Alteration zone detection and analysis method and system

Abstract

The invention discloses an alteration zone detection and analysis method and system, and belongs to the field of unfavorable geological precursor feature recognition, and the method comprises the steps: obtaining physical and chemical process data in a geological environment through numerical simulation, and generating a spatial distribution image of an alteration zone; preprocessing the spatial distribution image to obtain a preprocessed image; alteration zone features are extracted from the preprocessed image; based on a supervised learning mode, training a convolutional neural network through alteration band features to obtain an alteration band detection model; and identifying the actual alteration band data through the alteration band detection model to obtain a detection result. The position, range and characteristics of the alteration zone can be accurately detected, and reliable technical support is provided for the fields of geological exploration, tunnel engineering, environment monitoring and the like.

Description

Technical Field

[0001] The present invention belongs to the technical field of adverse geological precursor feature recognition, and in particular relates to an alteration zone detection and analysis method and system. Background Art

[0002] With the development of computer technology, numerical simulation has become a powerful tool for studying geological processes. Through numerical simulation, the physical and chemical processes in geological bodies can be reproduced on computers to generate spatial distribution images of alteration zones. These images provide rich geological information, but how to effectively extract and analyze this information to achieve automated and high-precision alteration zone identification remains a challenge.

[0003] In recent years, machine learning and deep learning technologies have made significant progress in the fields of image recognition and pattern recognition. In particular, convolutional neural networks (CNNs) have been widely used in various image recognition tasks due to their high efficiency in image processing and powerful feature extraction capabilities. However, there is still a lack of research on the application of convolutional neural networks in the identification and analysis of geological alteration zones, especially in combination with numerical simulation data. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes an alteration zone detection and analysis method and system to solve the problems existing in the above prior art.

[0005] To achieve the above object, the present invention provides a method for detecting and analyzing alteration zones, comprising the following steps: Through numerical simulation, we can obtain data on physical and chemical processes in the geological environment and generate spatial distribution images of alteration zones; Preprocessing the spatial distribution image to obtain a preprocessed image; extracting alteration zone features from the preprocessed image; Based on the supervised learning method, a convolutional neural network is trained through the alteration zone characteristics to obtain an alteration zone detection model; The actual alteration zone data is identified through the alteration zone detection model to obtain the detection result.

[0006] Preferably, preprocessing the spatial distribution image includes: Performing denoising processing on the spatial distribution image by Gaussian filtering or median filtering to obtain a denoised image; The contrast of the denoised image is enhanced by histogram equalization or contrast limiting adaptive method to obtain an enhanced image; The enhanced image is normalized to scale the pixel values ​​to a range from 0 to 1.

[0007] Preferably, extracting alteration zone characteristics comprises: The texture features of the alteration zone are extracted through the gray-level co-occurrence matrix; the shape features of the alteration zone are extracted according to the geometric shape descriptor of the image; the edge features of the alteration zone are extracted using the Canny edge detection algorithm; Principal component analysis was used to reduce the dimension of the alteration zone characteristics.

[0008] Preferably, training a convolutional neural network using the alteration zone characteristics comprises: A convolutional neural network model is constructed using a deep learning framework, wherein the convolutional neural network model includes multiple convolutional layers, activation layers, pooling layers, and fully connected layers; wherein the deep learning framework includes TensorFlow and PyTorch; The extracted alteration zone features are input into the network as training data, and the network is trained using the cross entropy loss function and Adam optimizer.

[0009] Preferably, the method further comprises: Obtain misclassified samples in the detection results, retrain the alteration zone detection model through the misclassified samples, and update the alteration zone detection model.

[0010] The present invention also provides an alteration zone detection and analysis system, comprising: Numerical simulation module, used to obtain physical and chemical process data in geological environment through numerical simulation and generate spatial distribution images of alteration zones; A preprocessing module, used for preprocessing the spatial distribution image to obtain a preprocessed image; A feature extraction module, used for extracting alteration zone features from the preprocessed image; A model building module is used to train a convolutional neural network through the alteration zone characteristics based on a supervised learning method to obtain an alteration zone detection model; An identification module, used to identify actual alteration zone data through the alteration zone detection model to obtain a detection result; The optimization and updating module is used to obtain misclassified samples in the detection results, retrain the alteration zone detection model through the misclassified samples, and update the alteration zone detection model.

