Cement modified expansive soil ecological base material and preparation method thereof

By analyzing the mixing state through real-time video acquisition and deep learning technology, the mixing endpoint can be automatically determined, solving the problem of uneven mixing in traditional cement-modified expansive soil ecological substrates and achieving a more efficient preparation process.

CN118908632AActive Publication Date: 2024-11-08CHANGJIANG RIVER SCI RES INST CHANGJIANG WATER RESOURCES COMMISSION
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202410925777.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-08
Estimated Expiration
2044-07-11

AI Technical Summary

Technical Problem

The mixing time control of traditional cement-modified expansive soil ecological substrates relies on manual experience, which leads to uneven mixing and affects the stability and consistency of the preparation process.

Method used

The system uses a camera to collect real-time video of the stirring process and performs keyframe sampling and stirring texture feature analysis using video processing and analysis technology based on deep learning neural networks to automatically determine whether to stop stirring.

Benefits of technology

The automation level of the stirring process has been improved, the need for manual monitoring has been reduced, labor costs have been lowered, and a more intelligent preparation process has been achieved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118908632A_ABST
    Figure CN118908632A_ABST
Patent Text Reader

Abstract

The invention discloses a cement modified expansive soil ecological base material and a preparation method thereof.The cement modified expansive soil ecological base material comprises expansive soil, cement, ferrous sulfate, a water-retaining agent, organic fertilizer, peat and polyacrylamide, the mass of the cement accounts for 3%-10% of the mass of dry expansive soil, the mass of the ferrous sulfate accounts for 1%-2% of the mass of the dry expansive soil, and the mass of the water-retaining agent accounts for 1%-2% of the mass of the dry expansive soil. The mass of the water-retaining agent accounts for 0.05-0.2% of the mass of the dry expansive soil, the mass of the organic fertilizer accounts for 0.5-2% of the mass of the dry expansive soil, the mass of the peat accounts for 3-10% of the mass of the dry expansive soil, and the mass of the polyacrylamide accounts for 0.05-0.2% of the mass of the dry expansive soil. The material can inhibit expansion of expansive soil and enhance the overall stability of a slope filled with the expansive soil, and meanwhile, the material can be suitable for plant growth to achieve the purpose of ecological slope protection. The invention further provides a preparation method of the cement modified expansive soil ecological base material, and the preparation method is simple and easy to operate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent preparation, and more specifically, to a cement-modified expansive soil ecological substrate and a preparation method thereof. Background Art

[0002] Cement-modified expansive soil ecological substrate is a material used for soil improvement. It improves the physical and chemical properties of expansive soil by adding cement and other additives. This modified soil is usually used for vegetation restoration and ecological slope protection. It can improve soil stability, reduce soil expansion, and increase its bearing capacity.

[0003] In the preparation process of cement-modified expansive soil ecological substrate, the sieved expansive soil and cement need to be evenly mixed to ensure that the cement is evenly distributed throughout the expansive soil substrate to improve the overall performance and consistency of the substrate. However, the traditional control of mixing time usually relies on human experience or standard operating procedures. This method is not intelligent enough and can easily lead to uneven mixing, because the operator may not be able to accurately control the mixing time of each batch, thus affecting the stability of the mixing process.

[0004] Therefore, an optimized method for preparing a cement-modified expansive soil ecological substrate is desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a cement-modified expansive soil ecological substrate and a preparation method thereof, which collects the stirring state monitoring video in real time by a camera, and adopts the video processing and analysis technology based on deep learning neural network to perform key frame sampling and stirring texture feature analysis on the stirring state monitoring video, so as to automatically judge whether to stop stirring according to the texture time series aggregation representation characteristics of each stirring state key frame. In this way, the need for manual monitoring can be reduced, the labor cost can be reduced, the automation level of the stirring process can be improved, and a more intelligent cement-modified expansive soil ecological substrate preparation process can be achieved.

[0006] According to one aspect of the present application, a method for preparing a cement-modified expansive soil ecological substrate is provided, comprising:

[0007] The expansive soil raw material is dried, then rolled and sieved in sequence to obtain sieved expansive soil;

[0008] The sieved expansive soil is uniformly mixed with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture;

[0009] Determine the pH value of the cement expansive soil mixture, add ferrous sulfate accounting for 1% to 2% of the dry mass of the expansive soil to the cement expansive soil mixture, and adjust the pH value to 5.5-7.0 to obtain the adjusted cement expansive soil;

[0010] Adding 0.05% to 0.2% of a water retaining agent by dry weight of the expansive soil, 0.5% to 2% of an organic fertilizer by dry weight of the expansive soil, 3% to 10% of peat by dry weight of the expansive soil, and 0.05% to 0.2% of polyacrylamide by dry weight of the expansive soil into the adjusted cement expansive soil, and uniformly mixing to obtain a mixture;

[0011] Performing a light compaction test on the mixture to obtain the maximum dry density and optimal moisture content of the mixture;

[0012] Based on the optimal moisture content, water is added to the mixture to mix the mixture to prepare a soil sample with the optimal moisture content to obtain the cement-modified expansive soil ecological substrate;

[0013] The sieved expansive soil is uniformly mixed with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture, comprising:

[0014] Obtain stirring status monitoring video collected by the camera;

[0015] Performing key frame sampling on the stirring state monitoring video to obtain a time sequence of stirring state key frames;

[0016] Extracting texture features of each stirring state key frame in the time sequence of the stirring state key frames to obtain a time sequence of stirring texture feature maps;

[0017] Each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of stirring texture multi-scale semantic feature vectors;

[0018] Inputting the time series of the stirring texture multi-scale semantic feature vector into a node message propagation fusion network that fuses node importance to obtain a stirring texture temporal propagation aggregation representation vector as a stirring texture temporal propagation aggregation representation feature;

[0019] Based on the time series propagation aggregation representation feature of the stirring texture, a control instruction is obtained, and the control instruction is used to indicate whether to stop stirring.

