there are a series of challenges in the experimental process. First, there may be dark areas in some areas that cannot be reached by light, which is mainly due to the characteristics of laser irradiation. The propagation of lasers will be affected by the reflection and refraction of the material surface, resulting in some areas that cannot be directly irradiated by light, thus forming dark areas. These dark areas may contain key surface details, such as microstructures, cracks or defects, but cannot be captured by traditional monitoring equipment. Secondly, the vibration of the manufacturing equipment itself and the vibration of the surrounding environment are also an important influencing factor. In the laser additive manufacturing process, the movement of the laser head, the injection of materials, and the operation of other mechanical parts will cause tiny vibrations. These vibrations appear as image jitter and blur in the video data, which in turn affects the stability and clarity of the test data. The vibration phenomenon may be caused by many factors, including the accuracy of the manufacturing equipment, the adhesion of the material, the stability of the workbench, and the interference of the surrounding environment. . Summary of the invention

[0003] Purpose of the invention: In order to solve the problems of insufficient lighting and equipment vibration in the laser additive manufacturing process, the present invention provides an additive manufacturing deposited layer morphology measurement method based on image intelligent reconstruction. By integrating a variable-viewing angle camera and deep autoencoder technology, and combining a vibration elimination method, the influence of light and vibration can be effectively eliminated, thereby achieving high-precision testing of the surface morphology of the deposited layer.

[0004] Technical solution: To achieve the above purpose, the technical solution adopted by the present invention is:

[0005] A method for measuring the morphology of deposited layers in additive manufacturing based on intelligent image reconstruction can achieve fill-in lighting for dark areas during experiments by improving the monitoring equipment for laser metal deposition, thereby ensuring that the surface details of these key areas can also be accurately captured. Secondly, using deep learning technology, the collected video data can be processed and optimized to reduce the impact of vibration, thereby improving the stability and clarity of the test data. The present invention effectively solves the challenges faced by deposited layer morphology testing in the laser additive manufacturing process by combining improved monitoring equipment and deep learning technology, and improves the test quality and reliability. Specifically, it includes the following steps:

[0006] Step 1: Collect images and vibration data of the molten pool area.

[0007] Step 2: Input the collected molten pool area image and vibration data into a deep autoencoder for processing, and generate a reconstructed image through the deep autoencoder. The deep autoencoder includes an encoder, a decoder and a loss function. The encoder uses a convolutional neural network structure to convert the input image data and vibration data into a low-dimensional representation, and the decoder uses the low-dimensional representation to reconstruct a stable and accurate image.

[0008] In step 3, the quality and stability of the reconstructed image are evaluated using PSNR and SSIM. If the image quality does not meet the preset standard, the system will calculate the optimal lighting angle and camera position required to improve the image quality.

[0009] Step 4: Use the hollow shaft motor to adjust the angle and position of the LED light and camera to ensure that the light path evenly covers the surface of the molten pool and captures all possible light and dark areas.

[0010] Step 5: If the image quality reaches the preset standard, the reconstructed image is output.

[0011] Preferably, the loss function is:

[0012]

[0013] Among them, L total is the loss value, λ(t) is the weight parameter dynamically adjusted according to the training process, L recon is the reconstruction error loss, L stablity is the stability loss, η is used to adjust the contribution of the nonlinear interaction term to the total loss, α is the reconstruction error adjustment coefficient, β is the stability adjustment coefficient, γ is the error and stability related adjustment coefficient, δ is the deep feature mutual information adjustment coefficient, P(z|I) is the conditional probability distribution of the latent representation of the original image I after encoding, and P(z|K) is the conditional probability distribution of the latent representation of the reconstructed image K after encoding.

[0014] Preferably, the calculation formula of PSNR is:

[0015]

[0016] Among them, PSNR represents peak signal-to-noise ratio, MAX1 represents the maximum value of the color of the image point, f(V) represents the influence function of the deep autoencoder on the image quality, V is the vibration amplitude data received from the deep autoencoder, I represents the original image, K represents the reconstructed image, m represents the image height, and n represents the image width.

