Image recognition-based rock geology survey method, system, device and medium

Abstract

This invention discloses a method, system, equipment, and medium for rock and soil geological exploration based on image recognition, relating to the fields of geotechnical engineering and geological disaster prevention and control. The method includes image acquisition; image preprocessing; feature extraction; disease identification; and early warning output. Through high-precision image recognition technology, early warning of geological diseases is achieved, significantly improving the accuracy and efficiency of disease identification. Compared with traditional manual inspection methods, this invention avoids misjudgments caused by human factors, reducing the misjudgment rate. Simultaneously, through automated image acquisition and processing, it can quickly cover large areas prone to geological disasters, improving the efficiency of disease monitoring. Furthermore, by promptly generating early warning information and notifying management personnel, this invention helps to quickly take preventative measures, reducing disease losses and improving the continuity and reliability of disease monitoring.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering and geological disaster prevention technology, specifically to image recognition-based methods, systems, equipment, and media for geotechnical geological exploration. Background Technology

[0002] In the fields of geotechnical engineering and geological disaster prevention, accurate investigation of the structure and defects of soil and rock layers is a crucial foundation for ensuring engineering safety and disaster prevention and mitigation. Traditional geological investigation mainly relies on methods such as manual on-site inspections, borehole sampling, and laboratory testing. While these methods can obtain some geological information, they suffer from low efficiency, long cycles, and limited coverage. In recent years, with the development of image acquisition and computer vision technologies, image-based automated investigation methods have gradually gained attention, providing a new technological approach for soil and rock geological monitoring.

[0003] However, existing image-based geological exploration methods still have significant shortcomings in practical applications, mainly manifested in low accuracy of hazard identification. Because traditional techniques still rely heavily on manual labor and lack efficient automated identification and early warning mechanisms, they are prone to misjudgment or omission due to human error, making it difficult to achieve rapid and continuous monitoring of large areas prone to geological disasters. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, equipment and medium for rock and soil geological exploration based on image recognition, so as to solve the problems that the traditional technology still relies heavily on manual labor, lacks an efficient automated identification and early warning mechanism, is prone to misjudgment or omission due to human judgment errors, and is difficult to achieve rapid and continuous monitoring of large areas of frequent geological disasters.

[0005] To achieve the above objectives, the present invention provides the following technical solution, comprising the following steps:

[0006] Image acquisition involves installing image acquisition equipment in the geotechnical survey area to periodically collect cross-sectional image information of the geotechnical layers.

[0007] Image preprocessing involves performing noise reduction and enhancement operations on the acquired images to improve image quality.

[0008] Feature extraction involves extracting feature parameters related to rock and soil geology from the preprocessed image. These feature parameters include color, texture, and shape.

[0009] Disease identification utilizes a pre-trained deep learning model to analyze feature parameters and identify the type and extent of geological diseases in images;

[0010] The system generates early warning information based on the identification results and notifies management personnel through various means.

[0011] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, the image acquisition device is a high-definition camera, and the installation position and angle of the camera are adjusted according to the location of the exploration area;

[0012] The image preprocessing operations include median filtering, Gaussian filtering for noise reduction, and histogram equalization for enhancement.

[0013] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, the feature extraction includes the extraction of color features, texture features and shape features, wherein color features are obtained by calculating the statistics of different color channels in the image, texture features are obtained by the gray-level co-occurrence matrix method, and shape features are obtained by edge detection and contour extraction methods.

[0014] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, the image preprocessing includes: denoising the image using a first denoising method, namely a median filtering algorithm. The median filtering algorithm removes noise by replacing the value of each pixel in the image with the median value of its neighborhood. The specific steps are as follows:

[0015] Select a neighboring window, usually a 3×3 or 5×5 rectangular window;

[0016] Align the center of the window with each pixel in the image and extract the pixel values ​​within the window;

[0017] Sort the pixel values ​​within the window and take the median value as the new value for that pixel;

[0018] Repeat the above steps until the entire image has been processed.

[0019] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, the image is denoised using a second denoising method, namely, a Gaussian filtering algorithm. The Gaussian filtering algorithm smooths the image using a Gaussian kernel, thereby removing noise. The formula for Gaussian filtering is...