[0020] According to another aspect of the present application, a cement-modified expansive soil ecological substrate is provided, and the cement-modified expansive soil ecological substrate is prepared by the aforementioned preparation method of the cement-modified expansive soil ecological substrate.

[0021] Compared with the prior art, the present application provides a cement-modified expansive soil ecological substrate and a preparation method thereof, which collects a stirring state monitoring video in real time by a camera, and adopts a video processing and analysis technology based on a deep learning neural network to perform key frame sampling and stirring texture feature analysis on the stirring state monitoring video, thereby automatically judging whether to stop stirring according to the texture time series aggregation representation features of each stirring state key frame. In this way, the need for manual monitoring can be reduced, the labor cost can be reduced, the automation level of the stirring process can be improved, and a more intelligent cement-modified expansive soil ecological substrate preparation process can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0023] Figure 1 The present invention is a flow chart of a method for preparing a cement-modified expansive soil ecological substrate according to an embodiment of the present application.

[0024] Figure 2 The present invention is a flow chart of uniformly mixing the sieved expansive soil with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture in the preparation method of the cement-modified expansive soil ecological substrate according to an embodiment of the present application.

[0025] Figure 3 This is a schematic diagram of the structure of the cement-modified expansive soil ecological base material preparation method according to an embodiment of the present application, in which the sieved expansive soil is evenly mixed with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture.

[0026] Figure 4 The present invention is a flowchart of a method for preparing a cement-modified expansive soil ecological substrate according to an embodiment of the present application, in which each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of multi-scale semantic feature vectors of the stirring texture.

[0027] Figure 5 The present invention is a flowchart for training the texture feature extractor based on the hole convolution layer, the multi-scale semantic feature extraction module, the node message propagation fusion network of the fusion node importance, and the classifier-based mixing controller in the preparation method of the cement-modified expansive soil ecological substrate according to the embodiment of the present application. DETAILED DESCRIPTION

[0028] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, but rather these embodiments are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0029] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0030] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0031] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0032] Cement-modified expansive soil ecological substrate is a material used for soil improvement. It improves the physical and chemical properties of expansive soil by adding cement and other additives. This modified soil is usually used for vegetation restoration and ecological slope protection. It can improve soil stability, reduce soil expansion, and increase its bearing capacity.

[0033] In the preparation process of cement-modified expansive soil ecological substrate, the sieved expansive soil and cement need to be evenly mixed to ensure that the cement is evenly distributed throughout the expansive soil substrate to improve the overall performance and consistency of the substrate. However, the traditional control of mixing time usually relies on human experience or standard operating procedures. This method is not intelligent enough and can easily lead to uneven mixing, because the operator may not be able to accurately control the mixing time of each batch, thus affecting the stability of the mixing process.

[0034] Based on this, the present application proposes a method for preparing a cement-modified expansive soil ecological substrate, which comprises: drying the expansive soil raw material, rolling and sieving it in sequence to obtain the sieved expansive soil; uniformly mixing the sieved expansive soil with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture; determining the pH value of the cement expansive soil mixture, adding ferrous sulfate accounting for 1% to 2% of the dry mass of the expansive soil to the cement expansive soil mixture, and adjusting the pH value to 5.5-7.0 to obtain the adjusted cement expansive soil; and adding the adjusted A water retaining agent accounting for 0.05-0.2% of the dry mass of the expansive soil, an organic fertilizer accounting for 0.5-2% of the dry mass of the expansive soil, a peat accounting for 3-10% of the dry mass of the expansive soil, and a polyacrylamide accounting for 0.05-0.2% of the dry mass of the expansive soil are added to cement expansive soil, and the mixture is uniformly mixed to obtain a mixture; a light compaction test is performed on the mixture to obtain the maximum dry density and the optimal moisture content of the mixture; and, based on the optimal moisture content, water is added to the mixture to mix and prepare a soil sample with the optimal moisture content to obtain the cement modified expansive soil ecological substrate.

[0035] In the process of uniformly mixing the sieved expansive soil with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture, a camera is used to collect a stirring state monitoring video in real time, and a video processing and analysis technology based on a deep learning neural network is used to perform key frame sampling and stirring texture feature analysis on the stirring state monitoring video, so as to automatically determine whether to stop stirring according to the texture time series aggregation representation features of each stirring state key frame. In this way, the need for manual monitoring can be reduced, the labor cost can be reduced, the automation level of the stirring process can be improved, and a more intelligent cement-modified expansive soil ecological substrate preparation process can be achieved.

