Method and device for predicting physicochemical properties of coal gangue biologically modified soil
Through the improved ResNet model combined with the CBAM attention mechanism, the problem of uneven corruption in the biomodified soiling process of coal gangue is solved, and rapid and accurate physical and chemical characteristics prediction is achieved, reducing detection costs and time.
Patent Information
- Application Number
- CN202510368567.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-26
- Publication Date
- 2025-07-04
AI Technical Summary
The prior art has uneven degree of corruption in the biomodified soiling process of coal gangue, resulting in the need of multi-point sampling and testing, which consumes time and resources, making it difficult to achieve rapid and accurate physical and chemical characteristics prediction.
The improved ResNet model is adopted, combined with the 3×3 convolutional layer of the BottleNeck block and the CBAM attention mechanism layer, and a deep learning model is constructed to predict the physical and chemical characteristics of coal gangue biomodified soil, and to achieve fast and accurate characteristic prediction through image recognition.
It improves the accuracy and speed of prediction of physical and chemical characteristics of biologically modified soil of coal gangue, reduces detection costs and time, and improves the detailed identification ability of the model.
Smart Images
Figure CN120260753A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of resource recycling, and particularly to a method and device for predicting the physical and chemical properties of coal gangue bio-modified soil. Background Art
[0002] Coal gangue is derived from strata and is homologous to soil. In line with the concept of "coming from the soil and returning to the soil", it is undoubtedly one of the best ways to utilize coal gangue by modifying it into soil through biological fermentation. Currently, the transformation of coal gangue into soil is achieved by biological fermentation. However, in the process, there is a problem of uneven degree of decomposition, and it is necessary to detect the physical and chemical properties of soil at different parts through the method of multi-point sampling and testing. This process consumes time, energy and resources, and there is an urgent need to develop a method for predicting the physical and chemical properties of coal gangue bio-modified soil that is fast, accurate and low-cost.
[0003] In recent years, researchers have used means such as images or near-infrared spectra to construct intelligent models. The advantage is that the degree of decomposition of organic agricultural and forestry waste piles can be predicted through changes in images or spectra. However, this method can only provide category information, such as classifying images as decomposed or undecomposed. However, in the prediction of the properties of coal gangue bio-modified soil, in addition to judging whether it is decomposed, it is also necessary to predict quantifiable physical and chemical properties. To achieve the prediction of physical and chemical properties, it is necessary to modify the network activation function and the fully connected layer to transform the model applicable to image classification into a model applicable to image regression. However, through experiments, it is found that simply modifying the activation function and the fully connected layer is not sufficient to achieve the purpose of accurately predicting characteristic parameters. Summary of the Invention
[0004] The purpose of the present application is to provide a method and device for predicting the physical and chemical properties of coal gangue bio-modified soil, which can accurately and quickly predict the physical and chemical properties of coal gangue bio-modified soil.
[0005] To achieve the above purpose, the present application provides the following solutions:
[0006] In the first aspect, the present application provides a method for predicting the physical and chemical properties of coal gangue bio-modified soil, including:
[0007] Obtain an image to be predicted of coal gangue bio-modified soil;
[0008] Input the to-be-predicted image of the coal gangue biologically modified soil into the improved ResNet model to predict the physical and chemical properties of the coal gangue biologically modified soil. The improved ResNet model is a model obtained by improving the ResNet101 model. The improvement includes using the 3×3 convolutional layer of the Bottleneck block in the ResNet101 model for downsampling operations and connecting the CBAM attention mechanism layer between every two adjacent Bottleneck blocks after the maximum pooling layer of the ResNet101 model.
[0009] In a second aspect, the present application provides a device for predicting the physical and chemical properties of coal gangue biologically modified soil, including:
[0010] An image acquisition module, configured to acquire the to-be-predicted image of the coal gangue biologically modified soil;
[0011] A prediction module, configured to input the to-be-predicted image of the coal gangue biologically modified soil into the improved ResNet model to predict the physical and chemical properties of the coal gangue biologically modified soil. The improved ResNet model is a model obtained by improving the ResNet101 model. The improvement includes using the 3×3 convolutional layer of the Bottleneck block in the ResNet101 model for downsampling operations and connecting the CBAM attention mechanism layer between every two adjacent Bottleneck blocks after the maximum pooling layer of the ResNet101 model.