[0020]

[0021] Where σ is the standard deviation of the Gaussian kernel, which controls the smoothness of the filter; x and y represent the coordinate positions of pixels in the image, that is, x and y are the offsets relative to the center of the Gaussian kernel; and e is the base of the natural logarithm, which is generally taken as 2.718.

[0022] The specific steps are as follows: select a Gaussian kernel, usually a 3×3 or 5×5 matrix;

[0023] Calculate the weights of the Gaussian kernel, and calculate the weight of each position according to the Gaussian formula;

[0024] Align the center of the Gaussian kernel with each pixel in the image, and calculate the new value of that pixel as a weighted sum of the pixel values ​​within the Gaussian kernel.

[0025] Repeat the above steps until the entire image has been processed.

[0026] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, the image is enhanced by using a histogram equalization method. Histogram equalization makes the contrast and brightness of the image more uniform by adjusting the histogram distribution of the image.

[0027] The specific steps are as follows: calculate the histogram h(i) of the image, where i is the pixel value and h(i) is the number of times pixel value i appears;

[0028] Calculate the cumulative histogram c(i).

[0029]

[0030] Calculate the new pixel value s(i).

[0031]

[0032] Among them, c min It is the minimum value of the cumulative histogram, and n is the total number of pixels in the image;

[0033] The new pixel value s(i) is mapped back to the original pixel value range to obtain the enhanced image.

[0034] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, the feature extraction includes calculating statistical quantities such as the mean and variance of different color channels in the image as color features; the specific steps are as follows:

[0035] Calculate the mean μ of the red, green, and blue color channels respectively. R μ G μ B and variance

[0036]

[0037]

[0038] Among them, R i G i B i is the red, green, and blue channel values ​​of the i-th pixel, and n is the total number of pixels in the image;

[0039] Texture features are extracted using the gray-level co-occurrence matrix method;

[0040] The specific steps are as follows: calculate the gray-level co-occurrence matrix P(d,θ), where d is the distance between pixels and θ is the direction; the element P(i,j;d,θ) of the gray-level co-occurrence matrix represents the number of times a pixel with gray value i and a pixel with gray value j appear under the direction θ and distance d.

[0041] The statistical characteristics of the gray-level co-occurrence matrix, including contrast, correlation, energy, and homogeneity, are calculated using the following formulas:

[0042] Contrast Ratio:

[0043]

[0044] Correlation:

[0045]

[0046] Where, μ i and μ j These are the mean values ​​of grayscale values ​​i and j, respectively, and σ i and σ j These are the standard deviations of grayscale values ​​i and j, respectively.

[0047] As a preferred embodiment of the image recognition-based rock and soil geological exploration method of the present invention, shape features are obtained through edge detection and contour extraction methods;

[0048] The specific steps are as follows:

[0049] Use an edge detection algorithm to detect edges in an image;

[0050] Use a contour extraction algorithm to extract edge contours;

[0051] Calculate the geometric features of the contour, including area, perimeter, and shape factor.

[0052] As a preferred embodiment of the image recognition-based rock and soil geological exploration system of the present invention, it includes: an image acquisition module: used to install image acquisition equipment in the rock and soil exploration area and periodically acquire cross-sectional image information of the rock and soil layers;

[0053] Image preprocessing module: used to perform noise reduction and enhancement operations on the acquired images to improve image quality;

[0054] Feature extraction module: used to extract feature parameters related to rock and soil geology from the preprocessed image, including color, texture, and shape;

[0055] Disease identification module: Used to analyze feature parameters using a pre-trained deep learning model to identify the type and severity of geological diseases in images;

[0056] Early warning output module: used to generate early warning information based on the identification results and notify staff and managers via SMS and APP push;

[0057] The image acquisition device in the image acquisition module is a high-definition camera, and the installation position and angle of the camera are adjusted according to the location of the survey area.

[0058] As a preferred embodiment of the image recognition-based rock and soil geological exploration system of the present invention, the image preprocessing module includes denoising the image using a median filtering algorithm;

[0059] The image is denoised using a Gaussian filtering algorithm.