[0036] Figure 1 Flow chart of the preparation method of cement-modified expansive soil ecological substrate according to the embodiment of the present application. Figure 1As shown, according to the preparation method of the cement-modified expansive soil ecological substrate of the embodiment of the present application, the method comprises: S110, drying the expansive soil raw material, rolling and sieving in sequence to obtain the sieved expansive soil; S120, uniformly mixing the sieved expansive soil with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture; S130, determining the pH value of the cement expansive soil mixture, and adding ferrous sulfate accounting for 1% to 2% of the dry mass of the expansive soil to the cement expansive soil mixture, and adjusting the pH value to 5.5-7.0 to obtain the adjusted cement expansive soil; S140, mixing the sieved expansive soil with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture. The adjusted cement expansive soil is added with a water retaining agent accounting for 0.05% to 0.2% of the dry mass of the expansive soil, an organic fertilizer accounting for 0.5% to 2% of the dry mass of the expansive soil, a peat accounting for 3% to 10% of the dry mass of the expansive soil, and a polyacrylamide accounting for 0.05% to 0.2% of the dry mass of the expansive soil, and the mixture is uniformly mixed to obtain a mixture; S150, a light compaction test is performed on the mixture to obtain the maximum dry density and the optimal moisture content of the mixture; and, S160, based on the optimal moisture content, water is added to the mixture to mix and prepare a soil sample with the optimal moisture content to obtain the cement modified expansive soil ecological substrate.

[0037] Figure 2 The present invention is a flow chart of uniformly mixing the sieved expansive soil with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture in the preparation method of the cement-modified expansive soil ecological substrate according to an embodiment of the present application. Figure 3 The present invention is a schematic diagram of the structure of the cement-modified expansive soil ecological substrate preparation method according to the embodiment of the present application, wherein the sieved expansive soil and cement accounting for 3% to 10% of the dry mass of the expansive soil are uniformly mixed to obtain a cement expansive soil mixture. Specifically, in the embodiment of the present application, Figure 2 and Figure 3As shown, the sieved expansive soil is uniformly mixed with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture, including: S210, obtaining a stirring state monitoring video collected by a camera; S220, performing key frame sampling on the stirring state monitoring video to obtain a time series of stirring state key frames; S230, respectively performing texture feature extraction on each stirring state key frame in the time series of the stirring state key frames to obtain a time series of stirring texture feature maps; S240, respectively passing each stirring texture feature map in the time series of the stirring texture feature map through a multi-scale semantic feature extraction module to obtain a time series of stirring texture multi-scale semantic feature vectors; S250, inputting the time series of the stirring texture multi-scale semantic feature vectors into a node message propagation fusion network of fusion node importance to obtain a stirring texture time series propagation aggregation representation vector as a stirring texture time series propagation aggregation representation feature; and, S260, obtaining a control instruction based on the stirring texture time series propagation aggregation representation feature, the control instruction being used to indicate whether to stop stirring.

[0038] In step S210, a monitoring video of the mixing state collected by a camera is obtained. It should be understood that, considering that the monitoring video of the mixing state is recorded in real time by a camera, video data showing the state changes of the expansive soil and the cement mixture during the mixing process are displayed. Based on this, in the technical solution of the present application, the monitoring video of the mixing state collected by the camera is obtained, the uniformity of the mixing can be checked, and the uneven areas or agglomeration phenomena that may occur during the mixing process can be discovered in time, which provides a basis and support for the precise control of the mixing time.

[0039] In step S220, the stirring state monitoring video is key frame sampled to obtain a time sequence of stirring state key frames. Accordingly, considering that the stirring state monitoring video contains a large number of frame images, not every frame is equally important for monitoring and analyzing the stirring state, and some background information irrelevant to the stirring process is also included. Based on this, in the technical solution of the present application, the stirring state monitoring video is key frame sampled to obtain a time sequence of stirring state key frames. It should be understood that through key frame sampling, representative stirring state frame images can be selected as key frames, and these stirring state key frames can retain important stirring state information and changes over time, so as to more accurately reflect the temporal state changes of the entire stirring process.

[0040] In step S230, texture feature extraction is performed on each stirring state key frame in the time sequence of the stirring state key frame to obtain the time sequence of the stirring texture feature map. Specifically, in an embodiment of the present application, texture feature extraction is performed on each stirring state key frame in the time sequence of the stirring state key frame to obtain the time sequence of the stirring texture feature map, including: each stirring state key frame in the time sequence of the stirring state key frame is respectively passed through a texture feature extractor based on a hole convolution layer to obtain the time sequence of the stirring texture feature map. It should be understood that considering that each stirring state key frame in the time sequence of the stirring state key frame contains information about the stirring texture feature during the stirring process, that is, the size and distribution of the expansive soil particles, and the crushing of the particles during the stirring process. The hole convolution layer can help the network better learn the subtle texture features in the image, thereby improving the representation ability and discrimination of the texture features, which is very important for distinguishing the texture features under different stirring time states. Therefore, in the technical solution of the present application, each stirring state key frame in the time series of the stirring state key frames is respectively passed through a texture feature extractor based on a hole convolution layer to capture and extract the local texture features in each key frame, thereby obtaining a time series of stirring texture feature maps.

