Particle tracking treatment method suitable for transparent sample, preparation process and seepage experiment device thereof

Abstract

The invention discloses a particle tracking treatment method suitable for a transparent sample, a preparation process and a seepage experiment device thereof. The method comprises the following steps: generating and acquiring first data corresponding to particles in a transparent sample, preprocessing the first data, and constructing a data set corresponding to the first data; creating a first model corresponding to the first data, and extracting, processing and generating second data corresponding to particles based on the first model; based on the second data, performing extraction processing to generate third data corresponding to the data set, and based on the third data, generating fourth data corresponding to the particles; wherein the fourth data are movement speed data and trajectory data of the particles, and the preparation process, the seepage experiment device, the system and the platform corresponding to the method can improve the detection and tracking precision of particle movement, and can effectively eliminate background interference and accurately identify the particles in a complex flow field.

Description

Technical Field

[0001] The present invention belongs to the technical field of seepage experiments, and in particular relates to a particle tracking processing method, a preparation process and a seepage experiment device suitable for transparent samples. Background Art

[0002] At present, seepage experiments are an important means to study the movement laws of fluids in porous media and are widely used in civil engineering, environmental science, petroleum engineering and other fields.

[0003] Traditional seepage experimental methods often rely on manual observation and recording of particle motion trajectories, which is inefficient and easily affected by human factors, making it difficult to obtain high-precision data. In recent years, with the development of computer vision and deep learning technologies, image processing-based methods have gradually become a new trend in seepage experimental research. Using high-speed cameras to capture particle motion in real time, combined with image processing algorithms, can effectively improve the accuracy and real-time performance of particle detection. In addition, the preparation methods of transparent samples are also constantly evolving, allowing researchers to better observe the interaction between fluids and particles.

[0004] However, existing technologies still have limitations in transparent sample preparation, flow experimental setup design, and the effectiveness of particle tracking velocimetry. For example, traditional sample preparation methods can result in irregular particle morphology, affecting experimental results; and existing particle tracking algorithms still face challenges in handling complex backgrounds, varying lighting conditions, and overlapping particles.

[0005] Therefore, in view of the above technical problems and defects, it is urgent to design and develop a particle tracking processing method, preparation process and seepage experimental device suitable for transparent samples. Summary of the Invention

[0006] In order to overcome the shortcomings and difficulties of the above-mentioned existing technologies, the present invention provides a particle tracking processing method, preparation process and seepage experimental device suitable for transparent samples, aiming to improve the efficiency and accuracy of seepage experiments and provide more reliable data support for related research.

[0007] The first purpose of the present invention is to provide a particle tracking and processing method suitable for transparent samples; the second purpose of the present invention is to provide a preparation process suitable for transparent samples; the third purpose of the present invention is to provide a seepage experimental device suitable for transparent samples; the fourth purpose of the present invention is to provide a particle tracking and processing system suitable for transparent samples; the fifth purpose of the present invention is to provide a particle tracking and processing platform suitable for transparent samples.

[0008] The first object of the present invention is achieved in that the method comprises the steps of:

[0009] Generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to the particle distribution;

[0010] Creating a first model corresponding to the first data, and extracting and processing second data corresponding to the particles based on the first model; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels;

[0011] Based on the second data, extraction processing generates third data corresponding to the data set, and based on the third data, generates fourth data corresponding to the particle; wherein, the third data is the position information data of the particle; and the fourth data is the movement speed data and trajectory data of the particle.

[0012] Furthermore, the generating and acquiring first data corresponding to particles in the transparent sample, preprocessing the first data, and constructing a data set corresponding to the first data further includes:

[0013] Generate and obtain first data corresponding to particles in a transparent sample and having different backgrounds and different lighting conditions;

[0014] De-noising the first data and marking particle position information corresponding to the first data; wherein the de-noising is Gaussian filtering de-noising;

[0015] Based on the data set and in combination with the particle position information, fifth data corresponding to the first data is generated; the fifth data is image size data.

[0016] Furthermore, the step of creating a first model corresponding to the first data and extracting and processing the generated second data corresponding to the particles based on the first model further includes:

[0017] Establishing a multi-layer convolution layer and a pooling layer corresponding to the first model, and extracting and generating second data corresponding to the particles based on the multi-layer convolution layer and the pooling layer;

[0018] Based on the skip connection and in combination with the Sigmoid activation function, the ownership of the pixel corresponding to the first data is determined.

[0019] Furthermore, the step of creating a first model corresponding to the first data and extracting and processing the generated second data corresponding to the particles based on the first model further includes:

[0020] Creating a second model corresponding to the first data; wherein the second model is a prediction and evaluation model;

[0021] According to the binary cross entropy loss function, sixth data corresponding to the first data is calculated and generated; wherein the sixth data is the error data between the prediction result of the second model and the actual annotation; the calculation formula is as follows:

[0022]

[0023] Where L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample.