[0011] Preferably, the preprocessing module comprises: A denoising unit, configured to perform denoising processing on the spatial distribution image by Gaussian filtering or median filtering to obtain a denoised image; An image enhancement unit, used for enhancing the contrast of the denoised image by a histogram equalization or a contrast limiting adaptive method to obtain an enhanced image; The normalization unit is used to perform normalization processing on the enhanced image and scale the pixel values ​​to a range from 0 to 1.

[0012] Preferably, the feature extraction module comprises: The extraction unit is used to extract the texture features of the alteration zone through the gray level co-occurrence matrix; extract the shape features of the alteration zone according to the geometric shape descriptor of the image; and extract the edge features of the alteration zone using the Canny edge detection algorithm; A dimension reduction unit is used to reduce the dimension of the alteration zone characteristics using principal component analysis.

[0013] Preferably, the model building module comprises: A construction unit is used to construct a convolutional neural network model using a deep learning framework, wherein the convolutional neural network model includes multiple convolutional layers, activation layers, pooling layers, and fully connected layers; wherein the deep learning framework includes TensorFlow and PyTorch; The training unit is used to input the extracted alteration zone features into the network as training data, and use the cross entropy loss function and Adam optimizer to train the network.

[0014] The present invention also provides a computer program product, comprising a computer-readable storage medium and a computer program code for executing the method; wherein the program code comprises: an interface code of a numerical simulation software for generating a spatial distribution image of an alteration zone; an implementation code of an image preprocessing algorithm; an implementation code of a feature extraction algorithm; a construction and training code of a convolutional neural network model; and an implementation code of a deep learning iterative optimization algorithm.

[0015] Compared with the prior art, the present invention has the following advantages and technical effects: The present invention discloses a method and system for detecting and analyzing alteration zones, including: obtaining physical and chemical process data in a geological environment through numerical simulation to generate a spatial distribution image of the alteration zone; preprocessing the spatial distribution image to obtain a preprocessed image; extracting alteration zone features from the preprocessed image; training a convolutional neural network through alteration zone features based on a supervised learning method to obtain an alteration zone detection model; identifying actual alteration zone data through the alteration zone detection model to obtain a detection result. The present invention can accurately detect the location, range and features of the alteration zone, and provide reliable technical support for the fields of geological exploration, tunnel engineering and environmental monitoring.

[0016] The present invention significantly reduces the need for field sampling and laboratory analysis through automated numerical simulation and deep learning technology. The powerful feature extraction and recognition capabilities of convolutional neural networks improve the accuracy of alteration zone identification. It reduces the manpower and material consumption in traditional methods and reduces the overall cost. It can also quickly respond to real-time data in geological exploration to provide support for decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 This is a flow chart of a method for detecting and analyzing alteration zones according to an embodiment of the present invention; Figure 2 This is a neural network topology diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0018] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0019] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here. Embodiment 1:

[0020] like Figure 1-2 As shown, this embodiment provides an alteration zone detection and analysis method, including the following steps: Through numerical simulation, we can obtain data on physical and chemical processes in the geological environment and generate spatial distribution images of alteration zones; Preprocessing the spatial distribution image to obtain a preprocessed image; Extract alteration zone features from preprocessed images; Based on supervised learning, the convolutional neural network is trained through the alteration zone characteristics to obtain the alteration zone detection model; The actual alteration zone data is identified through the alteration zone detection model to obtain the detection results.

[0021] The specific steps include: Step 1: Generate alteration zone image by numerical simulation; Use commercial numerical simulation software (such as COMSOL Multiphysics or ANSYS Fluent) to simulate the physical and chemical processes in the geological environment. Simulation parameters include temperature, pressure, fluid flow, and chemical reaction rates.

[0022] Set the boundary conditions and initial conditions of the simulation area to reflect the actual geological environment.

[0023] The simulation software is run to generate spatial distribution images of alteration zones, which are presented in the form of high-resolution two-dimensional or three-dimensional data.

[0024] Step 2: Image preprocessing; Import the images generated by numerical simulation into image processing software (such as MATLAB or Python's OpenCV library).

[0025] De-noising the image, using algorithms such as Gaussian filtering or median filtering to reduce random noise in the image.

[0026] Enhance image contrast using histogram equalization or contrast limited adaptive method (CLAHE).