[0041] In step S240, each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of multi-scale semantic feature vectors of the stirring texture. Accordingly, considering that each stirring texture feature map in the time series of the stirring texture feature map expresses global and local semantic feature information about the stirring texture, such as the overall distribution of cement and expansive soil in the entire stirring container and the crushing and refinement of particles. Based on this, in order to more comprehensively understand and grasp the scene understanding of the stirring texture at each time point, in the technical solution of the present application, each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of multi-scale semantic feature vectors of the stirring texture. In detail, the multi-scale semantic feature extraction module extracts the global stirring texture semantic feature information and the local stirring texture semantic feature information in each stirring texture feature map respectively to capture the background characteristics of the overall stirring temporal state and the temporal changes of the local texture characteristics of the stirring state respectively, and then fuses the extracted global and local semantic features of the stirring texture to comprehensively consider the overall stirring state and local texture characteristics in the stirring texture feature map at each time point, thereby obtaining a time series of multi-scale semantic feature vectors of the stirring texture.

[0042] Figure 4The present invention is a flowchart of a method for preparing a cement-modified expansive soil ecological substrate according to an embodiment of the present application, wherein each stirring texture feature map in the time series of the stirring texture feature map is respectively subjected to a multi-scale semantic feature extraction module to obtain a time series of a stirring texture multi-scale semantic feature vector. Specifically, in an embodiment of the present application, Figure 4 As shown, each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of stirring texture multi-scale semantic feature vectors, including: S310, extracting a stirring texture global semantic feature vector from the stirring texture feature map; S320, extracting a stirring texture local semantic feature vector from the stirring texture feature map; and, S330, fusing the stirring texture global semantic feature vector and the stirring texture local semantic feature vector to obtain the stirring texture multi-scale semantic feature vector.

[0043] More specifically, in an embodiment of the present application, a global semantic feature vector of a stirring texture is extracted from the stirring texture feature map, including: upsampling the stirring texture feature map to obtain an upsampled stirring texture feature map; performing global mean pooling processing on each feature matrix along the channel dimension in the upsampled stirring texture feature map to obtain an upsampled stirring texture semantic channel feature vector; and performing point convolution encoding on the upsampled stirring texture semantic channel feature vector to obtain the global semantic feature vector of the stirring texture.

[0044] More specifically, in an embodiment of the present application, extracting a local semantic feature vector of the stirring texture from the stirring texture feature map includes: upsampling the stirring texture feature map to obtain an upsampled stirring texture feature map; performing point convolution encoding on the upsampled stirring texture feature map to obtain a channel modulated upsampled stirring texture feature map; and performing two-dimensional convolution encoding on the channel modulated upsampled stirring texture feature map to obtain the local semantic feature vector of the stirring texture.

[0045] In step S250, the time series of the stirring texture multi-scale semantic feature vector is input into the node message propagation fusion network of the fusion node importance to obtain the stirring texture time series propagation aggregation representation vector as the stirring texture time series propagation aggregation representation feature. It should be understood that, considering that stirring is a dynamic process, the time series of the stirring texture multi-scale semantic feature vector reflects the semantic features of the stirring texture at each time point, and each stirring texture multi-scale semantic feature vector also has mutual correlation and dependency. Therefore, in order to more clearly understand the evolution of the stirring texture features over time, so as to better capture the temporal dynamic characteristics of the stirring process, in the technical solution of the present application, the time series of the stirring texture multi-scale semantic feature vector is input into the node message propagation fusion network of the fusion node importance to obtain the stirring texture time series propagation aggregation representation vector. It is worth mentioning that the node message propagation is a commonly used technology in graph neural networks. By transmitting information on the graph structure, each node can gather information from surrounding nodes, thereby better characterizing the features of the entire graph. That is to say, in this scheme, the time series of the stirring texture feature vector can be regarded as a graph structure, and the stirring texture multi-scale semantic feature vector of each time step can be regarded as a node. The node message propagation fusion network propagation through the fusion node importance can make the network more flexibly process the stirring texture multi-scale semantic features of different nodes by considering the importance and contribution of each node, so as to better characterize and understand the characteristics of the stirring texture state and the overall temporal changes.

[0046] Specifically, in an embodiment of the present application, the time series of the stirred texture multi-scale semantic feature vector is input into a node message propagation fusion network that fuses the node importance to obtain a stirred texture temporal propagation aggregate representation vector as a stirred texture temporal propagation aggregate representation feature, including: using a scoring function to measure the saliency factors of the first n-1 stirred texture multi-scale semantic feature vectors in the time series of the stirred texture multi-scale semantic feature vector to obtain n-1 stirred texture multi-scale semantic saliency factors; using the n-1 stirred texture multi-scale semantic saliency factors as weights, calculating the position-weighted sum of the first n-1 stirred texture multi-scale semantic feature vectors in the time series of the stirred texture multi-scale semantic feature vector to aggregate the stirred texture multi-scale semantic features. The first n-1 stirring texture multi-scale semantic feature vectors in the time series of the stirring texture multi-scale semantic feature vector are used to obtain the stirring texture multi-scale semantic aggregation representation vector of the rear node set; the nth stirring texture multi-scale semantic feature vector in the time series of the stirring texture multi-scale semantic feature vector is linearly modulated using a preset hyperparameter to obtain the modulated nth stirring texture multi-scale semantic feature vector; the positional sum of the modulated nth stirring texture multi-scale semantic feature vector and the rear node set stirring texture multi-scale semantic aggregation representation vector is calculated to obtain the stirring texture multi-scale semantic full-node aggregation representation vector; and the stirring texture multi-scale semantic full-node aggregation representation vector is input into a multi-layer perceptron model to obtain the stirring texture temporal propagation aggregation representation vector.

