Visualization system for monitoring production of hydraulic machine based on full polyp technology

Abstract

The invention relates to the technical field of hydraulic machines, in particular to a visualization system for monitoring production of a hydraulic machine based on a full polyp technology, which comprises a monitoring system and a hydraulic machine, and is characterized in that the monitoring system comprises an image acquisition device, an upper computer in communication connection with the image acquisition device, and a laser generator in communication connection with the upper computer; the image acquisition equipment is used for acquiring a holographic image with interference fringes of the whole hydraulic machine, the upper computer is used for processing the holographic image with interference fringes acquired in the image acquisition equipment to generate a three-dimensional holographic image of the hydraulic machine, and the laser generator projects the processed three-dimensional holographic image of the hydraulic machine into space. According to the invention, the internal structure and the operation state of the hydraulic machine can be accurately monitored, and the complex structure and the dynamic change in the hydraulic machine can be visually displayed.

Description

Technical Field

[0001] The present invention relates to the technical field of hydraulic presses, and particularly to a visualization system for monitoring the production of hydraulic presses based on holographic imaging technology. Background Art

[0002] Currently, in industrial production, as an important processing equipment, the stability of the performance and the accuracy of the control of hydraulic presses are directly related to production efficiency and product quality. However, there are some deficiencies in traditional hydraulic press monitoring systems, such as low monitoring accuracy, incomplete data collection, and untimely fault warning. These problems may cause faults during the operation of hydraulic presses, affecting production efficiency and product quality. Traditional hydraulic press monitoring systems mainly rely on parameters such as pressure, temperature, and speed collected by sensors. Although these parameters can reflect part of the operating state of hydraulic presses, they cannot intuitively display the complex internal structure and dynamic changes of hydraulic presses.

[0003] Therefore, the present invention proposes a visualization system for monitoring the production of hydraulic presses based on holographic imaging technology, which can achieve precise monitoring of the internal structure and operating state of hydraulic presses and more intuitively display the complex internal structure and dynamic changes of hydraulic presses. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems existing in the prior art, and to propose a visualization system for monitoring the production of hydraulic presses based on holographic imaging technology.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions: A visualization system for monitoring the production of hydraulic presses based on holographic imaging technology, including a monitoring system and a hydraulic press. The monitoring system includes an image acquisition device, a host computer communicatively connected to the image acquisition device, and a laser generator communicatively connected to the host computer. The image acquisition device is used to acquire the holographic image of the entire hydraulic press with interference fringes. The host computer is used to process the holographic image with interference fringes acquired by the image acquisition device and generate a three-dimensional holographic image of the hydraulic press. The laser generator projects the three-dimensional holographic image of the hydraulic press into space.

[0006] Further, the host computer includes an image imaging module, a phase extraction module, and a three-dimensional reconstruction module. The image imaging module processes the holographic image with interference fringes acquired by the image acquisition device. The phase extraction module extracts the information in the digital holographic image processed by the image imaging module. The three-dimensional reconstruction module reconstructs the holographic two-dimensional image in the phase extraction module and generates a three-dimensional holographic image of the hydraulic press.

[0007] Further, the image acquisition device is a laser camera, which generates a laser beam. The laser beam is divided into an object beam and a reference beam. After the object beam and the reference beam are superimposed, a holographic image with alternating bright and dark interference fringes is generated.

[0008] Further, the host computer further includes a digital sensor, which records the information of the interference fringes in the holographic image with interference fringes. The image imaging module processes the information of the interference fringes in the digital sensor to form a stable digital holographic image.

[0009] Further, the digital holographic image includes phase information and amplitude information. After the phase extraction module extracts the phase information and amplitude information, a holographic two-dimensional image is obtained. Then, the holographic two-dimensional image with phase information and amplitude information is transmitted into the three-dimensional reconstruction module.