[0024] Furthermore, the extracting and processing based on the second data to generate third data corresponding to the data set, and generating fourth data corresponding to the particles based on the third data, further includes:

[0025] Based on the data set, dividing and processing the data set into a training set and a validation set corresponding to the first data, and generating seventh data corresponding to the validation set; wherein the seventh data is performance data of the evaluation model on the validation set;

[0026] According to the seventh data, the parameter data corresponding to the second model is corrected in real time;

[0027] Create a third model corresponding to the first data, and analyze and process the first data in real time based on the third model; wherein the third model is a training model.

[0028] The second object of the present invention is achieved by: a transparent sample prepared by the preparation process is used to implement the particle tracking processing method, and the preparation process includes the following steps:

[0029] Generate and acquire eighth data corresponding to the transparent sample, and create a pore geometry database corresponding to the eighth data; wherein the eighth data is pore morphology data of the porous medium;

[0030] Calculating and generating ninth data corresponding to the eighth data, and generating corresponding tenth data based on the ninth data; wherein the ninth data is a probability density distribution of a pore target shape index; and the tenth data is shape index data;

[0031] Based on the pore geometry database, respectively retrieving pore morphologies corresponding to the matching process and the scaling and orientation process;

[0032] Combined with the overlap detection method, the pore geometry is assigned to the sample domain, and the reconstructed pore morphology and the target pore morphology are analyzed for similarity.

[0033] The pore morphology is redistributed until a specified tolerance is met, and a three-dimensional model corresponding to the transparent specimen is generated based on the reconstructed porous medium by combining 3D printing technology and transparent materials.

[0034] The third object of the present invention is achieved as follows: the device is used to implement the particle tracking processing method, the device includes a water circulation system for circulating water supply; a seepage generating device for achieving seepage and a detection device for particle tracking processing;

[0035] The water circulation system is connected to water pipes at both ends of the seepage generating device, one end of which is connected to the water supply tank and the other end is connected to the sand-water collection and separation system; the water inlet pipe end of the water circulation system is connected to a water pump, and the flow meter and valve installed on the water pipe are used to adjust the seepage flow rate and measure the water flow rate in real time; the end of the water outlet pipe is provided with a screen for solid-liquid separation;

[0036] The seepage generating device includes a particle supply system and an erosion box; the particle supply system is used to continuously supply particles to the erosion box; the interior of the erosion box is divided into three parts by a water seepage partition, the middle part of which is used to place sample particles and the two ends are used to place glass beads; the two ends of the erosion box are respectively provided with an exhaust hole;

[0037] The detection equipment consists of a differential pressure gauge and an optical imaging system; the differential pressure gauge is respectively arranged at the left and right ends of the middle section of the seepage generating device; the optical imaging system consists of a color-variable light source, a high-speed camera and an image processing device.

[0038] The fourth object of the present invention is achieved as follows: the system is used to implement the particle tracking processing method applicable to transparent samples, and the system includes:

[0039] a data generation and construction unit, configured to generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to the particle distribution;

[0040] a data creation and generation unit, configured to create a first model corresponding to the first data, and extract and process the first model to generate second data corresponding to the particles; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels;

[0041] A data processing and generation unit is used to extract, process and generate third data corresponding to the data set based on the second data, and generate fourth data corresponding to the particles based on the third data; wherein the third data is the position information data of the particles; and the fourth data is the movement speed data and trajectory data of the particles.

[0042] Furthermore, the data generation construction unit further includes:

[0043] A first generating module is used to generate and obtain first data corresponding to particles in the transparent sample and having different backgrounds and different lighting conditions;

[0044] A first processing module is configured to perform denoising on the first data and label the particle position information corresponding to the first data; wherein the denoising is performed by Gaussian filtering;

[0045] a second generating module, configured to generate fifth data corresponding to the first data based on the data set and in combination with the particle position information; the fifth data being image size data;

[0046] And / or, the data creation and generation unit further includes:

[0047] A first establishing module, configured to establish a multi-layer convolution layer and a pooling layer corresponding to the first model, and extract and generate second data corresponding to the particles based on the multi-layer convolution layer and the pooling layer;

[0048] A first determination module is configured to determine the ownership of the pixel corresponding to the first data based on a skip connection and in combination with a Sigmoid activation function;

[0049] A second building module is configured to create a second model corresponding to the first data; wherein the second model is a prediction and evaluation model;

[0050] The first calculation module is configured to calculate and generate sixth data corresponding to the first data based on a binary cross entropy loss function; wherein the sixth data is error data between the prediction result of the second model and the actual annotation; the calculation formula is as follows:

[0051]

[0052] Where L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample;

[0053] And / or, the data processing and generating unit further includes:

[0054] a second processing module, configured to divide the data set into a training set and a validation set corresponding to the first data, and generate seventh data corresponding to the validation set; wherein the seventh data is performance data of the evaluation model on the validation set;

[0055] a third processing module, configured to modify parameter data corresponding to the second model in real time according to the seventh data;

[0056] The fourth processing module is used to create a third model corresponding to the first data, and analyze and process the first data in real time based on the third model; wherein the third model is a training model.