[0012] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned method for predicting the physical and chemical properties of coal gangue biologically modified soil.
[0013] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the above-mentioned method for predicting the physical and chemical properties of coal gangue biologically modified soil.
[0014] In a fifth aspect, the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the above-mentioned method for predicting the physical and chemical properties of coal gangue biologically modified soil.
[0015] According to the specific embodiments provided by the present application, the following technical effects are disclosed:
[0016] The present application provides a method and device for predicting the physical and chemical properties of coal gangue biologically modified soil, obtaining an image to be predicted of the coal gangue biologically modified soil; inputting the image to be predicted of the coal gangue biologically modified soil into an improved ResNet model to predict the physical and chemical properties of the coal gangue biologically modified soil; the improved ResNet model is a model obtained by improving the ResNet101 model; the improvement includes using the 3×3 convolutional layer of the BottleNeck block in the ResNet101 model for downsampling operation and connecting a CBAM attention mechanism layer between every two adjacent BottleNeck blocks after the maximum pooling layer of the ResNet101 model. Among them, combining ResNet and CBAM can make full use of the deep feature extraction ability of ResNet and the ability of CBAM to adaptively increase the weight of important features, construct an efficient and accurate deep learning model, and improve the accuracy and rapidity of prediction. Using the 3×3 convolutional layer of the BottleNeck block in the ResNet101 model for downsampling operation can reduce the information loss when executing the 1×1 convolutional layer, improve the detail recognition ability of the model, and thus improve the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.
[0018] Figure 1 It is an application environment diagram of a method for predicting the physical and chemical properties of coal gangue biologically modified soil in an embodiment of the present application;
[0019] Figure 2 It is a flowchart of a method for predicting the physical and chemical properties of coal gangue biologically modified soil provided in an embodiment of the present application;
[0020] Figure 3 It is a schematic diagram of the CBAM attention mechanism structure provided in an embodiment of the present application;
[0021] Figure 4 It is a schematic diagram of the structure of a ResNet-CBAM network model provided in an embodiment of the present application;
[0022] Figure 5 It is a schematic diagram of an image enhancement example provided in an embodiment of the present application;
[0023] Figure 6 It is a schematic diagram of a test set example image provided in an embodiment of the present application;
[0024] Figure 7 Schematic diagram of functional modules of a device for predicting physical and chemical properties of coal gangue bio - modified soil provided by an embodiment of the present application;
[0025] Figure 8 Schematic diagram of the structure of a computer device provided by an embodiment of the present application. Specific embodiments
[0026] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
[0027] To make the above - mentioned objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the drawings and specific embodiments.
[0028] The method for predicting the physical and chemical properties of coal gangue bio - modified soil provided by the embodiments of the present application can be applied to the application environment as shown in Figure 1 . Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be set up separately, integrated on the server, placed in the cloud or on other servers. The terminal can send the image of the coal gangue bio - modified soil to be predicted to the server. After receiving the image of the coal gangue bio - modified soil to be predicted, the server inputs the image of the coal gangue bio - modified soil to be predicted into the improved ResNet model to predict the physical and chemical properties of the coal gangue bio - modified soil; the improved ResNet model is a model obtained by improving the ResNet101 model; the improvement includes performing down - sampling operations using the 3×3 convolutional layer of the BottleNeck block in the ResNet101 model and connecting a CBAM attention mechanism layer between every two adjacent BottleNeck blocks after the maximum pooling layer of the ResNet101 model. The server can feedback the predicted physical and chemical properties of the coal gangue bio - modified soil to the terminal. In addition, in some embodiments, the method for predicting the physical and chemical properties of the coal gangue bio - modified soil can also be implemented by the server or the terminal alone. For example, the terminal can directly predict the physical and chemical properties of the coal gangue bio - modified soil for the image of the coal gangue bio - modified soil to be predicted, or the server can obtain the image of the coal gangue bio - modified soil to be predicted from the data storage system and perform the prediction of the physical and chemical properties of the coal gangue bio - modified soil.
[0029] Among them, the terminal can be, but is not limited to, various desktop computers, laptop computers, smartphones, tablet computers, Internet of Things devices, and portable wearable devices. The server can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.