[0017] Preferably, the calculation formula of SSIM is:

[0018]

[0019] Among them, SSIM(x,y) represents the structural similarity of images x and y, μ x is the mean value of image x, μ y is the mean value of image y. x is the standard deviation of image x, σ y is the standard deviation of image y, σ xy is the covariance. c1 and c2 are constants used to maintain stability, g(V) is the mean coefficient in the adjusted denominator, and h(V) is the variance term coefficient in the adjusted denominator.

[0020] Preferably, the encoder includes a vibration feature encoder and an image feature encoder, and the vibration feature encoder includes a first input layer 1, a first RNN layer 1, a first RNN layer 2, a first fully connected layer 1, and a first output feature vector layer connected in sequence. The vibration feature encoder includes a second input layer, a second convolutional layer 1, a second pooling layer 1, a second convolutional layer 2, a second pooling layer 2, a second convolutional layer 3, a second fully connected layer, and a second output feature vector layer connected in sequence.

[0021] Preferably, the decoder includes a vibration feature decoder and an image feature decoder, and the vibration feature decoder includes a third input layer 1, a third fully connected layer 1, a third RNN layer 1, a third RNN layer 2, and a third output layer connected in sequence. The vibration feature decoder includes a fourth input layer, a fourth fully connected layer, a fourth deconvolution layer 1, a fourth upsampling layer 1, a fourth deconvolution layer 2, a fourth upsampling layer 2, a fourth deconvolution layer 3, and a fourth output layer connected in sequence.

[0022] Preferably, according to the evaluation results, the parameters of the codec are adjusted through a feedback adjustment mechanism, and the optimized parameters are used to further improve the image processing process.

[0023] Another object of the present invention is to provide an additive manufacturing deposited layer morphology measurement device based on image intelligent reconstruction, which is used to implement the additive manufacturing deposited layer morphology measurement method based on image intelligent reconstruction, including a laser head, a hollow shaft motor, a measured substrate, an LED lamp, a laser emitter and a remelting molten pool morphology monitoring sensor and a camera system with a variable viewing angle. The camera system with a variable viewing angle includes a side monitoring CCD and a line laser monitoring CCD. The laser head is installed on the top of the equipment, and the hollow shaft motor is installed on the laser head. The line laser monitoring CCD, LED lamp, laser emitter, and side monitoring CCD are installed on the hollow shaft motor. The hollow shaft motor is used to drive the line laser monitoring CCD, LED lamp, laser emitter, and side monitoring CCD to rotate around the laser head.

[0024] Preferably, there are two LED lamps arranged at 180°.

[0025] Preferably, the hollow shaft motor is arranged at the lower part of the laser head and keeps a certain distance from the nozzle.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. Stabilize the image through the deep autoencoder, so that even in the presence of vibration and other interference, it can still output a stable image. Such image output can be directly used for display and further processing, ensuring the stability of the image and the accuracy of processing.

[0028] 2. By integrating a variable-angle camera and deep autoencoder technology, the present invention can eliminate dark areas that cannot be reached by light, and at the same time process vibration interference, thereby obtaining high-quality and stable surface morphology data of the deposited layer, thereby improving the accuracy and reliability of the test.

[0029] 3. By using a variable viewing angle camera, the present invention can obtain image data of the surface of the sediment layer at different angles, providing more comprehensive information for analysis and helping to more accurately understand the morphological characteristics of the sediment layer. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is the front view of the device designed by the present invention.

[0031] Figure 2 It is a side tilt view of the device designed by the present invention.

[0032] Figure 3 It is a structural framework diagram of the autoencoder proposed in the present invention.

[0033] Figure 4 It is a structural framework diagram of the encoder part of the present invention.

[0034] Figure 5 It is a structural framework diagram of the decoder part of the present invention.

[0035] Figure 6 It is a flow chart of the quality assessment module of the present invention.