[0047] More specifically, in an embodiment of the present application, a scoring function is used to measure the salient factors of the first n-1 stirring texture multi-scale semantic feature vectors in the time series of the stirring texture multi-scale semantic feature vector to obtain n-1 stirring texture multi-scale semantic salient factors, including: respectively calculating the matrix product between each weight coefficient matrix and the first n-1 stirring texture multi-scale semantic feature vectors to obtain the first n-1 stirring texture multi-scale semantic weighted feature vectors; adding the corresponding bias vector to each stirring texture multi-scale semantic weighted feature vector in the first n-1 stirring texture multi-scale semantic weighted feature vectors to obtain the first n-1 stirring texture multi-scale semantic weighted feature vectors; The invention relates to a method for obtaining a plurality of stirred texture multi-scale semantic bias feature vectors; inputting the first n-1 stirred texture multi-scale semantic bias feature vectors into a sigmoid function for activation to obtain the first n-1 stirred texture multi-scale semantic activation feature vectors; calculating the product between each stirred texture multi-scale semantic activation feature vector in the first n-1 stirred texture multi-scale semantic activation feature vectors and a weight coefficient vector to obtain the first n-1 stirred texture multi-scale semantic associations; and inputting the first n-1 stirred texture multi-scale semantic associations into a softmax function to obtain the n-1 stirred texture multi-scale semantic saliency factors.

[0048] In an embodiment of the present application, specifically, the time series of the stirring texture multi-scale semantic feature vector is input into a node message propagation fusion network of fused node importance to obtain the stirring texture temporal propagation aggregation representation vector, including: using the node message propagation fusion network of fused node importance, processing the time series of the stirring texture multi-scale semantic feature vector with the following propagation fusion formula to obtain the stirring texture temporal propagation aggregation representation vector; wherein the propagation fusion formula is:

[0049]

[0050]

[0051] Among them, v i represents the i-th stirring texture multi-scale semantic feature vector in the time series of the stirring texture multi-scale semantic feature vector, W i represents the weight coefficient matrix, B i is the bias vector, represents the weight coefficient vector, sigmoid(·) represents the sigmoid function, softmax(·) represents the softmax function, e(v i ) represents the stirring texture multi-scale semantic saliency factor at each position in the sequence of stirring texture multi-scale semantic saliency factors, n is the number of feature vectors in the time series of the stirring texture multi-scale semantic feature vector, v n represents the nth stirring texture multi-scale semantic feature vector in the time series of the stirring texture multi-scale semantic feature vector, ∈ is the preset hyperparameter, MLP(·) represents a multi-layer perceptron, v c A representation vector for the temporal propagation aggregation of the stirring texture.

[0052] In step S260, a control instruction is obtained based on the stirring texture time-series propagation aggregation representation feature, and the control instruction is used to indicate whether to stop stirring. Specifically, in an embodiment of the present application, a control instruction is obtained based on the stirring texture time-series propagation aggregation representation feature, including: inputting the stirring texture time-series propagation aggregation representation vector into a classifier-based stirring controller to obtain the control instruction. That is, the stirring texture time-series propagation aggregation representation feature obtained by node message fusion using the time series of the stirring texture multi-scale semantic feature vector is used for classification processing to automatically determine whether to stop stirring. In this way, the need for manual monitoring can be reduced, the labor cost can be reduced, the automation level of the stirring process can be improved, and a more intelligent cement-modified expansive soil ecological substrate preparation process can be achieved.

[0053] It is worth mentioning that those skilled in the art should be aware that before applying a deep neural network model for inference, the deep neural network model must be trained so that the deep neural network can implement specific functional capabilities.

[0054] Specifically, in an embodiment of the present application, a training step is also included: for training the texture feature extractor based on the hole convolution layer, the multi-scale semantic feature extraction module, the node message propagation fusion network of the fusion node importance, and the classifier-based stirring controller.

[0055] Figure 5 The present invention is a flowchart of training the texture feature extractor based on the hole convolution layer, the multi-scale semantic feature extraction module, the node message propagation fusion network of the fusion node importance, and the mixing controller based on the classifier in the preparation method of the cement-modified expansive soil ecological substrate according to the embodiment of the present application. Figure 5 As shown, the training step includes: S410, obtaining training data, wherein the training data includes a stirring state monitoring video collected by a camera, and a real control instruction, wherein the real control instruction is a real value of whether to stop stirring; S420, performing key frame sampling on the training stirring state monitoring video to obtain a time sequence of training stirring state key frames; S430, passing each training stirring state key frame in the time sequence of the training stirring state key frames through the texture feature extractor based on the hole convolution layer to obtain a time sequence of training stirring texture feature maps; S440, passing each training stirring texture feature map in the time sequence of the training stirring texture feature map through the multi-scale semantic feature extraction module to obtain a time sequence of training stirring texture multi-scale semantic feature vectors; Column; S450, input the time series of the training stirring texture multi-scale semantic feature vector into the node message propagation fusion network of the fusion node importance to obtain the training stirring texture time series propagation aggregation representation vector; S460, pass the training stirring texture time series propagation aggregation representation vector through the classifier-based stirring controller to obtain the training control instruction; S470, calculate the cross entropy loss function value between the training control instruction and the real control instruction to obtain the classification loss function value; and, S480, based on the classification loss function value and through gradient descent back propagation to train the texture feature extractor based on the void convolution layer, the multi-scale semantic feature extraction module, the node message propagation fusion network of the fusion node importance and the classifier-based stirring controller.