[0057] The fifth object of the present invention is achieved as follows: it includes a processor, a memory and a particle tracking processing platform control program applicable to transparent samples; wherein the particle tracking processing platform control program applicable to transparent samples is executed by the processor, the particle tracking processing platform control program applicable to transparent samples is stored in the memory, and the particle tracking processing platform control program applicable to transparent samples implements the particle tracking processing method applicable to transparent samples.

[0058] The present invention generates and obtains first data corresponding to particles in a transparent sample through a method, preprocesses the first data and constructs a data set corresponding to the first data; wherein, the first data is image data corresponding to the particle distribution; creates a first model corresponding to the first data, and based on the first model, extracts and processes to generate second data corresponding to the particles; wherein, the first model is a convolutional neural network model; the second data is feature data of particles at different levels; based on the second data, extracts and processes to generate third data corresponding to the data set, and based on the third data, generates fourth data corresponding to the particles; wherein, the third data is the position information data of the particles; the fourth data is the movement speed data and trajectory data of the particles, as well as the preparation process, seepage experimental device, system and platform corresponding to the method, which can improve the detection and tracking accuracy of particle motion, and can effectively eliminate background interference in complex flow fields and accurately identify particles. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 A schematic diagram of the process steps of a particle tracking processing method applicable to transparent samples of the present invention;

[0061] Figure 2 1. A schematic flow chart of an embodiment of a particle tracking processing method applicable to transparent samples according to the present invention;

[0062] Figure 3 A schematic diagram illustrating the accuracy of an embodiment of a particle tracking processing method applicable to transparent samples of the present invention;

[0063] Figure 4 A schematic diagram of the process steps of a preparation process for a transparent sample according to the present invention;

[0064] Figure 5 This is a schematic diagram of the actual effect of a transparent sample preparation process applicable to the present invention;

[0065] Figure 6 This is a schematic diagram of the structure of a seepage experiment device suitable for transparent samples according to the present invention;

[0066] Figure 7 This is a schematic diagram of the architecture of a particle tracking processing system suitable for transparent samples according to the present invention;

[0067] Figure 8 Schematic diagram of the particle tracking processing platform architecture suitable for transparent samples of the present invention. DETAILED DESCRIPTION

[0068] In order to better understand the purpose, technical solutions and advantages of the present invention, the present invention is further described below with reference to the accompanying drawings and specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0069] The present invention may also be implemented or applied through other different specific examples, and the details in this specification may also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0070] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly.

[0071] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of such features. Secondly, the technical solutions between the various embodiments can be combined with each other, but this must be based on the fact that ordinary technicians in this field can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0072] Preferably, the particle tracking processing method for transparent samples of the present invention is applied to one or more terminals or servers. The terminal is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0073] The terminal can be a computing device such as a desktop computer, notebook, PDA, cloud server, etc. The terminal can interact with the client through a keyboard, mouse, remote control, touchpad, or voice control device.

[0074] The present invention provides a particle tracking processing method, system and platform suitable for transparent samples.

[0075] like Figure 1 FIG. 1 is a flow chart of a particle tracking processing method applicable to transparent samples provided by an embodiment of the present invention.

[0076] In this embodiment, the particle tracking processing method for transparent samples can be applied to a terminal with a display function or a fixed terminal, and the terminal is not limited to a personal computer, a smart phone, a tablet computer, a desktop computer or an all-in-one computer equipped with a camera, etc.

[0077] The particle tracking processing method for transparent specimens can also be applied in a hardware environment consisting of a terminal and a server connected to the terminal via a network. The network includes, but is not limited to, a wide area network, a metropolitan area network, or a local area network. The particle tracking processing method for transparent specimens according to the embodiments of the present invention can be executed by a server, a terminal, or both.

[0078] For example, for a terminal that needs to perform particle tracking processing for transparent samples, the particle tracking processing function for transparent samples provided by the method of the present invention can be directly integrated into the terminal, or a client for implementing the method of the present invention can be installed. For another example, the method provided by the present invention can also be run on a server or other device in the form of a software development kit (SDK). The SDK provides an interface for the particle tracking processing function for transparent samples, and the terminal or other device can implement the particle tracking processing function for transparent samples through the provided interface. The present invention is further explained below with reference to the accompanying drawings.

[0079] like Figure 1 As shown, the present invention provides a particle tracking processing method applicable to transparent samples, the method comprising the following steps:

[0080] S01. Generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to particle distribution;

[0081] S02. Creating a first model corresponding to the first data, and extracting and processing second data corresponding to the particles based on the first model; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels;

[0082] S03. Based on the second data, extract and process to generate third data corresponding to the dataset, and based on the third data, generate fourth data corresponding to the particles; wherein the third data is the particle position information data, and the fourth data is the particle velocity data and trajectory data. Specifically, using the trained model, the image data collected during the experiment is analyzed in real time to extract particle position information. The particle velocity and trajectory are calculated by analyzing the particle displacement between consecutive frames.