[0030] In an exemplary embodiment, to predict the physical and chemical properties of coal gangue bio-modified soil, while constructing a deep feature extraction and understanding model, the recognition ability of the model for important features and the acquisition ability of detailed features are improved, which is expected to strengthen the recognition and understanding of the key detailed features of coal gangue bio-modified soil, thereby improving the performance of the intelligent model in the task of predicting the physical and chemical properties of coal gangue bio-modified soil and achieving rapid and accurate prediction of the physical and chemical properties of coal gangue bio-modified soil. To facilitate the popularization of this technology, as Figure 2 shown, a method for predicting the physical and chemical properties of coal gangue bio-modified soil is provided. Using this method, users can quickly obtain the prediction results with only simple operations. This method is executed by a computer device, specifically, it can be executed alone by a computer device such as a terminal or a server, or jointly executed by a terminal and a server. In the embodiment of the present application, taking this method applied to Figure 1 the server in it as an example for illustration, it includes the following steps 101 to step 102.
[0031] Step 101, obtain the image of the coal gangue bio-modified soil to be predicted.
[0032] Step 102, input the image of the coal gangue bio-modified soil to be predicted into the improved ResNet model, and predict the physical and chemical properties of the coal gangue bio-modified soil; the improved ResNet model is a model obtained by improving the ResNet101 model; the improvement includes using the 3×3 convolutional layer of the Bottleneck block in the ResNet101 model for downsampling operations and connecting the CBAM attention mechanism layer between every two adjacent Bottleneck blocks after the maximum pooling layer of the ResNet101 model.
[0033] As Figure 3 shown, the CBAM self-attention mechanism combines the channel attention mechanism Mc and the spatial attention mechanism Ms, performs Attention on channels and spaces respectively, and adaptively adjusts the weights of each channel and each position in the feature map, which helps to increase the proportion of the detailed features of the model image in the model, thereby improving the accuracy of feature extraction and the overall performance of the model.
[0034] Implementing the above steps 101 to 102, the improved ResNet model is an improvement based on the ResNet101 model. Among them, the ResNet101 model solves the problem of gradient disappearance in deep neural networks by introducing residual blocks, enabling the network to be trained at a deeper level. Combining ResNet with CBAM can make full use of the deep feature extraction ability of ResNet and the ability of CBAM to adaptively increase the weights of important features to construct an efficient and accurate deep learning model, improving the efficiency and accuracy of predicting the physical and chemical properties of coal gangue bio-modified soil. In addition, the structures of conv2_x, conv3_x, conv4_x, and conv5_x of the ResNet101 model will be modified, and the downsampling will be transferred from the 1×1 convolutional layer to the 3×3 convolutional layer to obtain Newconv2_x, Newconv3_x, Newconv4_x, and Newconv5_x. The new structure can reduce the information loss of each 1×1 convolutional layer and improve the detail recognition ability of the model. The structures of conv2_x, conv3_x, conv4_x, and conv5_x are Bottleneck blocks.
[0035] In another exemplary embodiment of this application, in addition to the improvements given above, the improvements also include: modifying the AdaptiveAvgPool2d layer and the fully connected layer in the ResNet101 model into an AvgPool2d layer and a convolutional layer respectively.
[0036] Changing the AdaptiveAvgPool2d and the fully connected layer in the ResNet101 model to an AvgPool2d and a 1×1, 12 convolutional layer. The AdaptiveAvgPool2d and the fully connected layer in the ResNet101 model can well extract the overall features of the image, but in the images of coal gangue bio-modified soil, more attention needs to be paid to local detail features, and the proportion of overall features is relatively small. Therefore, it is replaced with a combination of an AvgPool2d and a 1×1, 12 convolutional layer that is more suitable for identifying local features to identify local detail features and improve the prediction accuracy.
[0037] In another exemplary embodiment of this application, in addition to the improvements given above, the improvements also include: replacing the softmax output layer used to complete the classification task in the ResNet101 model with an MSE output layer.
[0038] Replacing the softmax output layer used to complete the classification task in the ResNet101 model with an MSE output layer suitable for regression tasks. The mean squared error MSE is the most commonly used error in regression loss, and its formula is:
[0039]
[0040] In the formula, n is the number of images, and f(x i ) is the model prediction value, and y i is the actual value.