[0010] Further, the extraction formula for the amplitude information extracted by the phase extraction module is: , where A(x, y) represents the amplitude information, representing the intensity of the light field at the position (x, y); U(x, y) represents the complex amplitude distribution, representing the complex value of the light field at the spatial position (x, y); Re[U(x, y)] represents the real part of the complex amplitude, and Im[U(x, y)] represents the imaginary part of the complex amplitude. The extraction formula for extracting the phase information is: , where represents the phase information, representing the phase angle of the light field at the position (x, y); represents the imaginary part of the complex amplitude, representing the sine component of the light field; represents the real part of the complex amplitude, representing the cosine component of the light field.

[0011] Further, the three-dimensional reconstruction module uses the holographic two-dimensional image to construct a three-dimensional holographic image of the hydraulic press with higher accuracy. The three-dimensional reconstruction module includes a model-driven deep learning module, and the model-driven deep learning module includes a dataset training module, a feature extraction module, and an image enhancement module.

[0012] Further, the dataset training module pre-inserts the pre-trained data related to the hydraulic press with phase information and amplitude information to obtain a holographic three-dimensional image. The feature extraction module automatically extracts features from the holographic three-dimensional image to obtain a holographic three-dimensional image one. The image enhancement module is used to improve the resolution and clarity of the holographic three-dimensional image one to obtain a three-dimensional holographic image of the hydraulic press.

[0013] Furthermore, the feature extraction module uses a convolutional neural network to extract features. The convolutional neural network can automatically learn and extract features from raw data. In holographic three-dimensional image processing, the convolutional neural network extracts local features through the combination of convolutional layers, pooling layers, and fully connected layers, and gradually abstracts high-level features through a hierarchical structure.

[0014] Furthermore, the three-dimensional holographic image of the hydraulic press projects a three-dimensional holographic image of the hydraulic press through a laser generator.

[0015] Compared with the existing technology, a visualization system for monitoring the production of a hydraulic press based on holographic imaging technology provided by the present invention has the following advantages: 1. Accurately monitor the internal structure and dynamic changes of the hydraulic press. The holographic image technology can intuitively display the complex internal structure and dynamic changes of the hydraulic press, providing a more comprehensive monitoring perspective to help operators and maintenance personnel better understand the operating status of the equipment. 2. Improve the accuracy of fault diagnosis. Through the three-dimensional holographic image of the hydraulic press, potential fault points can be more intuitively discovered, maintenance and repair can be carried out in advance, equipment downtime can be reduced, and production efficiency can be improved. 3. Enhance the training effect of operators. The three-dimensional holographic image of the hydraulic press can be used as a training tool to help new employees quickly familiarize themselves with the internal structure and operating principle of the equipment, and improve their operation skills and maintenance capabilities. 4. Improve the transparency of the production process, provide more intuitive production data and equipment status information for management, facilitate production planning and resource allocation, and optimize the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic diagram of the framework structure of the present invention; Figure 2 is a schematic diagram of the framework structure of the monitoring system of the present invention; Figure 3 is a schematic diagram of the working process of the model-driven deep learning module of the present invention; Figure 4 is a schematic diagram of the working process of the present invention; In the figure: 1. Image acquisition device, 2. Host computer, 21. Image imaging module, 22. Phase extraction module, 23. Three-dimensional reconstruction module, 24. Model-driven deep learning module, 25. Dataset training module, 26. Feature extraction module, 27. Image enhancement module, 3. Laser generator, 4. Hydraulic press. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] 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. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0019] Embodiment 1, as Figures 1-4 shown, a visualization system for monitoring the production of a hydraulic press based on holographic technology includes a monitoring system and a hydraulic press 4. The monitoring system includes an image acquisition device 1, a host computer 2 communicatively connected to the image acquisition device 1, and a laser generator 3 communicatively connected to the host computer 2. The image acquisition device 1 is used to acquire a holographic image of the entire hydraulic press 4 with interference fringes. The host computer 2 is used to process the holographic image with interference fringes acquired by the image acquisition device 1. The laser generator 3 projects the processed three-dimensional holographic image of the hydraulic press into space.