[0036] Figure 7 It is a flow chart of the feedback regulation mechanism of the present invention. DETAILED DESCRIPTION

[0037] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0038] An additive manufacturing deposited layer morphology measurement device based on image intelligent reconstruction is used to implement the additive manufacturing deposited layer morphology measurement method based on image intelligent reconstruction, including a laser head 100, a hollow shaft motor 200, a substrate to be measured 300, an LED lamp 202, a laser emitter 203, a remelting molten pool morphology monitoring sensor 400, and a camera system with a variable viewing angle. The camera system with a variable viewing angle includes a side monitoring CCD 204 and a line laser monitoring CCD 202. The laser head 100 is installed on the top of the device and is responsible for emitting laser-melted powder to form a molten pool 301 and create a deposited layer 302 on the substrate. The camera system with a variable viewing angle consists of a side monitoring CCD and a laser monitoring CCD, which are connected to the laser head through a hollow shaft motor to capture molten pool images from multiple angles. The hollow shaft motor 200 is mounted on the laser head 100, and the line laser monitoring CCD 201, LED light 202, laser emitter 203, and side monitoring CCD 204 are mounted on the hollow shaft motor 200. The line laser emitter and the line laser monitoring CCD are arranged at 90 degrees to ensure that the line laser is accurately pointed to the molten pool area. The hollow shaft motor 200 is used to drive the line laser monitoring CCD 201, LED light 202, laser emitter 203, and side monitoring CCD 204 to rotate around the laser head 100.

[0039] The lighting device is composed of two LED lights connected to the laser head through a hollow shaft motor. The two LED lights are arranged at 180 degrees to evenly illuminate the molten pool area. An 808 filter is installed in front of the lens, and the lighting light path is directed to the molten pool.

[0040] The hollow shaft motor is located at the lower part of the laser head and keeps a certain distance from the nozzle to adjust the angle and position of the lighting system and the camera system. The hollow shaft motor can drive the lighting equipment, the variable viewing angle camera system and the line laser generator to rotate around the laser head. By improving the laser head lighting equipment, the dark area that cannot be illuminated during the laser additive manufacturing process can be eliminated; the vibration amplitude data during the experiment can be collected; the collected image data can be processed and optimized using a deep autoencoder to reduce the impact of vibration on the test data; and the stabilized image can be monitored in real time and output.

[0041] The angle and position of the two LED lights and variable-angle camera are adjusted by a hollow shaft motor to ensure multi-angle and comprehensive lighting coverage and image capture.

[0042] The back side of the substrate to be measured is provided with a remelting pool morphology monitoring sensor, and also includes a computer, which obtains the vibration amplitude of the camera and the monitoring platform during the monitoring process by subtracting the height of the deposited layer after cooling.

[0043] During the manufacturing process, the hollow shaft motor 200 can drive the LED lamp 202 , the side monitoring CCD 204 , the line laser monitoring CCD 202 , and the laser emitter 203 to rotate around the laser head 100 .

[0044] The remelting molten pool morphology monitoring sensor adopts an ultrasonic detector, which is located on the back of the substrate and is used to detect the three-dimensional morphology of the molten pool, thereby realizing high-precision deposition layer morphology testing.

[0045] The camera system of the variable-angle visual measurement device can dynamically adjust the position and angle of the camera to capture all possible light and dark areas as well as detailed information.

[0046] During the test, the laser head is first started to form the deposited layer, and the lighting system and camera system are activated at the same time. The lighting system is adjusted by the hollow shaft motor to ensure that the molten pool area is evenly illuminated and reduce the dark area problem caused by insufficient light. The camera system captures images from different angles for subsequent image analysis.

[0047] The captured image data and real-time measured vibration data are input to a deep autoencoder. The autoencoder first cleans and normalizes the data through a preprocessing layer, and then processes it through the encoder part, including the encoding of vibration and image features. After feature fusion in the latent space, the decoder reconstructs the image and vibration features to produce a stabilized image.

[0048] The quality assessment module evaluates the quality of the reconstructed image and evaluates whether the current lighting conditions are adequate based on the image quality indicators output by the autoencoder. If the image quality does not meet the preset standard, the optimal lighting angle and camera position required to improve the image quality are calculated. Based on the evaluation results, the feedback adjustment mechanism adjusts the parameters of the codec, and the optimized parameters are used to further improve the image processing process.

[0049] A method for measuring the morphology of deposited layers in additive manufacturing based on intelligent image reconstruction, such as Figure 3-7 As shown, the specific steps include:

[0050] Step 1: Collect images and vibration data of the molten pool area.