[0041] In another exemplary embodiment of the present application, as Figure 4 shown, a ResNet-CBAM network model structure that focuses on identifying image detail features and can predict multiple physical and chemical properties of coal gangue bio-modified soil. The improved ResNet model includes: an input layer, an initial convolutional layer (corresponding to Figure 4 the 7×7 Convolution in Figure 4 ), a max pooling layer (corresponding to Figure 4 the 3×3 Maxpool in Figure 4 ), a first CBAM attention mechanism layer, a first Bottleneck (corresponding to Figure 4 the Newconv2_x in Figure 4 ), a second CBAM attention mechanism layer, a second Bottleneck block (corresponding to Figure 4 the 1×1 Convolution in
[0042] In another exemplary embodiment of the present application, the input image format that the improved ResNet model can identify is only 224*224. Compressing the original image to 224*224 will inevitably cause detail distortion, and cropping 1 image from the original image will waste a lot of image information. In order to capture more image details, the original image can be regularly cropped into multiple high-definition detail images of 224×224, and then input into the improved ResNet model in sequence to obtain multiple groups of physical and chemical properties of coal gangue bio-modified soil, and then integrated into 1 group of prediction results by a random forest model, so that the information in the original image is fully utilized, greatly improving the accuracy and reliability of the prediction. Therefore, as an optional implementation manner, in step 102, inputting the coal gangue bio-modified soil image to be predicted into the improved ResNet model to predict the physical and chemical properties of the coal gangue bio-modified soil specifically includes:
[0043] (1) Crop the image to be predicted of the biologically modified gangue soil to obtain several sub-images.
[0044] (2) Input each sub-image into the improved ResNet model respectively to predict the physical and chemical properties of the biologically modified gangue soil corresponding to each sub-image.
[0045] (3) Input the physical and chemical properties of the biologically modified gangue soil corresponding to each sub-image into the random forest model for integration to predict the physical and chemical properties of the biologically modified gangue soil.
[0046] In another exemplary embodiment of the present application, when using the improved ResNet model to predict the image to be predicted of the biologically modified gangue soil, the improved ResNet model needs to be trained first. The specific training process is as follows.
[0047] (S1): Collect image sample data. Extract samples of the biologically modified gangue soil at different times (such as 1 day, 5 days, 10 days, 20 days, 30 days, 40 days), and take photos of the samples with a camera. To increase the richness of the data set, set multiple shooting parameters: the distance between the camera and the sample (4 cm, 10 cm, 20 cm, 40 cm), the exposure parameter (0, -2, -4), and the sensitivity (50, 250, 800, 4000).
[0048] (S2): Detect the physical and chemical properties of the biologically modified gangue soil. According to the method specified in "Grade of Cultivated Land Quality GBT33469-2016", detect the soil mechanical composition, soil bulk density, organic matter content, sulfate ion concentration, electrical conductivity, and pH of the sample; to judge the harmless compost maturity of the sample, according to the method specified in "Organic Fertilizer NY525-2021", detect the seed germination rate of the sample; to judge whether there is carbon residue in the sample, according to the method specified in "Industrial Analysis Method of Coal GB / T 212-2008", detect the fixed carbon content and calorific value of the sample.
[0049] (S3): Construct a training data set. Label the image data set of the biologically modified gangue soil, and each sample image corresponds to a set of detected values of the physical and chemical properties of the biologically modified gangue soil. Taking images Img_0001, Img_0002, and Img_2171 as examples, as shown in Table 1.
[0050] Table 1 Example of Detected Values of Physical and Chemical Properties in the Training Set
[0051]
[0052] (S4): Train the improved ResNet model.
[0053] Based on the coal gangue biologically modified soil image dataset in step (S3), train the improved ResNet model, including the following steps:
[0054] 1) Divide the dataset: Divide the coal gangue biologically modified soil image dataset into a training set and a validation set in a ratio of 7:3. The training set includes 1,520 images, and the validation set includes 651 images.
[0055] 2) Image enhancement: The coal gangue biologically modified soil image dataset is augmented through data enhancement means such as random rotation, random mirroring, random warping, and random adjustment of brightness, contrast, and saturation, as Figure 5 shown.
[0056] 3) Hyperparameter setting: Set the batch size of the improved ResNet model to 4, the number of training times to 300, and the learning rate to 0.0001.