[0060] Image enhancement is achieved using histogram equalization.

[0061] The feature extraction module includes,

[0062] Calculate the mean and variance, and other statistical measures, of different color channels in the image as color features;

[0063] Texture features are extracted using the gray-level co-occurrence matrix method;

[0064] Shape features are obtained through edge detection and contour extraction methods;

[0065] The output methods of the early warning output module include SMS and APP push, and the early warning information includes disease type, location of occurrence, and severity.

[0066] The beneficial effects of this invention are as follows: Through high-precision image recognition technology, early warning of geological hazards is achieved, significantly improving the accuracy and efficiency of hazard identification; compared with traditional manual inspection methods, this invention avoids misjudgments caused by human factors, reducing the misjudgment rate; simultaneously, through automated image acquisition and processing, it can quickly cover large areas prone to geological disasters, improving the efficiency of hazard monitoring; furthermore, by generating early warning information and notifying management personnel in a timely manner, this invention helps to quickly take preventive measures, reducing hazard losses; and it improves the continuity and reliability of hazard monitoring. Attached Figure Description

[0067] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0068] Figure 1This is a schematic diagram of the overall process of the image recognition-based rock and soil geological exploration method of the present invention. Detailed Implementation

[0069] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0070] As attached Figure 1 As shown:

[0071] Example:

[0072] Reference Figure 1 As shown in the figure, the image recognition-based rock and soil geological exploration method includes the following steps: image acquisition, installing image acquisition equipment in the rock and soil exploration area, and periodically acquiring cross-sectional image information of the rock and soil layers. Here, the rock and soil exploration area includes natural outcrops, engineering excavation slopes or artificial trenches. Install image acquisition equipment to periodically acquire surface outcrops or cross-sectional exposure image information of the rock and soil layers.

[0073] Image preprocessing involves performing noise reduction and enhancement operations on the acquired images to improve image quality.

[0074] Feature extraction involves extracting feature parameters related to rock and soil geology from the preprocessed image. These feature parameters include color, texture, and shape.

[0075] Disease identification utilizes a pre-trained deep learning model to analyze feature parameters and identify the type and severity of geological diseases in images; and early warning output generates early warning information based on the identification results and notifies management personnel through various means.

[0076] Specifically, the image acquisition device is a high-definition camera, and the installation position and angle of the camera are adjusted according to the location of the survey area;

[0077] Image preprocessing operations include median filtering, Gaussian filtering for noise reduction, and histogram equalization for enhancement.

[0078] Furthermore, feature extraction includes the extraction of color features, texture features, and shape features. Color features are obtained by calculating the statistics of different color channels in the image, texture features are obtained using the gray-level co-occurrence matrix method, and shape features are obtained using methods such as edge detection and contour extraction.

[0079] Specifically, image preprocessing includes using a first denoising method, namely median filtering algorithm, to denoise the image. The median filtering algorithm removes noise by replacing the value of each pixel in the image with the median value of the neighborhood of that pixel.

[0080] The specific steps are as follows: select a neighborhood window, usually a 3×3 or 5×5 rectangular window;

[0081] Align the center of the window with each pixel in the image and extract the pixel values ​​within the window;

[0082] Sort the pixel values ​​within the window and take the median value as the new value for that pixel;

[0083] Repeat the above steps until the entire image has been processed.

[0084] Specifically, the second denoising method, Gaussian filtering, is used to denoise the image. Gaussian filtering smooths the image using a Gaussian kernel, thereby removing noise. The formula for Gaussian filtering is...

[0085]

[0086] Where σ is the standard deviation of the Gaussian kernel, controlling the smoothness of the filter; x and y represent the coordinates of pixels in the image, i.e., x and y are the offsets relative to the center of the Gaussian kernel; and e is the base of the natural logarithm, typically taken as 2.718. The specific steps are as follows:

[0087] Choose a Gaussian kernel, usually a 3×3 or 5×5 matrix;

[0088] Calculate the weights of the Gaussian kernel, and calculate the weight of each position according to the Gaussian formula;

[0089] Align the center of the Gaussian kernel with each pixel in the image, and calculate the new value of that pixel as a weighted sum of the pixel values ​​within the Gaussian kernel.