[0027] The image is normalized and the pixel values ​​are scaled to the range of 0 to 1 to facilitate subsequent feature extraction and neural network training.

[0028] Step 3: Feature extraction; The features of the alteration zone are extracted from the preprocessed image, including texture features (such as gray-level co-occurrence matrix), shape features (such as geometric shape descriptors of the image) and edge features (using the Canny edge detection algorithm).

[0029] First, the extraction of texture features is achieved through the gray-level co-occurrence matrix (GLCM). The gray-level co-occurrence matrix is ​​a statistical method that analyzes the gray-level value relationship between pixels in an image. In practice, we first convert the color image into a gray-level image, which can reduce the computational complexity while retaining enough information for texture analysis. Next, we calculate the gray-level co-occurrence matrix of the image at different directions (such as 0°, 45°, 90°, and 135°) and different distances (usually 1 pixel). These matrices contain the gray-level values ​​of pixel pairs and the spatial relationship between them, so that the texture characteristics of the image can be quantified. From these matrices, we can extract a variety of texture features, including energy (indicating the uniformity of the image), contrast (indicating the difference in gray-level values ​​in the image), homogeneity (indicating the degree of clustering of similar gray-level values ​​in the image), and correlation (indicating the degree of linear dependence between pixel pairs).

[0030] Secondly, the extraction of shape features is achieved through the geometric shape descriptors of the image. The purpose of this step is to identify and describe the geometric properties of the alteration zone, such as area, perimeter, centroid, and major axis length. In order to extract these features, we first need to convert the image into a binary image, which can clearly distinguish the alteration zone from the background. After binarization, we can use the regional attribute analysis tool in the image processing software to calculate the geometric features of each connected area. These features are crucial to understanding the spatial distribution and morphological characteristics of the alteration zone.

[0031] Finally, the edge feature extraction is achieved through the Canny edge detection algorithm. The Canny algorithm is a popular edge detection method that identifies edges in images through multi-stage noise reduction, gradient calculation, non-maximum suppression, and hysteresis threshold processing. After applying the Canny algorithm to the grayscale image, we can obtain edge images of the alteration zone, which clearly show the contours of the alteration zone. Edge features are crucial for subsequent tasks such as image segmentation, target recognition, and shape analysis.

[0032] Use principal component analysis (PCA) or other dimensionality reduction techniques to reduce feature dimensions to improve data processing efficiency.

[0033] Step 4: Build and train a convolutional neural network; Use a deep learning framework (such as TensorFlow or PyTorch) to build a convolutional neural network model, which includes multiple convolutional layers, activation layers (such as ReLU), pooling layers, and fully connected layers.

[0034] The extracted features are input into the network as training data, and the network is trained using the cross entropy loss function and Adam optimizer.

[0035] Through multiple iterations of training, the network achieves satisfactory accuracy on the validation set.

[0036] As a practical measurement method, the alteration feature detection equipment includes an image collector, an X-ray fluorescence spectrometer and an X-ray diffraction analyzer; an infrared temperature sensor; The image collector is used to collect images of target surrounding rocks; X-ray fluorescence spectrometer is mainly used to determine the composition and content of elements in the alteration zone; X-ray diffractometers are mainly used to analyze information such as the crystal structure of minerals in alteration zones; Infrared temperature sensors are mainly used to detect the temperature of target surrounding rocks; Collect samples of rock and process them.

[0037] The X-ray fluorescence spectrum measured by the X-ray fluorescence spectrometer is transmitted to the computer and processed using the corresponding analysis software. Based on the energy of the characteristic X-ray fluorescence, the type of elements present in the sample can be determined (each element has its own specific fluorescence energy). By analyzing the intensity of the fluorescence, the content of each element can be calculated. It is usually necessary to use a standard sample to establish a calibration curve, and then determine the content of the element in the unknown sample based on the calibration curve.

[0038] The X-ray diffraction patterns obtained by X-ray diffractometer detection are transferred to the computer and processed and analyzed using professional X-ray diffraction analysis software. They are compared with known mineral standard diffraction patterns to further confirm the type and structure of the mineral.

[0039] Aim the infrared temperature sensor at the rock part that needs to be tested to detect the temperature of the rock in real time.