[0056] It should be understood that here, considering the multi-scale image semantic feature non-correspondence of the source image semantics of the stirring state key frame based on the individual training stirring texture multi-scale semantic feature vectors in the time series of the training stirring texture multi-scale semantic feature vectors, when the node message propagation fusion of the fusion node importance is fused, the training stirring texture temporal propagation aggregation representation vector will also have the image semantic distribution knowledge offset caused by the node importance propagation difference with the image semantic feature distribution, thereby causing the uncertainty of its image semantic feature distribution relative to the class probability understanding of the classifier-based stirring controller, reducing the speed of classification training and the accuracy of classification results.

[0057] Based on this, in a preferred embodiment, the training stirring texture timing propagation aggregate representation vector is input into the classifier-based stirring controller to obtain a training control instruction, including: determining the stop class probability value corresponding to the indication of stopping stirring obtained by the training stirring texture timing propagation aggregate representation vector through the classifier-based stirring controller, and subtracting the stop class probability value from one to obtain a non-stop class probability value; respectively calculating a power function of each eigenvalue of the training stirring texture timing propagation aggregate representation vector with the stop class probability value as an exponent, and a power function with the non-stop class probability value as an exponent, to obtain a training stirring texture timing propagation aggregate stop class vector and a training stirring texture timing propagation aggregate non-stop class vector; and comparing the training stop class probability value with the training stirring texture timing propagation aggregate representation vector. The first training stirring texture timing propagation aggregation non-stop class vector is point multiplied with the training stirring texture timing propagation aggregation stop class vector to obtain the first training stirring texture timing propagation aggregation cross class vector, and the non-stop class probability value is point multiplied with the training stirring texture timing propagation aggregation stop class vector to obtain the second training stirring texture timing propagation aggregation cross class vector; after point multiplying the first training stirring texture timing propagation aggregation cross class vector with the second training stirring texture timing propagation aggregation cross class vector, further point multiplication is performed with the stop class probability value and the training stirring texture timing propagation aggregation stop class vector to obtain the optimized training stirring texture timing propagation aggregation representation vector; and the optimized training stirring texture timing propagation aggregation representation vector is input into the classifier-based stirring controller to obtain the training control instruction.

[0058] That is, when the training stirring texture time series propagation aggregate representation vector is classified through a classifier-based stirring controller, in order to achieve unsupervised domain adaptation from the feature distribution domain of the training stirring texture time series propagation aggregate representation vector to the probability distribution domain of the classification probability, the stop class probability value and the non-stop class probability value obtained by the training stirring texture time series propagation aggregate representation vector through the classifier-based stirring controller are used as domain agents, and the training stirring texture time series propagation aggregate representation vector is respectively based on the moving average of the probability distribution of the training stirring texture time series propagation aggregate stop class vector and the training stirring texture time series propagation aggregate non-stop class vector through the power function class distribution interaction as an exponential, and the knowledge transfer from unlabeled classification features to labeled probability distribution is achieved by superimposing the feature domain probability distribution knowledge of the training stirring texture time series propagation aggregate representation vector itself, thereby promoting the classification operation of the training stirring texture time series propagation aggregate representation vector through the classifier-based stirring controller, that is, improving the classification training speed of the classifier-based stirring controller and the accuracy of control instructions. In this way, the need for manual monitoring can be reduced, labor costs can be reduced, and the level of automation of the mixing process can be improved, thereby achieving a more intelligent cement-modified expansive soil ecological substrate preparation process.

[0059] In an embodiment of the present application, specifically, when the training stirring texture temporal propagation aggregation representation vector is input into a classifier-based stirring controller to obtain a training control instruction, the training stirring texture temporal propagation aggregation representation vector is optimized to obtain an optimized training stirring texture temporal propagation aggregation representation vector, including: optimizing the training stirring texture temporal propagation aggregation representation vector according to the following optimization formula to obtain the optimized training stirring texture temporal propagation aggregation representation vector; wherein the optimization formula is:

[0060] V′=(p⊙V 1-p )⊙[(1-)⊙V p ]⊕(p⊙V p )

[0061] Wherein, V is the training stirring texture temporal propagation aggregate representation vector, p is the stop class probability value corresponding to the stop stirring obtained by the training stirring texture temporal propagation aggregate representation vector through the classifier-based stirring controller, 1- is the non-stop class probability value, V 1-p is the training stirring texture sequential propagation aggregation without stopping the class vector, V p is the training stirring texture temporal propagation aggregation stop class vector, ⊙ is the position point multiplication, ⊕ is the position point addition, and V′ is the optimized training stirring texture temporal propagation aggregation representation vector.