[0090] Repeat the above steps until the entire image has been processed.

[0091] Furthermore, the image is enhanced using histogram equalization. Histogram equalization adjusts the histogram distribution of the image to make the contrast and brightness of the image more uniform. The specific steps are as follows: calculate the histogram h(i) of the image, where i is the pixel value and h(i) is the number of times the pixel value i appears.

[0092] Calculate the cumulative histogram c(i).

[0093]

[0094] Calculate the new pixel value s(i).

[0095]

[0096] Among them, c min It is the minimum value of the cumulative histogram, and n is the total number of pixels in the image;

[0097] The new pixel value s(i) is mapped back to the original pixel value range to obtain the enhanced image.

[0098] Specifically, feature extraction includes calculating statistical measures such as the mean and variance of different color channels in the image as color features; the specific steps are as follows:

[0099] Calculate the mean μ of the red, green, and blue color channels respectively. R μ G μ B and variance

[0100]

[0101]

[0102] Among them, R i G i B i is the red, green, and blue channel values ​​of the i-th pixel, and n is the total number of pixels in the image;

[0103] Texture features are extracted using the gray-level co-occurrence matrix method. The specific steps are as follows: calculate the gray-level co-occurrence matrix P(d,θ), where d is the distance between pixels and θ is the direction; the element P(i,j;d,θ) of the gray-level co-occurrence matrix represents the number of times a pixel with gray value i and gray value j occur under the direction θ and distance d.

[0104] Calculate the statistical characteristics of the gray-level co-occurrence matrix, including contrast, correlation, energy, and homogeneity; the specific formula is as follows:

[0105] Contrast Ratio:

[0106]

[0107] Correlation:

[0108]

[0109] Where, μ i and μ j These are the mean values ​​of grayscale values ​​i and j, respectively, and σ i and σ j These are the standard deviations of grayscale values ​​i and j, respectively;

[0110] In addition, energy:

[0111]

[0112] Homogeneity:

[0113]

[0114] Where L is the number of gray levels in the image.

[0115] Furthermore, shape features are obtained through edge detection and contour extraction methods; the specific steps are as follows:

[0116] Edge detection algorithms are used to detect edges in images. These algorithms identify regions with significant brightness variations, which typically correspond to object boundaries. In geological lesion detection, edge detection helps identify the boundaries of lesion areas, providing a foundation for subsequent contour extraction and feature calculation. The Canny edge detection algorithm is commonly chosen due to its good noise resistance and edge localization accuracy. It's important to note that different parameters need to be adjusted for different conditions, such as the threshold in the Canny algorithm. These parameters should be optimized based on the specific image quality and lesion characteristics. Furthermore, it is recommended to denoise the image before edge detection to improve accuracy.

[0117] Edge contours are extracted using contour extraction algorithms, which are used to extract continuous contour lines from edge detection results. These contour lines can be further used to calculate geometric features to help identify and classify diseased areas. The extracted contours may contain a large number of noisy contours, which need to be filtered by geometric features such as area and perimeter to remove small noisy contours. To simplify the contours, polygon approximation methods can be used to approximate the contours, reducing the number of contour points and improving computational efficiency.

[0118] The geometric features of the calculated profile, including area, perimeter, and shape factor, are used to describe the shape and size of the diseased area, providing important information for disease classification and severity assessment. The specific calculations are as follows:

[0119] Area: Calculates the area of ​​the region enclosed by the contour using the coordinates of the contour points; in OpenCV, the contourArea function can be used to calculate the area.

[0120] Perimeter: Calculates the total length of the profile. In OpenCV, the arcLength function can be used to calculate the perimeter.

[0121] Shape factor: The shape factor is used to describe the shape of the outline, such as roundness, rectangularity, etc.