[0020] Through this technical solution, the image acquisition device 1 in the monitoring system acquires the information of the body and the interior of the hydraulic press 4. The acquired holographic image with interference fringes is then transmitted to the host computer 2. After being processed by the host computer 2, a high-quality three-dimensional holographic image of the hydraulic press is formed. Then, through the laser generator 3, the three-dimensional holographic image of the hydraulic press is projected out completely and clearly. Not only can the external shape of the entire hydraulic press 4 be seen, but also the internal image of the hydraulic press 4 can be intuitively seen. For operators and maintenance personnel, they can clearly understand the internal and external shapes of the hydraulic press 4 and the operating state of the hydraulic press 4, and timely discover potential problems.

[0021] The hydraulic press 4 includes a hydraulic press 4 body composed of an upper beam, a slider, a lower beam, a main cylinder for sliding the slider, etc., an electrical control system, and an oil circuit system connecting the main cylinder. The oil circuit system includes various hydraulic valves. It can be understood that the hydraulic press 4 includes but is not limited to four-column hydraulic presses, frame hydraulic presses, and single-column hydraulic presses.

[0022] Embodiment 2: The image acquisition device 1 is a laser camera. The laser camera generates a laser beam, which is divided into an object beam and a reference beam. After the object beam and the reference beam are superimposed, interference fringes with alternating light and dark are generated. After the digital information of the hydraulic press 4 is constructed by the computer imaging module, it is transmitted into the phase extraction module 22.

[0023] In this embodiment, the image acquisition device 1 is installed around the hydraulic press 4. In order to collect more detailed images of the hydraulic press 4, multiple image acquisition devices 1 can be installed. The image acquisition device 1 can use devices that can collect images, such as a high-definition camera or a laser camera. Since the laser camera has the function of penetrating objects, the shooting effect is relatively good, and a laser camera is used in this embodiment.

[0024] The laser camera first emits a laser beam through the built-in laser generator 3. The laser generator 3 is a prior art. The laser beam can be monochromatic or multi-color, with high directivity and monochromaticity. The laser beam can propagate in the air or through transmission media such as optical fibers. Then, the propagation and adjustment of the light are carried out. After the laser beam is emitted from the laser camera, through the focusing effect of the lens, it becomes a ray. This ray can be transformed, adjusted, and split through the action of a series of optical elements (such as a laser beam expander, a beam splitter, etc.) to finally form a suitable irradiation light. After the irradiation light irradiates the object to be photographed, phenomena such as scattering, reflection, and absorption will occur. The object to be photographed will have different reactions to the incident light. Then, the reflected beam is received. The receiver built in the laser camera will receive the reflected light and record the time and intensity of the reflected beam. Since the propagation speed of the laser beam is known, the distance between the laser beam and the surface of the hydraulic press 4 can be calculated according to the time difference. By calculating this distance, the distance between the laser camera and the hydraulic press 4 can be appropriately placed.

[0025] The laser beam is further divided into an object beam and a reference beam. After the object beam and the reference beam are superimposed, interference occurs to form a holographic image with interference fringes.

[0026] When the laser camera has completely collected the holographic image with interference fringes of the hydraulic press 4 inside and outside, this holographic image with interference fringes is then transmitted to the host computer 2, and the host computer 2 processes it into a three-dimensional holographic image of the hydraulic press that is more intuitive to view. The connection between the laser camera and the host computer 2 can use a serial connection, an Ethernet connection, or a wireless network connection. Using an Ethernet connection or a wireless network connection can reduce the layout of the lines.

[0027] Embodiment 3, as Figures 1-2As shown, the host computer 2 includes an image imaging module 21, a phase extraction module 22, and a three-dimensional reconstruction module 23. The image imaging module 21 processes the holographic image with interference fringes collected by the image acquisition device 1. The phase extraction module 22 extracts the information from the digitized holographic image processed by the image imaging module 21. The three-dimensional reconstruction module 23 reconstructs the holographic two-dimensional image in the phase extraction module 22 and generates a three-dimensional holographic image of the hydraulic press.