[0083] The generating and acquiring first data corresponding to particles in the transparent sample, preprocessing the first data and constructing a data set corresponding to the first data further includes:

[0084] S011. Generate and obtain first data corresponding to particles in the transparent sample and under different backgrounds and lighting conditions; by collecting a large number of data sets continuously captured during the experiment, covering different backgrounds, lighting conditions, and particle distributions.

[0085] S012. De-noise the first data and mark the particle position information corresponding to the first data; wherein the de-noising process is Gaussian filtering de-noising process; that is, de-noising and other processes are performed on the collected image, and the positions of the particles in the image are marked.

[0086] S013. Based on the data set and in combination with the particle position information, generate fifth data corresponding to the first data; the fifth data is image size data.

[0087] The step of creating a first model corresponding to the first data and extracting and processing the generated second data corresponding to the particles based on the first model further includes:

[0088] S021. Establishing a multi-layer convolution layer and a pooling layer corresponding to the first model, and extracting and generating second data corresponding to the particles based on the multi-layer convolution layer and the pooling layer;

[0089] S022. Determine the identity of the pixels corresponding to the first data based on skip connections and a Sigmoid activation function. That is, after downsampling, skip connections are added to preserve details when restoring the image; the output layer uses a Sigmoid activation function to determine whether each pixel belongs to a particle or background.

[0090] The step of creating a first model corresponding to the first data and extracting and processing the generated second data corresponding to the particles based on the first model further includes:

[0091] S023. Create a second model corresponding to the first data; wherein the second model is a prediction and evaluation model;

[0092] S024. Calculate and generate sixth data corresponding to the first data according to the binary cross entropy loss function; wherein the sixth data is the error data between the prediction result of the second model and the actual annotation; the calculation formula is as follows:

[0093]

[0094] Where L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample. That is, the binary cross entropy loss function is selected to calculate the error between the model prediction result and the actual annotation.

[0095] The extracting process generates third data corresponding to the data set based on the second data, and generates fourth data corresponding to the particle based on the third data, further comprising:

[0096] S031. Based on the data set, divide and process it into a training set and a validation set corresponding to the first data, and generate seventh data corresponding to the validation set; wherein the seventh data is performance data of the evaluation model on the validation set;

[0097] S032. Correcting parameter data corresponding to the second model in real time based on the seventh data;

[0098] S033. Create a third model corresponding to the first data, and analyze and process the first data in real time based on the third model; wherein the third model is a training model.

[0099] Specifically, in the embodiment of the present invention, Figure 2-Figure 3 As shown in the figure, an improved particle tracking velocimetry method is proposed, which uses a high-frame-rate high-speed camera to capture the rapid movement of particles. At the same time, a deep learning-based image processing algorithm is developed, and a convolutional neural network (CNN) is used to automatically detect and track particles.

[0100] The improved particle tracking velocimetry method comprises the following steps:

[0101] S31. By collecting a large number of data sets continuously captured during the experiment, covering different backgrounds, lighting conditions, and particle distributions, the collected images are processed by denoising and the positions of the particles in the images are marked.

[0102] S32. Build a convolutional neural network model with strong feature extraction capabilities and target segmentation accuracy: set up multiple convolutional layers and pooling layers to extract different levels of features of particles; after downsampling, add skip connections to retain details when restoring the image; the output layer uses the Sigmoid activation function to determine whether each pixel belongs to a particle or background.

[0103] S33. Select the binary cross entropy loss function and calculate the error between the model prediction result and the actual annotation:

[0104]

[0105] Among them, L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample.

[0106] S34. Divide the dataset into a training set and a validation set. Evaluate the model's performance on the validation set and adjust model parameters to prevent overfitting. Use the trained model to analyze the image data collected during the experiment in real time to extract particle position information. Calculate the particle's velocity and trajectory by analyzing particle displacement between consecutive frames.

[0107] Example: The constructed convolutional neural network model takes a single-channel grayscale image of 256×256 pixels as input. Its encoder consists of four modules: the first module uses two 3×3 convolutional layers (both with "same" padding), the first layer outputs 64 feature channels and uses ReLU activation. The second module also outputs 64 feature channels, followed by a 2×2 max pooling layer to achieve spatial size reduction; the second module uses two 3×3 convolutional layers, increasing the number of output channels to 128, followed by a 2×2 max pooling layer; the third module uses two 3×3 convolutional layers to output 256 channels, which are downsampled by a 2×2 max pooling layer; the fourth module uses two 3×3 convolutional layers to output 512 channels, which are also downsampled by a 2×2 max pooling layer. The number of channels is further increased to 1024 in the bottleneck layer using two 3×3 convolutional layers, and a dropout of 0.5 is introduced to prevent overfitting. In the decoder, each upsampling module first uses a 2×2 transposed convolution for upsampling, then makes a jump connection with the feature map of the corresponding layer in the encoder, and then restores the image details through two 3×3 convolution layers, whose output channels are 512, 256, 128, and 64, respectively. Finally, a 1×1 convolution layer is used to reduce the output channel to 1, and a Sigmoid activation function is used to map each pixel to the interval [0,1] to generate a binary classification probability map. In terms of model training, the Adam optimizer is used, and the initial learning rate is set to 1×10 -4 , the loss function is binary cross entropy, and the batch size is set to 16.