[0051] Step 2: Input the collected molten pool area image and vibration data into the deep autoencoder for processing, and generate a reconstructed image through the deep autoencoder.

[0052] The deep autoencoder optimizes the test accuracy of the deposited layer morphology by integrating image data and vibration data. The input layer of the autoencoder receives image data and vibration amplitude. The stable image generation process of the deep autoencoder includes processing the input image data and vibration amplitude using a pre-trained deep autoencoder and outputting a stabilized image.

[0053] During real-time monitoring, the current image data and vibration amplitude are continuously acquired and input into the deep autoencoder, which directly outputs the stabilized image for display and further processing.

[0054] The deep autoencoder includes an encoder, a decoder, and a loss function. The encoder uses a convolutional neural network structure to convert the input image data and vibration data into a low-dimensional representation. This process focuses on capturing complex image features to ensure accurate encoding of the data. The encoder part consists of two submodules: a convolutional neural network for extracting image features, and a recurrent neural network for processing vibration data to capture the vibration features of the time series. These features are combined in the feature fusion layer to ensure the effective integration of image and vibration information. Figure 4 As shown, the encoder includes a vibration feature encoder and an image feature encoder, and the vibration feature encoder includes a first input layer 1, a first RNN layer 1, a first RNN layer 2, a first fully connected layer 1, and a first output feature vector layer connected in sequence. The vibration feature encoder includes a second input layer, a second convolutional layer 1, a second pooling layer 1, a second convolutional layer 2, a second pooling layer 2, a second convolutional layer 3, a second fully connected layer, and a second output feature vector layer connected in sequence.

[0055] The decoder uses these low-dimensional representations to reconstruct a stable and accurate image, and uses advanced image restoration techniques to ensure the quality and detail fidelity of the reconstructed image. The decoder uses a transposed convolutional network to reconstruct the image from the fused features, and adjusts the reconstruction strategy based on the vibration data to suppress the distortion caused by vibration. Figure 5 As shown, the decoder includes a vibration feature decoder and an image feature decoder, and the vibration feature decoder includes a third input layer 1, a third fully connected layer 1, a third RNN layer 1, a third RNN layer 2, and a third output layer connected in sequence. The vibration feature decoder includes a fourth input layer, a fourth fully connected layer, a fourth deconvolution layer 1, a fourth upsampling layer 1, a fourth deconvolution layer 2, a fourth upsampling layer 2, a fourth deconvolution layer 3, and a fourth output layer connected in sequence.

[0056] The loss function includes mean square error to evaluate the difference between the reconstructed image and the original image, and total variation regularization to enhance the stability of the image, thereby generating stable and realistic image data. In this way, the autoencoder can effectively restore clear and high-quality images from images affected by vibration, improving the accuracy and reliability of sedimentary layer morphology testing.

[0057] In the process of generating stable images, in order to optimize this process, a comprehensive loss function is designed, including reconstruction error loss and stability loss. The reconstruction error is evaluated by the mean square error, focusing on minimizing the difference between the reconstructed image and the original image; the stability loss is measured by the total variation of the image, aiming to reduce image jitter and blur caused by vibration, thereby ensuring smooth output of the image. It aims to optimize the parameters of the deep autoencoder and generate stable and realistic image data. The loss function consists of two parts: reconstruction error loss L recon and stability loss L stablity In order to fully consider the independence and mutual influence of these two losses, a dynamic weight adjustment strategy is introduced: L total =λ(t)L recon +(1-λ(t))L stablity , where λ(t) is a weight parameter that is dynamically adjusted according to the training process. In order to further enhance the self-adjustment capability of the model, the present invention also introduces a combination containing nonlinear terms: L interation =αL recon 2 +βL stablity 2 +γL recon L stablity Among them, α, β, and γ are adjustment coefficients used to control the influence of each loss term in the autoencoder, so that the model can enhance stability while ensuring the reconstruction quality, and is optimized through experiments. Finally, the weight adjustment is combined with the nonlinear interaction effect to form a comprehensive loss function:

[0058] L total =λ(t)L recon +(1-λ(t))L stablity +η(αL recon 2 +βL stablity 2 +γL recon L stablity ),

[0059] Among them, η is used to adjust the contribution of nonlinear interaction terms to the total loss, which is adjusted through experiments.