[0028] The host computer 2 further includes a digital sensor. The digital sensor records the interference fringe information in the holographic image with interference fringes. The image imaging module 21 processes the interference fringe information in the digital sensor. The information includes the intensity distribution of the interference fringes, making it a stable digitized holographic image.

[0029] Specifically, the image acquisition device 1 collects a holographic image with interference fringes. The interference fringes directly record the intensity distribution of the interference fringes through the digital sensor. Then the image imaging module 21 processes the complex amplitude distribution of the reconstructed object light field recorded in the digital sensor, including intensity and phase. The reconstructed digitized holographic image of the hydraulic press 4 can be displayed in real time on the screen of the host computer 2. During the process of the image imaging module 21 generating the digitized holographic image, the Fresnel diffraction method can be used for calculation. Its calculation formula is: , where U(x, y) represents the complex amplitude distribution of the digitized holographic image, and the complete wavefront information of the object light field can be recovered from the holographic image with interference fringes; represents the phase accumulation after the light wave propagates a distance z; i is the imaginary unit, λ is the light wave wavelength, which determines the scale of the diffraction phenomenon; z is the propagation distance from the holographic image with interference fringes to the reconstruction plane, is used as a normalization factor to adjust the amplitude and phase of the complex amplitude; represents all positions on the plane of the holographic image with interference fringes for double integration; represents the complex amplitude distribution of the plane of the holographic image with interference fringes, which is composed of the interference fringe intensity and the complex conjugate of the reference light wave amplitude; represents the point from the plane of the holographic image with interference fringes to the point on the plane of the digitized holographic image the square of the spatial distance; The integration operation superimposes the contributions of all points to generate the complex amplitude distribution of the digitized holographic image; Finally, through the calculation of the image imaging module 21, a stable digital holographic image is generated from the holographic image with interference fringes collected by the image acquisition device 1.

[0030] In Embodiment 4, the digital holographic image includes phase information and amplitude information. After the phase extraction module 22 extracts the phase information and amplitude information, a holographic two-dimensional image is obtained, and then the holographic two-dimensional image with phase information and amplitude information is transmitted into the three-dimensional reconstruction module 23.

[0031] The extraction formula for the amplitude information extracted by the phase extraction module 22 is: , The extraction formula for the phase information extracted by the phase extraction module 22 is: , where A(x, y) represents the amplitude information, representing the intensity of the light field at the position (x, y); U(x, y) represents the complex amplitude distribution of the digital holographic image, representing the complex value of the light field at the spatial position (x, y); represents the phase information, representing the phase angle of the light field at the position (x, y); represents the imaginary part of the complex amplitude, representing the sine component of the light field; represents the real part of the complex amplitude, representing the cosine component of the light field. By calculating the square root of the sum of the squares of the real part of the complex amplitude and the imaginary part of the complex amplitude, the intensity information of the light field is obtained. By calculating the ratio of the imaginary part of the complex amplitude to the real part of the complex amplitude and taking the arctangent, the phase information of the light field is obtained.

[0032] The three-dimensional reconstruction module 23 is set in the host computer 2 in order to obtain a three-dimensional image with better quality.

[0033] The three-dimensional reconstruction module 23 constructs a three-dimensional holographic image of the hydraulic press with higher precision using the holographic two-dimensional image. The three-dimensional reconstruction module 23 further includes a model-driven deep learning module 24, and the model-driven deep learning module 24 includes a data set training module 25, a feature extraction module 26, and an image enhancement module 27.

[0034] As Figure 3 shown, the data set training module 25 pre-inserts the relevant data of the trained holographic two-dimensional image of the hydraulic press with phase information and amplitude information to obtain a holographic three-dimensional image. The feature extraction module 26 automatically extracts features from the holographic three-dimensional image to obtain a holographic three-dimensional image one. The image enhancement module 27 is used to improve the resolution and clarity of the holographic three-dimensional image one to obtain a three-dimensional holographic image of the hydraulic press.