[0108] To achieve the above object, the present invention also provides a preparation process suitable for transparent samples, such as Figure 4-Figure 5 As shown, the transparent sample prepared by the preparation process is used to implement the particle tracking processing method, and the preparation process includes the following steps:

[0109] S001. Generate and obtain eighth data corresponding to the transparent sample, and create a pore geometry database corresponding to the eighth data; wherein the eighth data is pore morphology data of the porous medium;

[0110] S002. Calculating and generating ninth data corresponding to the eighth data, and generating corresponding tenth data based on the ninth data; wherein the ninth data is a probability density distribution of a pore target shape index; and the tenth data is shape index data;

[0111] S003. Based on the pore geometry database, respectively retrieve pore morphologies corresponding to the matching process and the scaling and orientation process;

[0112] S004. In combination with the overlap detection method, the pore geometry is assigned to the sample domain, and similarity analysis is performed on the reconstructed pore morphology and the target pore morphology;

[0113] S005. Redistribute the pore morphology until a specified tolerance is met, and generate a three-dimensional model corresponding to the transparent sample based on the reconstructed porous medium by combining 3D printing technology and transparent materials.

[0114] Specifically, in an embodiment of the present invention, a method for preparing a transparent sample is proposed, which uses CT images to extract the pore morphology of a cubic porous sample from multiple directions, and forms a transparent sample with certain porosity based on 3D printing.

[0115] The transparent sample preparation method specifically comprises the following steps:

[0116] S11. Based on the original image of the porous medium, various pore morphologies of the specimen are extracted from multiple directions to establish a pore geometry database. In a laboratory environment, using continuous and stable lighting, a combination of filtering techniques and binarization methods are used to identify and extract pore morphologies. The shape characteristics of each pore outline obtained from the original image are quantified.

[0117] S12. To understand the shape variations of pore morphologies, we first calculated the probability density distribution of the target pore shape indices. For each void, we randomly determined shape indices, including aspect ratio, circularity, orientation, and size, based on the obtained probability density distribution. We then retrieved matching pore morphologies from an established database using the shape indices (aspect ratio and circularity). The selected pore morphologies were then scaled and oriented based on the other shape indices (orientation and size).

[0118] S13. Based on the overlap detection method, the pore geometry is assigned to the sample domain. Then, a similarity analysis is performed between the reconstructed pore morphology and the target pore morphology, and the pore morphology is reassigned until the specified tolerance is met.

[0119] S14. Based on the reconstructed porous medium, a three-dimensional model is generated using 3D printing technology and transparent materials.

[0120] To achieve the above object, the present invention also provides a seepage experiment device suitable for transparent samples, such as Figure 6 As shown, the device is used to implement the particle tracking processing method, and the device includes a water circulation system for circulating water supply; a seepage generating device for achieving seepage and a detection device for particle tracking processing;

[0121] The water circulation system is connected to water pipes at both ends of the seepage generating device, one end of which is connected to the water supply tank and the other end is connected to the sand-water collection and separation system; the water inlet pipe end of the water circulation system is connected to a water pump, and the flow meter and valve installed on the water pipe are used to adjust the seepage flow rate and measure the water flow rate in real time; the end of the water outlet pipe is provided with a screen for solid-liquid separation;

[0122] The seepage generating device includes a particle supply system and an erosion box; the particle supply system is used to continuously supply particles to the erosion box; the interior of the erosion box is divided into three parts by a water seepage partition, the middle part of which is used to place sample particles and the two ends are used to place glass beads; the two ends of the erosion box are respectively provided with an exhaust hole;

[0123] The detection equipment consists of a differential pressure gauge and an optical imaging system; the differential pressure gauge is respectively arranged at the left and right ends of the middle section of the seepage generating device; the optical imaging system consists of a color-variable light source, a high-speed camera and an image processing device.

[0124] Specifically, in an embodiment of the present invention, an experimental device for relatively high-speed seepage is proposed, which specifically includes three parts: a water circulation system, a seepage generating device and a detection device: the water circulation system can realize water circulation and flow rate adjustment according to the flow meter and valve; the seepage generating device places the prepared transparent sample for seepage experiment, and is equipped with a fine particle supply system to continuously transport particles; the detection equipment can obtain the head difference, particle movement and its parameters through a differential pressure meter, a high-speed camera and an image processing device.

[0125] The experimental device for relatively high-speed seepage includes three parts: a water circulation system, a seepage generating device, and a detection device.

[0126] The water circulation system connects water pipes at both ends of the seepage generator: one end connects to the water supply tank, and the other connects to the sand and water collection and separation system. A water pump is connected to the inlet pipe for continuous water supply, and a flow meter and valve installed on the pipe allow for real-time adjustment of the seepage flow rate and measurement of water flow rate. A screen is placed at the end of the outlet pipe to separate solids and liquids. Solids are retained on the screen, while liquids are collected back into the water supply tank, achieving a circulating water supply.