[0057] 4) Train the improved ResNet model: Calculate the loss of the validation set for each training. After 300 trainings, save the parameter weights with the lowest validation set loss.
[0058] (S5): Couple the improved ResNet model with the random forest model.
[0059] (1) Image preprocessing. Crop each captured image into 9 detailed high-definition images of 224×224. The rule for cropping the original image is:
[0060] 1) Crop the upper left, lower left, upper right, and lower right regions of the original image to obtain 4 detailed images of 224×224, numbered T 1.1 、T 1.2 、T 1.3 、T 1.4 respectively.
[0061] 2) Crop the center of the original image to obtain 1 image of 800×800.
[0062] 3) Crop the middle, upper left, lower left, upper right, and lower right of the 800×800 image to obtain 5 images of 224×224 respectively, numbered T 2.1 、T 2.2 、T 23 、T 2.4 、T 2.5 respectively, for a total of 9 sub-images.
[0063] (2) Construct the random forest model dataset. Input the 9 sub-images into the improved ResNet model in sequence, and output 9 groups of prediction results P1 - P9. P1 - P9 are used as input data, and the physical and chemical property detection values corresponding to the original image are used as output data.
[0064] 3. Train a random forest model. Based on the random forest model dataset, train a random forest model that can integrate 9 groups of prediction results into 1 group of final prediction results.
[0065] (S6) Evaluate the prediction effect.
[0066] (1) Use a test set containing 100 images to test the performance of the "ResNet-CBAM / random forest" coupled model. The test set contains images TestImg1 - TestImg100. Taking TestImg1 - TestImg6 as an example, as Figure 6 shown.
[0067] (2) Input the test set into the "ResNet-CBAM / random forest" coupled model, and the output prediction results include: soil mechanical composition, soil bulk density, organic matter content, sulfate ion concentration, conductivity, pH, seed germination rate, fixed carbon content, calorific value.
[0068] (3) The prediction results of the physical and chemical properties of the coal gangue bio-modified soil for TestImg1 - TestImg6 and the corresponding actual detection results are shown in Table 2.
[0069] Table 2 Example of prediction results and detection results of the test set
[0070]
[0071]
[0072] (4) Evaluate the prediction effect. Calculate R2, RRMSE, and MAPE using all the predicted values and detected values obtained from the test set prediction to evaluate the model performance. The calculation results can be referred to in Table 3.
[0073] Among them, the formula for R2 (coefficient of determination) is:
[0074]
[0075] The formula for RRMSE (relative root mean square error) is:
[0076]
[0077] The formula for MAPE (mean absolute percentage error) is:
[0078]
[0079] In the formula, n is the number of images, f(x i ) is the model predicted value, y i is the actual detected value, is the mean of the actual detected values.
[0080] Table 3 Performance Evaluation Parameters of "ResNet-CBAM / Random Forest" Coupled Model
[0081]
[0082] R2 (coefficient of determination) reflects the fitting effect of the model. Generally, when it reaches above 0.8, the fitting effect of the model is considered good. As can be seen from Table 3, the R2 of all parameters is greater than 0.8, and the fitting effect is good.
[0083] MAPE (mean absolute percentage error) reflects the accuracy of model prediction. A value lower than 10% is considered to have a high prediction accuracy and small error, and 10% to 20% is considered an acceptable accuracy. As can be seen from Table 3, the average error of model prediction is within the acceptable range.
[0084] RRMSE (relative root mean square error) is similar to MAPE, but is more sensitive to extreme values. A value lower than 10% is considered to have a high prediction accuracy and small error, and 10% to 20% is considered an acceptable accuracy. As can be seen from Table 3, only the RRMSE of soil bulk density is greater than 20%, but the MAPE is only 12.41%, indicating that the average error of the model is small when predicting soil bulk density, but the error is large when predicting extreme values of soil bulk density, and the prediction effect is not ideal.
[0085] Based on comprehensive judgment, for the trained "ResNet-CBAM / Random Forest" coupled model, except for the unsatisfactory prediction effect of extreme values of soil bulk density, the fitting effect is good, the accuracy is high, and the error is small when predicting other physical and chemical properties. It can accurately predict the physical and chemical properties of coal gangue bio-modified soil, greatly reducing the manpower, material resources, financial resources and time consumed by experimental detection during the production process.