[0035] The dataset training module 25 trains the holographic two-dimensional images of the hydraulic press through deep learning. The large-scale training dataset contains the holographic two-dimensional images of the hydraulic press and the corresponding annotated holographic two-dimensional images, such as data on categories, boundaries, etc., and feeds them into a convolutional neural network for end-to-end supervised training, using the mean squared error (MSE) loss function for training.

[0036] The expression of the mean squared error (MSE) loss function can be expressed as , where N is the number of samples, x i is the true value of the i-th sample, and y i is the predicted value of the i-th sample. The mean squared error (MSE) loss function measures the overall accuracy of the model prediction by calculating the squared value of the difference between the true value of each sample and the predicted value of each sample, and then summing and averaging these squared values. The squared values ensure that all differences are positive, and larger differences have greater weights.

[0037] The dataset training module 25 synthesizes the holographic two-dimensional images into holographic three-dimensional images, and the feature extraction module 26 uses a convolutional neural network to extract features. The convolutional neural network can automatically learn and extract features from the original data. In the processing of holographic three-dimensional images, the convolutional neural network extracts local features through the combination of convolutional layers, pooling layers, and fully connected layers, and gradually abstracts high-level features through a hierarchical structure.

[0038] The feature extraction module 26 is mainly implemented through the combination of convolutional layers, pooling layers, and fully connected layers. The neurons between the convolutional layers, pooling layers, and fully connected layers are connected. The calculation formula for the convolutional neural network to process holographic three-dimensional images is as follows: , where I is the input holographic three-dimensional image, K is the convolutional kernel, (x, y) is a position on the output feature map of the holographic three-dimensional image, represents the convolution operation, I(m, n) represents the pixel value of the input image at position (m, n); K(x - m, y - n) represents the value of the convolutional kernel at position (x - m, y - n), where (m, n) are the coordinates of the convolutional kernel and (x, y) are the coordinates of the output image, represents the summation operation of all elements of the convolutional kernel, that is, the double summation of m and n.

[0039] In this embodiment, the input image is a holographic three-dimensional image, and the output image is a holographic three-dimensional image one. The convolution operation is a core part of the convolutional neural network. It is a mathematical operation that can be used to extract local features of the input image. We use the convolution operation to extract features in the holographic three-dimensional image. The specific operation process is as follows: First step, place the convolution kernel at the upper left corner of the holographic three-dimensional image and cover an area (the same size as the convolution kernel). Second step, calculate the dot product, which is the sum of the element-by-element multiplications of the convolution kernel and the covered area of the holographic three-dimensional image. This sum is the value of the feature of the extracted holographic three-dimensional image at the current position. Third step, according to the set stride, slide the convolution kernel to the right (or down) by a fixed distance, and repeat this step until the entire holographic three-dimensional image is covered. Fourth step, generate the feature map of the holographic three-dimensional image. Fifth step, repeat the second to fourth steps until the convolution kernel covers the entire holographic three-dimensional image, generating a complete feature map of the holographic three-dimensional image.

[0040] For example, assume the input image is a holographic three-dimensional image I with a size of m×n (m rows and n columns), and a convolution kernel with a size of K×L. The output of the convolution operation is a feature map of the holographic three-dimensional image, and its size may depend on the size of the input image, the size of the convolution kernel, as well as the stride and padding strategy. Its calculation expression is: , O(x,y) is the value of the feature map of the holographic three-dimensional image at position (x,y), I(x+m,y+n) is the value of the holographic three-dimensional image at position (x+m,y+n), and K(a - m,b - n) is the value of the convolution kernel at position (a - m,b - n), where a and b are the central indices of the convolution kernel respectively.

[0041] First, sum over the outer layer (sum over m), and the range of m is from -a to a, where a is half of the width of the convolution kernel K. This range ensures that the convolution kernel completely covers the input holographic three-dimensional image.