[0127] The seepage device consists of a particle supply system and an erosion box. The fine particle supply system continuously delivers particles to the erosion box. The interior of the erosion box is divided into three sections by a water-permeable partition. The middle section contains the sample particles, while glass beads are placed at both ends to eliminate uneven cross-sectional flow velocity caused by the sudden increase in the seepage cross-section. Furthermore, an exhaust hole is installed at each end of the erosion box. It is opened before the experiment to remove air from the device and remains closed during the experiment.

[0128] The detection equipment is mainly composed of a differential pressure gauge and an optical imaging system: the differential pressure gauges are respectively set at the left and right ends of the middle section of the seepage generating device to measure the head difference between the two ends; the optical imaging system is mainly composed of a variable color light source, a high-speed camera and an image processing device. Through experimental photography and identification of particles, the particle tracking velocimetry algorithm is used to calculate the coordinate position of the particles in each frame, thereby tracking their movement trajectory and calculating the movement speed, so as to determine the erosion, migration and deposition of fine particles in the sample.

[0129] To achieve the above object, the present invention also provides a particle tracking processing system suitable for transparent samples, such as Figure 7 As shown, the system is used to implement the particle tracking processing method applicable to transparent samples, and the system specifically includes:

[0130] a data generation and construction unit, configured to generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to the particle distribution;

[0131] a data creation and generation unit, configured to create a first model corresponding to the first data, and extract and process the first model to generate second data corresponding to the particles; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels;

[0132] A data processing and generation unit is used to extract, process and generate third data corresponding to the data set based on the second data, and generate fourth data corresponding to the particles based on the third data; wherein the third data is the position information data of the particles; and the fourth data is the movement speed data and trajectory data of the particles.

[0133] The data generation construction unit further includes:

[0134] A first generating module is used to generate and obtain first data corresponding to particles in the transparent sample and having different backgrounds and different lighting conditions;

[0135] A first processing module is configured to perform denoising on the first data and label the particle position information corresponding to the first data; wherein the denoising is performed by Gaussian filtering;

[0136] a second generating module, configured to generate fifth data corresponding to the first data based on the data set and in combination with the particle position information; the fifth data being image size data;

[0137] And / or, the data creation and generation unit further includes:

[0138] A first establishing module, configured to establish a multi-layer convolution layer and a pooling layer corresponding to the first model, and extract and generate second data corresponding to the particles based on the multi-layer convolution layer and the pooling layer;

[0139] A first determination module is configured to determine the ownership of the pixel corresponding to the first data based on a skip connection and in combination with a Sigmoid activation function;

[0140] A second building module is configured to create a second model corresponding to the first data; wherein the second model is a prediction and evaluation model;

[0141] The first calculation module is configured to calculate and generate sixth data corresponding to the first data based on a binary cross entropy loss function; wherein the sixth data is error data between the prediction result of the second model and the actual annotation; the calculation formula is as follows:

[0142]

[0143] Where L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample;

[0144] And / or, the data processing and generating unit further includes:

[0145] a second processing module, configured to divide the data set into a training set and a validation set corresponding to the first data, and generate seventh data corresponding to the validation set; wherein the seventh data is performance data of the evaluation model on the validation set;

[0146] a third processing module, configured to modify parameter data corresponding to the second model in real time according to the seventh data;

[0147] The fourth processing module is used to create a third model corresponding to the first data, and analyze and process the first data in real time based on the third model; wherein the third model is a training model.

[0148] In the system solution embodiment of the present invention, the method steps involved in the particle tracking processing suitable for transparent samples have been described above in detail. That is to say, the functional modules in the system are used to implement the steps or sub-steps in the above method embodiment, which will not be repeated here.

[0149] To achieve the above objectives, the present invention also provides a particle tracking processing platform suitable for transparent samples, such as Figure 8As shown, it includes a processor, a memory, and a particle tracking processing platform control program applicable to transparent samples; wherein, the particle tracking processing platform control program applicable to transparent samples is executed by the processor, and the particle tracking processing platform control program applicable to transparent samples is stored in the memory, and the particle tracking processing platform control program applicable to transparent samples implements the particle tracking processing method steps applicable to transparent samples. For example:

[0150] S01. Generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to particle distribution;

[0151] S02. Creating a first model corresponding to the first data, and extracting and processing second data corresponding to the particles based on the first model; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels;

[0152] S03. Based on the second data, extract and process to generate third data corresponding to the data set, and based on the third data, generate fourth data corresponding to the particle; wherein the third data is the position information data of the particle; and the fourth data is the movement speed data and trajectory data of the particle.

[0153] The specific details of the steps have been explained above and will not be repeated here.