[0122] Roundness can be calculated using the following formula:

[0123]

[0124] Where A represents the area and P represents the perimeter;

[0125] Rectangularity can be calculated using the following formula:

[0126]

[0127] Where A represents the area, and BRA represents the area of ​​the circumscribed rectangle, which is the area of ​​the smallest circumscribed rectangle of the outline;

[0128] Preferably, the image recognition-based rock and soil geological exploration system includes,

[0129] Image acquisition module: Used to install image acquisition equipment in the geotechnical survey area to periodically acquire cross-sectional image information of the geotechnical layer;

[0130] Image preprocessing module: used to perform noise reduction and enhancement operations on the acquired images to improve image quality;

[0131] Feature extraction module: used to extract feature parameters related to rock and soil geology from the preprocessed image, including color, texture, and shape;

[0132] Disease identification module: Used to analyze feature parameters using a pre-trained deep learning model to identify the type and severity of geological diseases in images;

[0133] Early warning output module: Used to generate early warning information based on the identification results and notify staff via SMS and APP push.

[0134] The image acquisition module uses a high-definition camera as its image acquisition device, and the camera's installation position and angle are adjusted according to the location of the survey area.

[0135] Furthermore, the image preprocessing module includes,

[0136] The image is denoised using a median filtering algorithm.

[0137] The image is denoised using a Gaussian filtering algorithm.

[0138] Image enhancement is achieved using histogram equalization.

[0139] The feature extraction module includes,

[0140] Calculate the mean and variance, and other statistical measures, of different color channels in the image as color features;

[0141] Texture features are extracted using the gray-level co-occurrence matrix method;

[0142] Shape features are obtained through edge detection and contour extraction methods;

[0143] The early warning output module outputs information via SMS and APP push notifications, including disease type, location, and severity.

[0144] Furthermore, if the functionality is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0145] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0146] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0147] Operational Procedure: First, high-definition cameras are installed in the geotechnical survey area to periodically collect cross-sectional images of the geotechnical layers. The camera's installation position and angle are adjusted according to the location of the survey area to ensure comprehensive coverage. It should be noted that this method is applicable to survey scenarios where the geotechnical layers are directly exposed. For survey areas with surface vegetation cover, surface vegetation must be cleared from selected points before image acquisition to expose the target geotechnical layer surface. In the image preprocessing stage, color segmentation is further used to identify and remove residual vegetation pixels, ensuring that subsequent feature extraction is based on valid geotechnical image regions. Next, the acquired images are preprocessed, including noise removal using a median filtering algorithm. This algorithm replaces the value of each pixel in the image with the median value of its neighborhood. Simultaneously, a Gaussian filtering algorithm is used to further smooth the image. Gaussian filtering applies a weighted average to the image using a Gaussian kernel. The weights of the Gaussian kernel are calculated according to a Gaussian distribution based on the distance between each pixel and the center point, with the standard deviation determining the weight distribution range. Furthermore, histogram equalization is used to enhance image contrast. This method adjusts the histogram distribution of the image to make the contrast and brightness more uniform. Specifically, it calculates the image histogram, then the cumulative histogram, calculates new pixel values ​​based on the cumulative histogram, and finally maps the new pixel values ​​back to the original pixel value range to obtain the enhanced image.

[0148] Before implementing this method, site preparation is required:

[0149] Remove surface vegetation: Before installing the image acquisition equipment, remove surface vegetation at the selected survey points, removing turf, shrubs and other coverings to directly expose the surface of the target soil and rock layer.

[0150] Section trimming: For natural outcrops or artificially excavated sections, tools are used to clean the loose soil and weathered layer on the surface of the section to expose the fresh rock and soil interface, ensuring that the acquired images can reflect the true lithology, structure and tectonic features.

[0151] In the image preprocessing stage, any remaining small amounts of vegetation or debris can be processed in the following ways:

[0152] Color segmentation: By utilizing the significant differences between vegetation and soil in color channels (such as Normalized Difference Vegetation Index NDVI or a simple green channel threshold), the image is segmented by color, and green vegetation areas are marked as background.

[0153] Mask removal: Generate a binary mask for the region marked as vegetation, set the pixel values ​​of the region to zero or ignore them before feature extraction, so that they do not participate in the subsequent calculation of color, texture and shape features.