[0062] In summary, the preparation method of the cement-modified expansive soil ecological substrate based on the embodiment of the present application is explained, which is collected by a camera in real time to monitor the stirring state, and the video processing and analysis technology based on the deep learning neural network is used to perform key frame sampling and stirring texture feature analysis on the stirring state monitoring video, so as to automatically judge whether to stop stirring according to the texture time series aggregation representation characteristics of each stirring state key frame. In this way, the need for manual monitoring can be reduced, the labor cost can be reduced, the automation level of the stirring process can be improved, and a more intelligent cement-modified expansive soil ecological substrate preparation process can be achieved.

[0063] Specifically, a cement-modified expansive soil ecological substrate is also provided, and the cement-modified expansive soil ecological substrate is prepared by the aforementioned preparation method of the cement-modified expansive soil ecological substrate.

[0064] The foregoing is merely an example of the principles of the present disclosure, and various modifications may be made by those skilled in the art without departing from the scope of the present disclosure. The above embodiments are presented for the purpose of illustration and not limitation. The present disclosure may also take many forms other than those explicitly described herein. Therefore, it is emphasized that the present disclosure is not limited to the explicitly disclosed methods and apparatus, but is intended to include variations and modifications within the spirit and scope of the appended claims.

Claims

1. A method for preparing a cement-modified expansive soil ecological substrate, characterized in that: include: The expansive soil raw material is dried, then rolled and sieved in sequence to obtain sieved expansive soil; The sieved expansive soil is uniformly mixed with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture; Determine the pH value of the cement expansive soil mixture, add ferrous sulfate accounting for 1% to 2% of the dry mass of the expansive soil to the cement expansive soil mixture, and adjust the pH value to 5.5-7.0 to obtain the adjusted cement expansive soil; Adding 0.05% to 0.2% of a water retaining agent by dry weight of the expansive soil, 0.5% to 2% of an organic fertilizer by dry weight of the expansive soil, 3% to 10% of peat by dry weight of the expansive soil, and 0.05% to 0.2% of polyacrylamide by dry weight of the expansive soil into the adjusted cement expansive soil, and uniformly mixing to obtain a mixture; Performing a light compaction test on the mixture to obtain the maximum dry density and optimal moisture content of the mixture; Based on the optimal moisture content, water is added to the mixture to mix the mixture to prepare a soil sample with the optimal moisture content to obtain the cement-modified expansive soil ecological substrate; The sieved expansive soil is uniformly mixed with cement accounting for 3% to 10% of the dry mass of the expansive soil to obtain a cement expansive soil mixture, comprising: Obtain stirring status monitoring video collected by the camera; Performing key frame sampling on the stirring state monitoring video to obtain a time sequence of stirring state key frames; Extracting texture features of each stirring state key frame in the time sequence of the stirring state key frames to obtain a time sequence of stirring texture feature maps; Each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of stirring texture multi-scale semantic feature vectors; Inputting the time series of the stirring texture multi-scale semantic feature vector into a node message propagation fusion network that fuses node importance to obtain a stirring texture temporal propagation aggregation representation vector as a stirring texture temporal propagation aggregation representation feature; Based on the time series propagation aggregation representation feature of the stirring texture, a control instruction is obtained, and the control instruction is used to indicate whether to stop stirring.

2. The method for preparing the cement-modified expansive soil ecological substrate according to claim 1, characterized in that: The method comprises: performing texture feature extraction on each stirring state key frame in the time sequence of the stirring state key frames to obtain a time sequence of stirring texture feature maps, including: passing each stirring state key frame in the time sequence of the stirring state key frames through a texture feature extractor based on a dilated convolutional layer to obtain a time sequence of the stirring texture feature maps.

3. The method for preparing the cement-modified expansive soil ecological substrate according to claim 2, characterized in that: Each stirring texture feature map in the time series of the stirring texture feature map is respectively passed through a multi-scale semantic feature extraction module to obtain a time series of stirring texture multi-scale semantic feature vectors, including: Extracting a global semantic feature vector of the stirring texture from the stirring texture feature map; Extracting a local semantic feature vector of the stirring texture from the stirring texture feature map; The stirring texture global semantic feature vector and the stirring texture local semantic feature vector are fused to obtain the stirring texture multi-scale semantic feature vector.

4. The method for preparing the cement-modified expansive soil ecological substrate according to claim 3, characterized in that: Extracting a global semantic feature vector of the stirring texture from the stirring texture feature map includes: Upsampling the stirring texture feature map to obtain an upsampled stirring texture feature map; Performing global mean pooling processing on each feature matrix along the channel dimension in the upsampled stirred texture feature map to obtain an upsampled stirred texture semantic channel feature vector; Point convolution encoding is performed on the upsampled stirring texture semantic channel feature vector to obtain the stirring texture global semantic feature vector.

5. The method for preparing the cement-modified expansive soil ecological substrate according to claim 4, characterized in that: Extracting a local semantic feature vector of the stirring texture from the stirring texture feature map includes: Upsampling the stirring texture feature map to obtain an upsampled stirring texture feature map; Performing point convolution encoding on the upsampled stirred texture feature map to obtain a channel modulated upsampled stirred texture feature map; Two-dimensional convolution coding is performed on the channel-modulated up-sampled stirring texture feature map to obtain the stirring texture local semantic feature vector.