[0042] Then, sum over the inner layer (sum over n), and the range of n is from -b to b, where b is half of the height of the convolution kernel K. This range also ensures that the convolution kernel completely covers the input holographic three-dimensional image.

[0043] For each pair of m,n, calculate the product of the value of the holographic three-dimensional image I at position (x + m,y + n) and the value of the convolution kernel K at position (a - m,b - n). The results of these products are accumulated to form the value of the holographic three-dimensional image feature map O(x,y) at position (x,y).

[0044] We can use GPU for GPU-accelerated computing. GPU-accelerated computing means using both the graphics processing unit (GPU) and the central processing unit (CPU) for scientific and fast computing.

[0045] After the convolution layer calculation, pooling layer calculation is also performed. The pooling layer is used to reduce the spatial size of the feature map. The pooling layer can extract the main information of the feature map, reduce the number of parameters and the amount of calculation, while retaining the most important feature information and removing redundancy and noise. The specific implementation process is as follows: The first step: Convert the holographic three-dimensional image feature map into matrix X: Convert the local area into a small matrix; The second step: Max pooling, extract the maximum value of each row of the large matrix, that is, obtain the maximum value of each local window; The third step: Output, map the generated holographic three-dimensional image of the hydraulic press 4 into the feature extraction module 26.

[0046] In the second step, all neurons of the pooling layer are connected to the neurons of the fully connected layer. After the pooling layer calculation, there is one or more fully connected layers. In the fully connected layer, all nodes of the previous layer are connected to each node of the next layer. This connection method enables the fully connected layer to perform more in-depth learning and extraction on the input features of the holographic three-dimensional image of the hydraulic press 4. The fully connected layer is usually used in the last few layers of the neural network for integrating and classifying the feature map of the holographic three-dimensional image of the hydraulic press 4. The fully connected layer maps the extracted features of the holographic three-dimensional image of the hydraulic press 4 into the feature extraction module 26.

[0047] In the third step, the feature extraction module 26 transmits the extracted features of the holographic three-dimensional image of the hydraulic press 4 to the image enhancement module 27. The image enhancement module 27 uses deep learning methods. The image enhancement module 27 includes a deconvolution algorithm and an anti-normalization algorithm. The deconvolution algorithm is used to restore the details and clarity of the holographic three-dimensional image. The deconvolution algorithm includes the following steps: The first step, initialization, select an initial estimated image f0 and a point spread function (PSF); usually the initial estimated image is the holographic three-dimensional image itself; The second step, iterative process, in each iteration k, update the estimated image f k+1 The formula is: , where q is the holographic three-dimensional image of the hydraulic press 4, h T is the transpose (or conjugate) of the point spread function (PSF), represents the convolution operation, which is the ratio of the blurred holographic three-dimensional image to the current estimated image for correcting the estimation; The third step, termination condition, usually set a fixed number of iterations, or stop the iteration when the improvement of the holographic three-dimensional image is no longer significant.

[0048] In another embodiment, the image enhancement module 27 can also adopt the GS algorithm. The GS algorithm specifically includes the following steps: First step, initialization: Prepare the input image, i.e., the amplitude information of the holographic three-dimensional image I of the hydraulic press 4. Initialize the phase information, which can usually be set to random phase or all-zero phase; Second step, frequency-domain iteration: Perform a Fourier transform on the current holographic three-dimensional image I (including the initial phase information) to obtain a frequency-domain image. Replace the amplitude of the frequency-domain image with the experimentally measured frequency-domain amplitude, while retaining the original phase information. Perform an inverse Fourier transform on the replaced frequency-domain image to return to the spatial domain; Third step, spatial-domain iteration: Replace the amplitude of the spatial-domain image with the experimentally measured spatial-domain amplitude, while retaining the phase information obtained after the inverse Fourier transform. Perform a Fourier transform on the updated image again to enter the next iteration; Fourth step, convergence judgment: Judge whether the phase converges, that is, whether the phase change is within a certain threshold range. If the convergence condition is reached, output the final phase information; otherwise, return to the second step to continue the iteration Finally, the image enhancement module 27 performs multiple calculations on the holographic three-dimensional image I to obtain a high-quality three-dimensional holographic image of the hydraulic press. The image enhancement module 27 is a key step in improving the visual effect of the three-dimensional holographic image I of the hydraulic press, aiming to make it more suitable for human eye observation or computer processing by adjusting features such as brightness, contrast, sharpness, and color of the three-dimensional holographic image I. Finally, the complete three-dimensional holographic image of the hydraulic press is stored in the three-dimensional reconstruction module 23 in the host computer 2.