[0154] Morphological filtering: For discontinuous areas in soil and rock images caused by vegetation obstruction, morphological closing operation (dilation followed by erosion) is used to fill the gaps and restore the continuity of the soil and rock surface. If the vegetation coverage area is too large (more than 50% of the total image pixels) or cannot be effectively segmented, the system will automatically determine the image as invalid and issue a prompt message, suggesting that the management personnel clean up the site and re-acquire the image.

[0155] After preprocessing, the feature extraction stage begins, extracting feature parameters related to soil geology from the image. For color features, the mean and variance of the red, green, and blue color channels are calculated. The mean reflects the overall brightness level of the color channel, while the variance represents the dispersion of the color distribution. These statistics can be used to preliminarily determine the impact of geological hazards on the local area. Texture feature extraction employs the gray-level co-occurrence matrix (GLCM) method. First, the GLCM is calculated, reflecting the spatial dependence of pixel gray values ​​in the image. Then, based on the GLCM, statistical features such as contrast, correlation, energy, and homogeneity are calculated. Contrast measures the brightness difference between pixel pairs in the image, correlation reflects the linear correlation between pixel pairs, energy represents the uniformity of the image texture, and homogeneity measures the similarity of pixel pairs in the image. Shape feature extraction uses an edge detection algorithm to detect edges in the image, then a contour extraction algorithm to extract the edge contours. Finally, the geometric features of the contours are calculated, including area, perimeter, and shape factor. Area represents the size of the region enclosed by the contour, perimeter is the total length of the contour line, and shape factors such as roundness and rectangularity describe the shape characteristics of the contour.

[0156] The extracted feature parameters are then input into a pre-trained deep learning model, which is trained using a large amount of geological disease image data and can accurately identify the type and severity of geological diseases in the images. Based on the model's identification results, the system generates early warning information and promptly notifies management personnel via SMS, APP push notifications, and other means. The early warning information includes key information such as disease type, location, and severity, enabling relevant personnel to quickly take appropriate prevention and control measures. The entire operation automates the process from image acquisition to disease early warning, improving the efficiency and accuracy of disease monitoring and providing strong support for the early detection and timely prevention and control of geological diseases.

[0157] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A rock and soil geological exploration method based on image recognition, characterized in that: The process includes the following steps: image acquisition, where image acquisition equipment is installed in the geotechnical survey area to periodically acquire cross-sectional image information of the geotechnical layers; Image preprocessing involves performing noise reduction and enhancement operations on the acquired images to improve image quality. Feature extraction involves extracting feature parameters related to rock and soil geology from the preprocessed image. These feature parameters include color, texture, and shape. Disease identification utilizes a pre-trained deep learning model to analyze feature parameters and identify the type and severity of geological diseases in images; and... The system generates early warning information based on the identification results and notifies management personnel through various means.

2. The image recognition-based rock and soil geological exploration method as described in claim 1, characterized in that: The image acquisition device is a high-definition camera, and the installation position and angle of the camera are adjusted according to the location of the survey area; The image preprocessing operations include median filtering, Gaussian filtering for noise reduction, and histogram equalization for enhancement.

3. The image recognition-based rock and soil geological exploration method as described in claim 2, characterized in that: The feature extraction includes the extraction of color features, texture features, and shape features. Color features are obtained by calculating the statistics of different color channels in the image, texture features are obtained using the gray-level co-occurrence matrix method, and shape features are obtained using edge detection and contour extraction methods.

4. The image recognition-based rock and soil geological exploration method as described in claim 3, characterized in that: The image preprocessing includes denoising the image using a first denoising method, namely, a median filtering algorithm. The median filtering algorithm removes noise by replacing the value of each pixel in the image with the median value of its neighborhood. The specific steps are as follows: Select a neighboring window, usually a 3×3 or 5×5 rectangular window; Align the center of the window with each pixel in the image and extract the pixel values ​​within the window; Sort the pixel values ​​within the window and take the median value as the new value for that pixel; Repeat the above steps until the entire image has been processed.