[0086] In this application, an alternative implementation discloses a rapid prediction method for the physical and chemical properties of coal gangue bio-modified soil based on the "ResNet-CBAM / Random Forest" coupling model. Using the sample images in the process of coal gangue soilification as input features, and leveraging the efficient learning ability of deep learning neural networks, it realizes the rapid and accurate prediction of the physical and chemical properties of coal gangue bio-modified soil. The ResNet-CBAM prediction model in the "ResNet-CBAM / Random Forest" coupling model is integrated by combining the powerful feature extraction and understanding ability of the ResNet network and the ability of the CBAM attention mechanism to adaptively increase the weights of important features. The integrated ResNet-CBAM prediction model shows significant advantages compared with the ResNet model. The CBAM module effectively enhances the model's attention to important features through channel and spatial attention mechanisms, can more accurately capture the detailed features of the target object, and improves the accuracy and robustness of the prediction. To capture more image details, the original image is regularly cropped into multiple high-definition detail images of 224×224, which are sequentially input into the ResNet-CBAM prediction model to obtain multiple groups of physical and chemical properties of coal gangue bio-modified soil, and then integrated into one group of prediction results by the random forest model.
[0087] The "ResNet-CBAM / Random Forest" coupling model obtained through training can comprehensively identify the detailed features in the images of coal gangue bio-modified soil, realize rapid and accurate prediction and preliminary judgment of physical and chemical properties, reduce the cost of repeated detection, and has the advantages of simple operation and rapid recognition, etc., and is easy to promote and apply.
[0088] This application also provides an application scenario that applies the above-mentioned prediction method for the physical and chemical properties of coal gangue bio-modified soil. Specifically: The prediction method for the physical and chemical properties of coal gangue bio-modified soil provided in this embodiment can be applied in the scenario of resource recovery and utilization evaluation. This scenario includes a data collection link, a physical and chemical property prediction link, and a resource recovery and utilization evaluation link; the data collection link is used to collect the images to be predicted of coal gangue bio-modified soilification; the physical and chemical property prediction link is used to predict the physical and chemical properties of coal gangue bio-modified soil based on the collected images to be predicted of coal gangue bio-modified soilification; the resource recovery and utilization evaluation link is used to conduct resource recovery and utilization evaluation based on the prediction of the physical and chemical properties of coal gangue bio-modified soil. The prediction method for the physical and chemical properties of coal gangue bio-modified soil provided in this embodiment belongs to the physical and chemical property prediction link.
[0089] Based on the same inventive concept, an embodiment of the present application further provides a device for predicting the physical and chemical properties of coal gangue biologically modified soil for implementing the method for predicting the physical and chemical properties of coal gangue biologically modified soil involved above. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the device for predicting the physical and chemical properties of coal gangue biologically modified soil provided below can refer to the limitations on the method for predicting the physical and chemical properties of coal gangue biologically modified soil in the above text, and will not be repeated here.
[0090] In an exemplary embodiment, as Figure 7 shown, a device for predicting the physical and chemical properties of coal gangue biologically modified soil is provided, including:
[0091] An image acquisition module M1, configured to acquire an image to be predicted of coal gangue biologically modified soil.
[0092] A prediction module M2, configured to input the image to be predicted of coal gangue biologically modified soil into an improved ResNet model to predict the physical and chemical properties of coal gangue biologically modified soil; the improved ResNet model is a model obtained by improving the ResNet101 model; the improvement includes performing downsampling operations using the 3×3 convolutional layer of the Bottleneck block in the ResNet101 model and connecting a CBAM attention mechanism layer between every two adjacent Bottleneck blocks after the max pooling layer of the ResNet101 model.
[0093] For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method part.
[0094] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 8As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the prediction data of the physical and chemical properties of coal gangue bio-modified soil. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for predicting the physical and chemical properties of coal gangue bio-modified soil.
[0095] Those skilled in the art can understand that Figure 8 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.
[0096] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.
[0097] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.
[0098] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.
[0099] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAMs), magnetoresistive random access memories (MRAMs), ferroelectric random access memories (FRAMs), phase change memories (PCMs), graphene memories, etc. Volatile memories can include random access memories (RAMs) or external caches, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.