[0060] In addition, in order to further improve the quality of image reconstruction, deep feature interaction information is introduced: Among them, P(z|I) and P(z|K) are the conditional probability distributions of the potential representations of the original image I and the reconstructed image K after encoding, respectively, so as to obtain the final loss function:

[0061]

[0062]

[0063] Among them, L total is the loss value, λ(t) is the weight parameter dynamically adjusted according to the training process, L recon is the reconstruction error loss, L stablity is the stability loss, η is used to adjust the contribution of the nonlinear interaction term to the total loss, which is adjusted through experiments, α is the reconstruction error adjustment coefficient, β is the stability adjustment coefficient, γ is the error and stability related adjustment coefficient, δ is the deep feature mutual information adjustment coefficient, P(z|I) is the conditional probability distribution of the latent representation of the original image I after encoding, and P(z|K) is the conditional probability distribution of the latent representation of the reconstructed image K after encoding.

[0064] By optimizing this composite loss function, the model is able to generate more stable and high-quality image data. This approach ensures that the generated images are both realistic and stable, effectively reflecting the detailed features of the sedimentary surface.

[0065] The dataset of the deep autoencoder comes from real-time monitoring data and image acquisition during the laser metal deposition process. Specifically, during the processing, the monitoring equipment records the images of the molten pool and the deposited layer. At the same time, the vibration amplitude can be indirectly obtained by performing a differential operation on the height of the deposited layer after cooling. The captured images are paired with the corresponding vibration amplitudes, and the real-time measured vibration amplitudes are annotated for each set of image data. Finally, the processed image data and the corresponding vibration data are integrated into a training set for the training of the deep autoencoder.

[0066] The specific steps to eliminate vibration are: first, use a variable-angle camera to capture images from multiple angles, and synchronously collect vibration data through a vibration sensor. After preprocessing, these data are input into a deep autoencoder, where the encoder part converts the image and vibration data into low-dimensional feature representations. In the decoder stage, these low-dimensional representations are used to reconstruct the image, and a comprehensive loss function is applied to optimize the stability and quality of the reconstructed image. This process includes dynamic weight adjustment and nonlinear dependencies to finely control the trade-off between image reconstruction and stability.

[0067] Step 3, such as Figure 6 and 7 , using PSNR and SSIM to evaluate the quality and stability of the reconstructed image. If the image quality does not meet the preset standard, the system will calculate the optimal lighting angle and camera position required to improve the image quality.

[0068] The specific steps to eliminate the dark areas of illumination are as follows: First, the image of the molten pool area is captured by a variable-angle camera system, and the relevant vibration data is collected by a line laser. Subsequently, these images and vibration data are input into the autoencoder for processing, and the autoencoder generates a reconstructed image and uses PSNR and SSIM to evaluate its quality and stability. Based on the image quality indicators output by the autoencoder, the system evaluates the current lighting conditions. If the image quality does not meet the preset standards, the system will calculate the optimal lighting angle and camera position required to improve the image quality. Then, the angle and position of the LED light and camera are adjusted by the hollow shaft motor to ensure that the light path evenly covers the surface of the molten pool and captures all possible dark areas of illumination. This process is repeated until the image quality is optimal. This feedback control mechanism with integrated autoencoder not only improves the automation of the test, but also ensures the accuracy and repeatability of the deposition layer morphology test under various operating conditions.

[0069] The calculation formula of PSNR is:

[0070]

[0071]

[0072] Among them, PSNR represents peak signal-to-noise ratio, MAX1 represents the maximum value of the color of the image point, f(V) represents the impact function of the deep autoencoder on the image quality to automatically adjust the evaluation parameters, V is the vibration amplitude data received from the deep autoencoder, I represents the original image, K represents the reconstructed image, m represents the image height, and n represents the image width.

[0073] Preferably, the calculation formula of SSIM is:

[0074]

[0075] Among them, SSIM(x,y) represents the structural similarity of images x and y, μ x is the mean value of image x, μ y is the mean value of image y. x is the standard deviation of image x, σ y is the standard deviation of image y, σ xy is the covariance. c1 and c2 are constants used to maintain stability, g(V) is the coefficient for adjusting the mean in the denominator, and h(V) is the coefficient for adjusting the variance term in the denominator. g(V) and h(V) are used to adjust the mean and variance terms in the denominator, reflecting the impact of vibration on image stability, which is obtained by statistical analysis of image data under different vibration conditions.