6. The method for preparing the cement-modified expansive soil ecological substrate according to claim 5, characterized in that: Inputting the time series of the stirring texture multi-scale semantic feature vector into the node message propagation fusion network of the fused node importance to obtain the stirring texture temporal propagation aggregation representation vector as the stirring texture temporal propagation aggregation representation feature, including: Using a scoring function to measure the salient factors of the first n-1 stirring texture multi-scale semantic feature vectors in the time series of the stirring texture multi-scale semantic feature vectors to obtain n-1 stirring texture multi-scale semantic salient factors; Taking the n-1 stirring texture multi-scale semantic saliency factors as weights, calculating the position-weighted sum of the first n-1 stirring texture multi-scale semantic feature vectors in the time series of the stirring texture multi-scale semantic feature vectors to aggregate the first n-1 stirring texture multi-scale semantic feature vectors in the time series of the stirring texture multi-scale semantic feature vectors to obtain a stirring texture multi-scale semantic aggregate representation vector of the rear node set; Using a preset hyperparameter, linearly modulate the nth stirring texture multi-scale semantic feature vector in the time series of the stirring texture multi-scale semantic feature vector to obtain the modulated nth stirring texture multi-scale semantic feature vector; Calculating the positional sum of the nth stirred texture multi-scale semantic feature vector after the modulation and the stirred texture multi-scale semantic aggregate representation vector of the rear node set to obtain a stirred texture multi-scale semantic full-node aggregate representation vector; The stirring texture multi-scale semantic full-node aggregate representation vector is input into a multi-layer perceptron model to obtain the stirring texture temporal propagation aggregate representation vector.

7. The method for preparing the cement-modified expansive soil ecological substrate according to claim 6, characterized in that: The scoring function is used to measure the salient factors of the first n-1 stirring texture multi-scale semantic feature vectors in the time series of the stirring texture multi-scale semantic feature vector to obtain n-1 stirring texture multi-scale semantic salient factors, including: Calculating the matrix products of each weight coefficient matrix and the first n-1 multi-scale semantic feature vectors of the stirring texture respectively to obtain the first n-1 multi-scale semantic weighted feature vectors of the stirring texture; Adding a corresponding bias vector to each of the first n-1 stirred texture multi-scale semantic weighted feature vectors to obtain first n-1 stirred texture multi-scale semantic biased feature vectors; Input the first n-1 stirring texture multi-scale semantic bias feature vectors into a sigmoid function for activation to obtain the first n-1 stirring texture multi-scale semantic activation feature vectors; Calculating the product of each stirring texture multi-scale semantic activation feature vector in the first n-1 stirring texture multi-scale semantic activation feature vectors and the weight coefficient vector to obtain the first n-1 stirring texture multi-scale semantic association degree; The first n-1 stirring texture multi-scale semantic associations are input into a softmax function to obtain the n-1 stirring texture multi-scale semantic saliency factors.

8. The method for preparing the cement-modified expansive soil ecological substrate according to claim 7, characterized in that: Based on the stirring texture temporal propagation aggregation representation feature, a control instruction is obtained, including: inputting the stirring texture temporal propagation aggregation representation vector into a classifier-based stirring controller to obtain the control instruction.

9. The method for preparing the cement-modified expansive soil ecological substrate according to claim 8, characterized in that: The method further includes a training step: for training the texture feature extractor based on the hole convolution layer, the multi-scale semantic feature extraction module, the node message propagation fusion network of the fusion node importance, and the stirring controller based on the classifier; Wherein, the training step includes: Acquire training data, wherein the training data includes a monitoring video of a stirring state collected by a camera, and a real control instruction, wherein the real control instruction is a real value of whether to stop stirring; Performing key frame sampling on the training stirring state monitoring video to obtain a time series of training stirring state key frames; Each training stirring state key frame in the time sequence of the training stirring state key frames is passed through the texture feature extractor based on the hole convolution layer to obtain a time sequence of training stirring texture feature maps; Each training stirring texture feature map in the time series of the training stirring texture feature map is respectively passed through the multi-scale semantic feature extraction module to obtain a time series of training stirring texture multi-scale semantic feature vectors; Inputting the time series of the training stirring texture multi-scale semantic feature vector into the node message propagation fusion network of the fusion node importance to obtain the training stirring texture time series propagation aggregation representation vector; Passing the training stirring texture temporal propagation aggregation representation vector through the classifier-based stirring controller to obtain training control instructions; Calculating a cross entropy loss function value between the training control instruction and the real control instruction to obtain a classification loss function value; The texture feature extractor based on the hole convolution layer, the multi-scale semantic feature extraction module, the node message propagation fusion network for fusion node importance, and the classifier-based stirring controller are trained based on the classification loss function value and through back propagation of gradient descent.

10. A cement-modified expansive soil ecological substrate, characterized in that: The cement-modified expansive soil ecological substrate is prepared by the preparation method of the cement-modified expansive soil ecological substrate according to any one of claims 1-9.

Citation Information

Patent Citations

  • Cement modified expansive soil, preparation method thereof and cement modified expansive soil ecological base material

    CN114031338A

  • Train brake valve cleaning agent and preparation method thereof

    CN117586838A

  • Preparation method and application of UHPC tunnel flue plate with fireproof function

    CN117863315A

  • New energy automobile speed distributed monitoring system and method

    CN118219851A

  • Equipment measurement data processing method based on deep learning

    CN118277888A