[0100] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.
[0101] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.
[0102] Specific examples are used in this article to elaborate on the principles and implementation manners of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.
Claims
1. A method for predicting the physical and chemical properties of coal gangue bio-modified soil, characterized in that, The method for predicting the physical and chemical properties of coal gangue biologically modified soil includes: Obtaining an image to be predicted of coal gangue biologically modified soil; Inputting the image to be predicted of coal gangue biologically modified soil into an improved ResNet model to predict the physical and chemical properties of coal gangue biologically modified soil; the improved ResNet model is a model obtained by improving the ResNet101 model; the improvement includes performing downsampling operations using the 3×3 convolutional layer of the Bottleneck block in the ResNet101 model and connecting a CBAM attention mechanism layer between every two adjacent Bottleneck blocks after the maximum pooling layer of the ResNet101 model.
2. The method for predicting the physical and chemical properties of coal gangue biologically modified soil according to claim 1, wherein The improvement also includes: modifying the AdaptiveAvgPool2d layer and the fully connected layer in the ResNet101 model to an AvgPool2d layer and a convolutional layer respectively.
3. The method for predicting the physical and chemical properties of coal gangue biologically modified soil according to claim 2, wherein The improvement also includes: replacing the softmax output layer for completing the classification task in the ResNet101 model with an MSE output layer.
4. The method for predicting the physical and chemical properties of coal gangue bio-modified soil according to claim 3, wherein, The improved ResNet model includes: an input layer, an initial convolutional layer, a maximum pooling layer, a first CBAM attention mechanism layer, a first Bottleneck block, a second CBAM attention mechanism layer, a second Bottleneck block, a third CBAM attention mechanism layer, a third Bottleneck block, a fourth CBAM attention mechanism layer, a fourth Bottleneck block, an AvgPool2d layer, a replacement convolutional layer, and an MSE output layer connected in sequence; the 3×3 convolutional layer in the first to fourth Bottleneck blocks performs downsampling operations; the replacement convolutional layer is a 1×1 convolutional layer; the 1×1 convolutional layer is a convolutional layer that replaces the fully connected layer of the ResNet101 model.
5. The method for predicting the physical and chemical properties of coal gangue biologically modified soil according to claim 1 or 4, characterized in that, Inputting the image to be predicted of coal gangue biologically modified soil into the improved ResNet model to predict the physical and chemical properties of coal gangue biologically modified soil specifically includes: Cropping the image to be predicted of coal gangue biologically modified soil to obtain several sub-images; Inputting each sub-image into the improved ResNet model respectively to predict the physical and chemical properties of coal gangue biologically modified soil corresponding to each sub-image; Inputting the physical and chemical properties of coal gangue biologically modified soil corresponding to each sub-image into a random forest model for integration to predict the physical and chemical properties of the coal gangue biologically modified soil.
6. The method for predicting the physical and chemical properties of coal gangue biologically modified soil according to claim 1, characterized in that, The physical and chemical properties of the coal gangue biologically modified soil include: soil mechanical composition, soil bulk density, organic matter content, sulfate ion concentration, electrical conductivity, pH, seed germination rate, fixed carbon content, and calorific value.
7. A device for predicting the physical and chemical properties of coal gangue bio-modified soil, characterized in that, The device for predicting the physical and chemical properties of coal gangue biologically modified soil includes: An image acquisition module for obtaining an image to be predicted of coal gangue biologically modified soil; A prediction module, configured to input the to-be-predicted image of the coal gangue biologically modified soil into an improved ResNet model to predict the physical and chemical properties of the coal gangue biologically modified soil; the improved ResNet model is a model obtained by improving the ResNet101 model; the improvement includes performing downsampling operations using the 3×3 convolutional layer of the Bottleneck block in the ResNet101 model and connecting a CBAM attention mechanism layer between every two adjacent Bottleneck blocks after the maximum pooling layer of the ResNet101 model.
8. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the method for predicting the physical and chemical properties of the coal gangue biologically modified soil according to any one of claims 1-6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for predicting the physical and chemical properties of the coal gangue biologically modified soil according to any one of claims 1-6.
10. A computer program product comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for predicting the physical and chemical properties of the coal gangue biologically modified soil according to any one of claims 1-6.