[0049] The three-dimensional reconstruction module 23 maps the three-dimensional holographic image of the hydraulic press in the laser generator 3, and the laser beam in the laser generator 3 projects the complete three-dimensional holographic image of the hydraulic press.

[0050] The laser beam in the laser generator 3 uses the same laser as the reproduction light for taking the holographic image of the hydraulic press 4, irradiates the image enhancement module 27 in the host computer 2, and through the diffraction principle, the reproduction light forms the same light wave as the original object on the three-dimensional holographic image of the hydraulic press, thus reconstructing the three-dimensional holographic image of the hydraulic press in space.

[0051] In another embodiment, femtosecond laser can also be used to break down air to form a three-dimensional holographic image of the hydraulic press. The femtosecond laser technology breaks down air through high-energy laser to form a luminous plasma, thus directly presenting the three-dimensional holographic image of the hydraulic press in the air. The specific steps are as follows: First step: Laser focusing: Use a high-energy femtosecond laser to focus the laser into the air, reaching an extremely high intensity of 100 terawatts per square centimeter, and break down the air to form a luminous plasma; Second step: Three-dimensional scanning: Use a three-dimensional scanner to scan the laser beam, arrange and combine the luminous dots, and form a three-dimensional holographic image of the hydraulic press in the air; Step 3: Image display. The staff can view the three-dimensional holographic image of the hydraulic press from all angles without dead angles, without the need for media such as a screen or water mist.

[0052] In the output space of the three-dimensional holographic image of the hydraulic press, we can visually observe the entire external shape of the hydraulic press 4 and the internal structure of the hydraulic press 4. This enables us to not only intuitively see aspects such as the position of the slider, the flow direction of the hydraulic oil in the hydraulic pipes, the working conditions of the hydraulic components, the movement of the piston inside the hydraulic cylinder, the position distribution of the rod cavity or the plug cavity, or the deformation of the frame of the hydraulic press 4, but also provide real-time dynamic operation diagrams of the hydraulic press 4 for training new employees. It more intuitively shows the internal changes of the hydraulic press 4 at different operation stages, strengthening the new employees' understanding and operation ability of the hydraulic press 4, so that they can master the operation skills faster. Compared with technicians viewing the model of the hydraulic press 4 in 3D software on a computer, by visually observing the three-dimensional holographic image of the hydraulic press, it is easier to find the installation positions of relevant components and is closer to the actual hydraulic press 4. Moreover, in case of a malfunction, the three-dimensional holographic image of the hydraulic press can also be referred to.

[0053] Workflow: As Figure 4 shown, first install the image acquisition device 1 around the hydraulic press 4. After the image acquisition device 1 acquires the holographic image of the hydraulic press 4 with interference fringes, transmit the holographic image of the hydraulic press 4 with interference fringes to the host computer 2. After being processed by the image imaging module 21, it becomes a stable digital holographic image. Then, the phase extraction module 22 extracts the phase information and amplitude information from the digital holographic image and transmits them to the three-dimensional reconstruction module 23. The three-dimensional reconstruction module 23 generates a high-precision three-dimensional holographic image of the hydraulic press through the dataset training module 25, the feature extraction module 26, and the image enhancement module 27, and then outputs it at the corresponding position in space through the laser generator 3, facilitating the staff to more intuitively see the three-dimensional holographic image and the internal structure of the hydraulic press.

[0054] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes, should be covered within the protection scope of the present invention.