[0040] Step 5: Acquisition and use of actual measured data; Using an image collector to collect images of target surrounding rocks; The elemental composition of the samples was analyzed using an X-ray fluorescence spectrometer, and the mineralogical composition of the samples was analyzed using an X-ray diffractometer (XRD). These data are combined with the image features generated by numerical simulation and used as the input of the convolutional neural network to identify and classify alteration zones.

[0041] Step 6: Deep learning iterative optimization; The actual measured alteration zone data is input into the trained convolutional neural network to obtain preliminary recognition results.

[0042] Analyze the recognition results to identify areas where the network performance is insufficient, such as misclassified samples.

[0043] These samples are used as new training data to retrain the network to improve recognition accuracy.

[0044] Step 7: Result analysis and application; Use the trained convolutional neural network to identify alteration zones on new geological data.

[0045] Analyze network output to determine the location, size, and likely mineral composition of alteration zones.

[0046] The analysis results are applied to geological exploration and mineral resource assessment to guide further exploration activities. Embodiment 2:

[0047] This embodiment also provides an alteration zone detection and analysis system, including: Numerical simulation module, used to obtain physical and chemical process data in geological environment through numerical simulation and generate spatial distribution images of alteration zones; A preprocessing module, used for preprocessing the spatial distribution image to obtain a preprocessed image; A feature extraction module is used to extract alteration zone features from preprocessed images; A model building module is used to train a convolutional neural network based on alteration zone characteristics based on supervised learning to obtain an alteration zone detection model; An identification module is used to identify actual alteration zone data through an alteration zone detection model to obtain detection results; The optimization and updating module is used to obtain misclassified samples in the detection results, retrain the alteration zone detection model through the misclassified samples, and update the alteration zone detection model.

[0048] Furthermore, the preprocessing module includes: A denoising unit, used for denoising the spatial distribution image by Gaussian filtering or median filtering to obtain a denoised image; An image enhancement unit, used for enhancing the contrast of the denoised image by a histogram equalization or a contrast limiting adaptive method to obtain an enhanced image; The normalization unit is used to normalize the enhanced image and scale the pixel values ​​to the range of 0 to 1.

[0049] Furthermore, the feature extraction module includes: The extraction unit is used to extract the texture features of the alteration zone through the gray level co-occurrence matrix; extract the shape features of the alteration zone according to the geometric shape descriptor of the image; and extract the edge features of the alteration zone using the Canny edge detection algorithm; Dimensionality reduction unit, used to reduce the dimensionality of alteration zone features using principal component analysis.

[0050] Furthermore, the model building module includes: A construction unit for building a convolutional neural network model using a deep learning framework, which includes multiple convolutional layers, activation layers, pooling layers, and fully connected layers; deep learning frameworks include TensorFlow and PyTorch; The training unit is used to input the extracted alteration zone features into the network as training data, and use the cross entropy loss function and Adam optimizer to train the network. Embodiment three:

[0051] This embodiment also provides a computer program product, including a computer-readable storage medium and a computer program code for executing the method; wherein the program code includes: an interface code of a numerical simulation software for generating a spatial distribution image of an alteration zone; an implementation code of an image preprocessing algorithm; an implementation code of a feature extraction algorithm; a construction and training code of a convolutional neural network model; and an implementation code of a deep learning iterative optimization algorithm.

[0052] The present invention proposes a method and system for detecting and analyzing alteration zones. The method first uses numerical simulation technology to simulate the characteristics of alteration zones, including key information such as the distribution of altered minerals, alteration intensity and spatial structure. Then, the simulated features are input into a convolutional neural network for preliminary learning through feature extraction to establish a mapping relationship between the characteristics of the alteration zone and geological phenomena, and the network training is completed to form an initial database. Next, the alteration zone data obtained by actual detection using the instrument is input into the network again for deep learning to continuously optimize the recognition and analysis capabilities of the model. The method can accurately detect the location, range and characteristics of the alteration zone, and significantly improve the accuracy and efficiency of the analysis. In addition, the present invention also constructs an alteration zone detection and analysis system integrating numerical simulation and CNN, which has a user-friendly interface and can provide reliable technical support for geological exploration, tunnel engineering and environmental monitoring. Through continuous data collection and model training, the system can adapt to different geological conditions, improve the accuracy of detection, and ensure data security and privacy protection.