[0076] Step 4: Use the hollow shaft motor to adjust the angle and position of the LED light and camera to ensure that the light path evenly covers the surface of the molten pool and captures all possible light and dark areas.

[0077] Step 5: If the image quality reaches the preset standard, the reconstructed image is output.

[0078] Step 6: According to the evaluation results, the parameters of the codec are adjusted through a feedback adjustment mechanism, and the optimized parameters are used to further improve the image processing process.

[0079] First, multiple images are collected using a variable-view camera, and the corresponding vibration amplitude data is obtained through a vibration sensor. Then, these image data are input into a deep autoencoder for data preprocessing to ensure data quality and consistency. The encoder part includes a vibration feature encoder and an image feature encoder to encode the input data into a low-dimensional representation. Then, feature fusion and dimensionality reduction are performed in the latent space to ensure that the data representation contains vibration and image features. The decoder part includes a vibration feature decoder and an image feature decoder to reconstruct image and vibration features from the low-dimensional representation. In the process of generating a stable image, the quality assessment module uses the PSNR and SSIM indicators to evaluate the quality of the reconstructed image. The feedback adjustment mechanism then adjusts the parameters of the encoder and decoder based on the evaluation results. This process generates stable and realistic image data by optimizing the parameters of the deep autoencoder, minimizing the reconstruction error loss and stability loss. The final stable image data can be used for additive manufacturing deposit layer morphology measurement based on image intelligent reconstruction, ensuring that the test data has high clarity and stability and accurately reflects the morphological characteristics of the deposit layer surface.

[0080] The purpose of improving the lighting equipment in the monitoring equipment is to eliminate the dark areas that cannot be illuminated during the laser metal deposition process; a variable-angle camera is installed in the test device to obtain surface image data of the deposited layer from multiple angles. The amplitude of vibration during the experiment is tested, and a deep autoencoder is used to create a reverse arc to keep the corrected image horizontal and solve the vibration problem during the experiment. By solving these problems, the surface morphology of the deposited layer can be tested with high precision.

[0081] This embodiment solves the problem of dark areas under laser irradiation by improving the lighting system of the monitoring equipment, and uses deep learning technology, especially deep autoencoders, to process and optimize the collected video data to reduce the impact of vibration on the test data and improve the stability and clarity of the test data. The advantages of this invention include solving the problems of insufficient lighting and vibration interference and improving the quality and reliability of test data.

[0082] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for measuring the morphology of deposited layers in additive manufacturing based on intelligent image reconstruction, characterized in that: The following steps are involved: Step 1, collecting images and vibration data of the molten pool area; Step 2: Input the collected molten pool area image and vibration data into a deep autoencoder for processing, and generate a reconstructed image through the deep autoencoder; the deep autoencoder includes an encoder, a decoder and a loss function, the encoder uses a convolutional neural network structure to convert the input image data and vibration data into a low-dimensional representation, and the decoder uses the low-dimensional representation to reconstruct a stable and accurate image; Step 3, and use PSNR and SSIM to evaluate the quality and stability of the reconstructed image; If the image quality does not meet the preset standard, the system will calculate the optimal lighting angle and camera position required to improve the image quality; Step 4: Use the hollow shaft motor to adjust the angle and position of the LED light and camera to ensure that the light path evenly covers the surface of the molten pool and captures all possible light and dark areas; Step 5: If the image quality reaches the preset standard, the reconstructed image is output.

2. The method for measuring the morphology of deposited layers in additive manufacturing based on intelligent image reconstruction according to claim 1, characterized in that: The loss function is: Among them, L total is the loss value, λ(t) is the weight parameter dynamically adjusted according to the training process, L recon is the reconstruction error loss, L stablity is the stability loss, η is used to adjust the contribution of the nonlinear interaction term to the total loss, α is the reconstruction error adjustment coefficient, β is the stability adjustment coefficient, γ is the error and stability related adjustment coefficient, δ is the deep feature mutual information adjustment coefficient, P(z|I) is the conditional probability distribution of the latent representation of the original image I after encoding, and P(z|K) is the conditional probability distribution of the latent representation of the reconstructed image K after encoding.