[0154] In an embodiment of the present invention, the built-in processor of the particle tracking processing platform for transparent specimens can be composed of an integrated circuit, such as a single packaged integrated circuit or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and a combination of various control chips. The processor utilizes various interfaces and circuits to connect various components, executes or runs programs or units stored in memory, and calls data stored in memory to perform various particle tracking processing functions for transparent specimens and process data.

[0155] The memory is used to store program codes and various data. It is installed in the particle tracking processing platform suitable for transparent samples and can automatically access programs or data at high speed during operation.

[0156] The memory includes read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electronically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0157] The present invention generates and obtains first data corresponding to particles in a transparent sample through a method, preprocesses the first data and constructs a data set corresponding to the first data; wherein, the first data is image data corresponding to the particle distribution; creates a first model corresponding to the first data, and based on the first model, extracts and processes to generate second data corresponding to the particles; wherein, the first model is a convolutional neural network model; the second data is feature data of particles at different levels; based on the second data, extracts and processes to generate third data corresponding to the data set, and based on the third data, generates fourth data corresponding to the particles; wherein, the third data is the position information data of the particles; the fourth data is the movement speed data and trajectory data of the particles, as well as the preparation process, seepage experimental device, system and platform corresponding to the method, which can improve the detection and tracking accuracy of particle motion, and can effectively eliminate background interference in complex flow fields and accurately identify particles. That is, the sample particle preparation method prepares transparent irregular particle samples through apparent shape characterization parameters, ensuring that the sample has an observable internal structure, which is conducive to the visual analysis of subsequent experiments; the experimental device can accurately obtain the motion information of the particles, improve the accuracy of the experimental data and the experimental efficiency; the improved particle tracking velocimetry method can realize real-time image analysis, improve the detection and tracking accuracy of particle motion, and can effectively eliminate background interference in complex flow fields and accurately identify particles.

[0158] In other words, the present invention utilizes CT images to extract pore morphology from multiple angles, combined with 3D printing technology to prepare transparent specimens. This allows the specimens to truly reflect the internal structure of the original porous medium, thereby improving experimental reproducibility and data reliability. The use of a high-frame-rate camera coupled with a deep-learning-based image processing algorithm to automatically detect and track particle motion significantly improves measurement accuracy and data acquisition efficiency, overcoming the limitations of existing technologies in analyzing fast-moving particles.

[0159] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A particle tracking processing method suitable for transparent samples, characterized in that: The method comprises the steps of: Generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to the particle distribution; Creating a first model corresponding to the first data, and extracting and processing second data corresponding to the particles based on the first model; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels; Based on the second data, extraction processing generates third data corresponding to the data set, and based on the third data, generates fourth data corresponding to the particle; wherein, the third data is the position information data of the particle; and the fourth data is the movement speed data and trajectory data of the particle.

2. The particle tracking processing method for transparent samples according to claim 1, characterized in that: The generating and acquiring first data corresponding to particles in the transparent sample, preprocessing the first data and constructing a data set corresponding to the first data further includes: Generate and obtain first data corresponding to particles in a transparent sample and having different backgrounds and different lighting conditions; De-noising the first data and marking particle position information corresponding to the first data; wherein the de-noising is Gaussian filtering de-noising; Based on the data set and in combination with the particle position information, fifth data corresponding to the first data is generated; the fifth data is image size data.

3. The particle tracking processing method for transparent samples according to claim 1, characterized in that: The step of creating a first model corresponding to the first data and extracting and processing the generated second data corresponding to the particles based on the first model further includes: Establishing a multi-layer convolution layer and a pooling layer corresponding to the first model, and extracting and generating second data corresponding to the particles based on the multi-layer convolution layer and the pooling layer; Based on the skip connection and in combination with the Sigmoid activation function, the ownership of the pixel corresponding to the first data is determined.

4. A particle tracking processing method applicable to transparent samples according to claim 1 or 3, characterized in that: The step of creating a first model corresponding to the first data and extracting and processing the generated second data corresponding to the particles based on the first model further includes: Creating a second model corresponding to the first data; wherein the second model is a prediction and evaluation model; According to the binary cross entropy loss function, sixth data corresponding to the first data is calculated and generated; wherein the sixth data is the error data between the prediction result of the second model and the actual annotation; the calculation formula is as follows: Where L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample.

5. The particle tracking processing method applicable to transparent samples according to claim 1, characterized in that: The extracting process generates third data corresponding to the data set based on the second data, and generates fourth data corresponding to the particle based on the third data, further comprising: Based on the data set, dividing and processing the data set into a training set and a validation set corresponding to the first data, and generating seventh data corresponding to the validation set; wherein the seventh data is performance data of the evaluation model on the validation set; According to the seventh data, the parameter data corresponding to the second model is corrected in real time; Create a third model corresponding to the first data, and analyze and process the first data in real time based on the third model; wherein the third model is a training model.