Claims

1. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology, including a monitoring system and a hydraulic press (4), characterized in that: The monitoring system includes an image acquisition device (1), a host computer (2) communicatively connected to the image acquisition device (1), and a laser generator (3) communicatively connected to the host computer (2). The image acquisition device (1) is used to acquire the holographic image of the hydraulic press (4) with interference fringes. The host computer (2) is used to process the holographic image of the hydraulic press (4) acquired by the image acquisition device (1) and generate a three-dimensional holographic image of the hydraulic press. The laser generator (3) projects the three-dimensional holographic image of the hydraulic press to the required position.

2. The visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 1, wherein: The host computer (2) includes an image imaging module (21), a phase extraction module (22), and a three-dimensional reconstruction module (23). The image imaging module (21) processes the holographic image with interference fringes acquired by the image acquisition device (1). The phase extraction module (22) extracts the information from the digital holographic image processed by the image imaging module (21). The three-dimensional reconstruction module (23) reconstructs the holographic two-dimensional image in the phase extraction module (22) and generates a three-dimensional holographic image of the hydraulic press.

3. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 1, characterized in that: The image acquisition device (1) is a laser camera, which generates a laser light source. The laser light source is divided into an object beam and a reference beam. After the object beam and the reference beam are superimposed, a holographic image with alternating bright and dark interference fringes is generated.

4. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 2, characterized in that: The host computer (2) further includes a digital sensor, which records the intensity distribution of the interference fringes in the holographic image with interference fringes. The image imaging module (21) processes the interference fringe information in the digital sensor to form a stable digital holographic image.

5. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 4, characterized in that: The digital holographic image includes phase information and amplitude information. After the phase extraction module (22) extracts the phase information and amplitude information, a holographic two-dimensional image is obtained, and then the holographic two-dimensional image with phase information and amplitude information is transmitted into the three-dimensional reconstruction module (23).

6. The visual system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 5, wherein: The extraction formula for the phase extraction module (22) to extract amplitude information is as follows: , where A(x,y) represents the amplitude information, which is the intensity of the optical field at the position (x,y); U(x,y) represents the complex amplitude distribution, which is the complex value of the optical field at the spatial position (x,y); Re[U(x,y)] represents the real part of the complex amplitude, and Im[U(x,y)] represents the imaginary part of the complex amplitude. The extraction formula for extracting phase information is as follows: , where represents the phase information, which is the phase angle of the optical field at the position (x, y); represents the imaginary part of the complex amplitude, which is the sine component of the optical field; represents the real part of the complex amplitude, which is the cosine component of the optical field.

7. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 5, characterized in that, The three-dimensional reconstruction module (23) constructs a three-dimensional holographic image of the hydraulic press with higher accuracy using the holographic two-dimensional image. The three-dimensional reconstruction module (23) includes a model-driven deep learning module (24), and the model-driven deep learning module (24) includes a dataset training module (25), a feature extraction module (26), and an image enhancement module (27).

8. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 7, characterized in that: The dataset training module (25) pre-inserts the pre-trained hydraulic press-related data with phase information and amplitude information to obtain a holographic three-dimensional image. The feature extraction module (26) automatically extracts features from the holographic three-dimensional image to obtain a holographic three-dimensional image one. The image enhancement module (27) is used to improve the resolution and clarity of the holographic three-dimensional image one to obtain a three-dimensional holographic image of the hydraulic press.

9. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 8, characterized in that: The feature extraction module (26) uses a convolutional neural network to extract. The convolutional neural network can automatically learn and extract features from the original data. In the processing of the holographic three-dimensional image, the convolutional neural network extracts local features through the combination of convolutional layers, pooling layers, and fully connected layers, and gradually abstracts high-level features through the hierarchical structure.

10. A visualization system for monitoring the production of a hydraulic press based on holographic imaging technology according to claim 8, characterized in that: The three-dimensional holographic image of the hydraulic press projects a three-dimensional holographic image of the hydraulic press through the laser generator.

Images