3. The method for measuring the deposited layer morphology of additive manufacturing based on image intelligent reconstruction according to claim 2, characterized in that: The calculation formula of PSNR is: Among them, PSNR represents peak signal-to-noise ratio, MAX1 represents the maximum value of the color of the image point, f(V) represents the influence function of the deep autoencoder on the image quality, V is the vibration amplitude data received from the deep autoencoder, I represents the original image, K represents the reconstructed image, m represents the image height, and n represents the image width.

4. The method for measuring the morphology of deposited layers in additive manufacturing based on intelligent image reconstruction according to claim 3, characterized in that: The calculation formula of SSIM is: Among them, SSIM(x,y) represents the structural similarity of images x and y, μ x is the mean value of image x, μ y is the mean value of image y; σ x is the standard deviation of image x, σ y is the standard deviation of image y, σ xy is the covariance; c1 and c2 are constants used to maintain stability, g(V) is the mean coefficient in the adjusted denominator, and h(V) is the variance term coefficient in the adjusted denominator.

5. The method for measuring the deposited layer morphology in additive manufacturing based on intelligent image reconstruction according to claim 4, characterized in that: The encoder includes a vibration feature encoder and an image feature encoder. The vibration feature encoder includes a first input layer 1, a first RNN layer 1, a first RNN layer 2, a first fully connected layer 1, and a first output feature vector layer connected in sequence; the vibration feature encoder includes a second input layer, a second convolutional layer 1, a second pooling layer 1, a second convolutional layer 2, a second pooling layer 2, a second convolutional layer 3, a second fully connected layer, and a second output feature vector layer connected in sequence.

6. The method for measuring the deposited layer morphology in additive manufacturing based on intelligent image reconstruction according to claim 5, characterized in that: The decoder includes a vibration feature decoder and an image feature decoder. The vibration feature decoder includes a third input layer 1, a third fully connected layer 1, a third RNN layer 1, a third RNN layer 2, and a third output layer connected in sequence; the vibration feature decoder includes a fourth input layer, a fourth fully connected layer, a fourth deconvolution layer 1, a fourth upsampling layer 1, a fourth deconvolution layer 2, a fourth upsampling layer 2, a fourth deconvolution layer 3, and a fourth output layer connected in sequence.

7. The method for measuring the deposited layer morphology in additive manufacturing based on intelligent image reconstruction according to claim 6, characterized in that: Based on the evaluation results, the parameters of the codec are adjusted through a feedback regulation mechanism, and the optimized parameters are used to further improve the image processing process.

8. An additive manufacturing deposition layer morphology measurement device based on image intelligent reconstruction, characterized in that: A method for measuring the morphology of a deposited layer in additive manufacturing based on intelligent image reconstruction as claimed in claim 1, comprising a laser head (100), a hollow shaft motor (200), a substrate to be measured (300), an LED lamp (202), a laser emitter (203), a remelting pool morphology monitoring sensor (400), and a camera system with a variable viewing angle, wherein the camera system with a variable viewing angle comprises a side monitoring CCD (204) and a line laser monitoring CCD (202); the laser head (100) is mounted on the top of the device, the hollow shaft motor (200) is mounted on the laser head (100), the line laser monitoring CCD (201), the LED lamp (202), the laser emitter (203), and the side monitoring CCD (204) are mounted on the hollow shaft motor (200), and the hollow shaft motor (200) is used to drive the line laser monitoring CCD (201), the LED lamp (202), the laser emitter (203), and the side monitoring CCD (204) to rotate around the laser head (100).

9. The device for measuring the deposited layer morphology of additive manufacturing based on intelligent image reconstruction according to claim 8, characterized in that: There are two LED lamps (202), which are arranged at 180 degrees.

10. The device for measuring the deposited layer morphology of additive manufacturing based on intelligent image reconstruction according to claim 9, characterized in that: The hollow shaft motor is arranged at the lower part of the laser head and keeps a certain distance from the nozzle.

Images