6. A preparation process suitable for transparent samples, characterized in that: The transparent sample prepared by the preparation process is used to implement the particle tracking processing method according to any one of claims 1 to 5, and the preparation process comprises the following steps: Generate and acquire eighth data corresponding to the transparent sample, and create a pore geometry database corresponding to the eighth data; wherein the eighth data is pore morphology data of the porous medium; Calculating and generating ninth data corresponding to the eighth data, and generating corresponding tenth data based on the ninth data; wherein the ninth data is a probability density distribution of a pore target shape index; and the tenth data is shape index data; Based on the pore geometry database, respectively retrieving pore morphologies corresponding to the matching process and the scaling and orientation process; Combined with the overlap detection method, the pore geometry is assigned to the sample domain, and the reconstructed pore morphology and the target pore morphology are analyzed for similarity. The pore morphology is redistributed until a specified tolerance is met, and a three-dimensional model corresponding to the transparent specimen is generated based on the reconstructed porous medium by combining 3D printing technology and transparent materials.

7. A seepage test device suitable for transparent samples, characterized in that: The device is used to implement the particle tracking processing method according to any one of claims 1 to 5, and the device includes a water circulation system for circulating water supply; a seepage generating device for achieving seepage and a detection device for particle tracking processing; The water circulation system is connected to water pipes at both ends of the seepage generating device, one end of which is connected to the water supply tank and the other end is connected to the sand-water collection and separation system; the water inlet pipe end of the water circulation system is connected to a water pump, and the flow meter and valve installed on the water pipe are used to adjust the seepage flow rate and measure the water flow rate in real time; the end of the water outlet pipe is provided with a screen for solid-liquid separation; The seepage generating device includes a particle supply system and an erosion box; the particle supply system is used to continuously supply particles to the erosion box; the interior of the erosion box is divided into three parts by a water seepage partition, the middle part of which is used to place sample particles and the two ends are used to place glass beads; the two ends of the erosion box are respectively provided with an exhaust hole; The detection equipment consists of a differential pressure gauge and an optical imaging system; the differential pressure gauge is respectively arranged at the left and right ends of the middle section of the seepage generating device; the optical imaging system consists of a color-variable light source, a high-speed camera and an image processing device.

8. A particle tracking processing system suitable for transparent samples, characterized in that: The system is used to implement the particle tracking processing method applicable to transparent samples as described in any one of claims 1 to 5, and the system includes: a data generation and construction unit, configured to generate and acquire first data corresponding to particles in a transparent sample, preprocess the first data, and construct a data set corresponding to the first data; wherein the first data is image data corresponding to the particle distribution; a data creation and generation unit, configured to create a first model corresponding to the first data, and extract and process the first model to generate second data corresponding to the particles; wherein the first model is a convolutional neural network model; and the second data is feature data of particles at different levels; A data processing and generation unit is used to extract, process and generate third data corresponding to the data set based on the second data, and generate fourth data corresponding to the particles based on the third data; wherein the third data is the position information data of the particles; and the fourth data is the movement speed data and trajectory data of the particles.

9. The particle tracking processing system for transparent samples according to claim 8, characterized in that: The data generation construction unit further includes: A first generating module is used to generate and obtain first data corresponding to particles in the transparent sample and having different backgrounds and different lighting conditions; A first processing module is configured to perform denoising on the first data and label the particle position information corresponding to the first data; wherein the denoising is performed by Gaussian filtering; a second generating module, configured to generate fifth data corresponding to the first data based on the data set and in combination with the particle position information; the fifth data being image size data; And / or, the data creation and generation unit further includes: A first establishing module, configured to establish a multi-layer convolution layer and a pooling layer corresponding to the first model, and extract and generate second data corresponding to the particles based on the multi-layer convolution layer and the pooling layer; A first determination module is configured to determine the ownership of the pixel corresponding to the first data based on a skip connection and in combination with a Sigmoid activation function; A second building module is configured to create a second model corresponding to the first data; wherein the second model is a prediction and evaluation model; The first calculation module is configured to calculate and generate sixth data corresponding to the first data based on a binary cross entropy loss function; wherein the sixth data is error data between the prediction result of the second model and the actual annotation; the calculation formula is as follows: Where L is the loss value, N is the number of samples, and y i is the true label of the i-th sample, is the predicted value of the i-th sample; And / or, the data processing and generating unit further includes: a second processing module, configured to divide the data set into a training set and a validation set corresponding to the first data, and generate seventh data corresponding to the validation set; wherein the seventh data is performance data of the evaluation model on the validation set; a third processing module, configured to modify parameter data corresponding to the second model in real time according to the seventh data; The fourth processing module is used to create a third model corresponding to the first data, and analyze and process the first data in real time based on the third model; wherein the third model is a training model.

10. A particle tracking processing platform suitable for transparent samples, characterized in that: It includes a processor, a memory and a particle tracking processing platform control program suitable for transparent samples; wherein, the particle tracking processing platform control program suitable for transparent samples is executed by the processor, the particle tracking processing platform control program suitable for transparent samples is stored in the memory, and the particle tracking processing platform control program suitable for transparent samples implements the particle tracking processing method suitable for transparent samples as described in any one of claims 1 to 5.

Images