5. The image recognition-based rock and soil geological exploration method as described in claim 4, characterized in that: The second denoising method, Gaussian filtering, is used to denoise the image. Gaussian filtering smooths the image using a Gaussian kernel, thereby removing noise. The formula for Gaussian filtering is: Where σ is the standard deviation of the Gaussian kernel, controlling the smoothness of the filter; x and y represent the coordinates of pixels in the image, i.e., x and y are the offsets relative to the center of the Gaussian kernel; and e is the base of the natural logarithm, typically taken as 2.

718. The specific steps are as follows: Choose a Gaussian kernel, usually a 3×3 or 5×5 matrix; Calculate the weights of the Gaussian kernel, and calculate the weight of each position according to the Gaussian formula; Align the center of the Gaussian kernel with each pixel in the image, and calculate the new value of that pixel as a weighted sum of the pixel values ​​within the Gaussian kernel. Repeat the above steps until the entire image has been processed.

6. The image recognition-based rock and soil geological exploration method as described in claim 5, characterized in that: Image enhancement is performed using histogram equalization. Histogram equalization adjusts the histogram distribution of an image to make its contrast and brightness more uniform. The specific steps are as follows: Calculate the histogram h(i) of the image, where i is the pixel value and h(i) is the number of times pixel value i appears; Calculate the cumulative histogram c(i). Calculate the new pixel value s(i). Among them, c min It is the minimum value of the cumulative histogram, and n is the total number of pixels in the image; The new pixel value s(i) is mapped back to the original pixel value range to obtain the enhanced image.

7. The image recognition-based rock and soil geological exploration method as described in claim 6, characterized in that: The feature extraction includes, Calculate the mean and variance, and other statistical measures, of different color channels in the image as color features; The specific steps are as follows: Calculate the mean μ of the red, green, and blue color channels respectively. R μ G μ B and variance Among them, R i G i B i is the red, green, and blue channel values ​​of the i-th pixel, and n is the total number of pixels in the image; Texture features are extracted using the gray-level co-occurrence matrix method; The specific steps are as follows: Calculate the gray-level co-occurrence matrix P(d,θ), where d is the distance between pixels and θ is the direction; the element P(i,j;d,θ) of the gray-level co-occurrence matrix represents the number of times a pixel with gray value i and a pixel with gray value j appear under the direction θ and distance d. Calculate the statistical characteristics of the gray-level co-occurrence matrix, including contrast, correlation, energy, and homogeneity; The specific formula is: Contrast: Correlation: Where, μ i and μ j These are the mean values ​​of grayscale values ​​i and j, respectively, and σ is the mean value of grayscale value i. i and σ j These are the standard deviations of grayscale values ​​i and j, respectively; Shape features are obtained through edge detection and contour extraction methods; The specific steps are as follows: Use an edge detection algorithm to detect edges in an image; Use a contour extraction algorithm to extract edge contours; Calculating the geometric features of the contour, including area, perimeter, and shape factor, is a method for rock and soil geological exploration based on image recognition, comprising the following steps: Image acquisition involves installing image acquisition equipment in the geotechnical survey area to periodically collect cross-sectional image information of the geotechnical layers. Image preprocessing involves performing noise reduction and enhancement operations on the acquired images to improve image quality. Feature extraction involves extracting feature parameters related to rock and soil geology from the preprocessed image. These feature parameters include color, texture, and shape. Disease identification utilizes a pre-trained deep learning model to analyze feature parameters and identify the type and severity of geological diseases in images; and... The system generates early warning information based on the identification results and notifies management personnel through various means.

8. A rock and soil geological exploration system based on image recognition, characterized in that: Includes an image acquisition module: used to install image acquisition equipment in the geotechnical survey area and periodically acquire cross-sectional image information of the geotechnical layers; Image preprocessing module: used to perform noise reduction and enhancement operations on the acquired images to improve image quality; Feature extraction module: used to extract feature parameters related to rock and soil geology from the preprocessed image, including color, texture, and shape; Disease identification module: Used to analyze feature parameters using a pre-trained deep learning model to identify the type and severity of geological diseases in images; Early warning output module: used to generate early warning information based on the identification results and notify staff via SMS and APP push; The image acquisition device in the image acquisition module is a high-definition camera, and the installation position and angle of the camera are adjusted according to the location of the survey area.

9. A computer device, characterized in that: It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images