[0053] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for detecting and analyzing alteration zones, characterized in that: The following steps are involved: Through numerical simulation, we can obtain data on physical and chemical processes in the geological environment and generate spatial distribution images of alteration zones; Preprocessing the spatial distribution image to obtain a preprocessed image; extracting alteration zone features from the preprocessed image; Based on the supervised learning method, a convolutional neural network is trained through the alteration zone characteristics to obtain an alteration zone detection model; The actual alteration zone data is identified through the alteration zone detection model to obtain the detection result.

2. The method according to claim 1, characterized in that: Preprocessing the spatial distribution image includes: Performing denoising processing on the spatial distribution image by Gaussian filtering or median filtering to obtain a denoised image; The contrast of the denoised image is enhanced by histogram equalization or contrast limiting adaptive method to obtain an enhanced image; The enhanced image is normalized to scale the pixel values ​​to a range from 0 to 1.

3. The method according to claim 1, characterized in that: The extraction of alteration zone characteristics includes: The texture features of the alteration zone are extracted through the gray-level co-occurrence matrix; the shape features of the alteration zone are extracted according to the geometric shape descriptor of the image; the edge features of the alteration zone are extracted using the Canny edge detection algorithm; Principal component analysis was used to reduce the dimension of the alteration zone characteristics.

4. The method according to claim 1, characterized in that Training the convolutional neural network using the alteration zone characteristics includes: A convolutional neural network model is constructed using a deep learning framework, wherein the convolutional neural network model includes multiple convolutional layers, activation layers, pooling layers, and fully connected layers; wherein the deep learning framework includes TensorFlow and PyTorch; The extracted alteration zone features are input into the network as training data, and the network is trained using the cross entropy loss function and Adam optimizer.

5. The method according to claim 1, characterized in that The method further comprises: Obtain misclassified samples in the detection results, retrain the alteration zone detection model through the misclassified samples, and update the alteration zone detection model.

6. An alteration zone detection and analysis system, characterized in that: include: Numerical simulation module, used to obtain physical and chemical process data in geological environment through numerical simulation and generate spatial distribution images of alteration zones; A preprocessing module, used for preprocessing the spatial distribution image to obtain a preprocessed image; A feature extraction module, used for extracting alteration zone features from the preprocessed image; A model building module is used to train a convolutional neural network through the alteration zone characteristics based on a supervised learning method to obtain an alteration zone detection model; An identification module, used to identify actual alteration zone data through the alteration zone detection model to obtain a detection result; The optimization and updating module is used to obtain misclassified samples in the detection results, retrain the alteration zone detection model through the misclassified samples, and update the alteration zone detection model.

7. The system according to claim 6, characterized in that The preprocessing module comprises: A denoising unit, configured to perform denoising processing on the spatial distribution image by Gaussian filtering or median filtering to obtain a denoised image; An image enhancement unit, used for enhancing the contrast of the denoised image by a histogram equalization or a contrast limiting adaptive method to obtain an enhanced image; The normalization unit is used to perform normalization processing on the enhanced image and scale the pixel values ​​to a range from 0 to 1.

8. The system according to claim 6, characterized in that The feature extraction module comprises: The extraction unit is used to extract the texture features of the alteration zone through the gray level co-occurrence matrix; extract the shape features of the alteration zone according to the geometric shape descriptor of the image; and extract the edge features of the alteration zone using the Canny edge detection algorithm; A dimension reduction unit is used to reduce the dimension of the alteration zone characteristics using principal component analysis.

9. The system according to claim 6, characterized in that The model building module includes: A construction unit is used to construct a convolutional neural network model using a deep learning framework, wherein the convolutional neural network model includes multiple convolutional layers, activation layers, pooling layers, and fully connected layers; wherein the deep learning framework includes TensorFlow and PyTorch; The training unit is used to input the extracted alteration zone features into the network as training data, and use the cross entropy loss function and Adam optimizer to train the network.

10. A computer program product, characterized in that It comprises a computer-readable storage medium and a computer program code for executing the method as claimed in any one of claims 1 to 5; wherein the program code comprises: an interface code of a numerical simulation software for generating a spatial distribution image of an alteration zone; an implementation code of an image preprocessing algorithm; an implementation code of a feature extraction algorithm; a construction and training code of a convolutional neural network model; and an implementation code of a deep learning iterative optimization algorithm.

Images