Mine resource exploration field information acquisition method and device based on big data and image recognition technology

Abstract

The invention provides a big data and image recognition technology-based mine resource exploration field information acquisition method and apparatus. The method comprises the steps of receiving an exploration site photo of related mine resources uploaded by a client; carrying out image identification and analysis on the uploaded investigation site photo, and automatically identifying stratum boundaries and features; the stratum naming suggestion is generated by predicting and classifying exploration data, and the exploration data comprises an exploration site photo, a geographic position and photographing time; and returning the generated boundaries and features of the different stratums and the stratigraphic naming suggestions to the client. According to the invention, the efficiency and precision of field data acquisition can be effectively improved, powerful support is provided for decision makers, and the method has a wide development prospect in the mine resource exploration industry.

Description

Technical Field

[0001] The present invention relates to the fields of mine exploration, big data and image recognition, and in particular to a method and device for collecting field information of mine resource exploration based on big data and image recognition technology. Background Art

[0002] Field information collection for mining resource exploration is one of the important links in the mining exploration process. The traditional field information collection method mainly relies on manual exploration, which has problems such as low information collection efficiency, high cost, and low accuracy. With the continuous development of big data and image recognition technology, these new technologies have provided new solutions for field information collection for mining resource exploration. How to use modern technology to improve the efficiency and accuracy of field information collection for mining resource exploration has become an urgent problem to be solved. Summary of the invention

[0003] The present invention provides a method and device for collecting field information of mining resource exploration based on big data and image recognition technology, which can effectively improve the efficiency and accuracy of field data collection, provide strong support for decision makers, and have broad development prospects in the mining resource exploration industry.

[0004] A method for collecting field information of mining resource exploration based on big data and image recognition technology comprises the following steps:

[0005] Receive on-site exploration photos of relevant mining resources uploaded by the client;

[0006] Perform image recognition and analysis on uploaded survey site photos to automatically identify stratum boundaries and features;

[0007] Generate stratigraphic naming suggestions by predicting and classifying survey data, wherein the survey data includes survey site photos, geographic locations, and photo taking times;

[0008] The generated boundaries and features of different strata and strata naming suggestions are returned to the client.

[0009] Furthermore, the client is a mobile phone or a tablet computer.

[0010] Furthermore, the uploaded survey site photos are subjected to image recognition and analysis to automatically identify stratum boundaries and features, specifically including:

[0011] All previous exploration color maps are used as input data. The input data is uniformly taken as the center point of the image 300*300 pixels, with a total of three channels of R, G, and B, and the pixel values ​​of each channel are normalized. Color images have 3 channels, so all color map data are stored in a four-dimensional array, and its form is (300,100,100,3). Recorded as X; the output data is the geotechnical type, the type number is 0, 1, 2, 3, 4...n, recorded as Y; the input data X and output data Y are randomly divided into 80% training set and 20% test set; the stacking model under the Keras module in TensorFlow2 is used to build a multi-layer convolutional neural network recognition model;

[0012] Use image datasets with annotated stratum boundaries and features for supervised learning and optimize the training of multi-layer convolutional neural network recognition models;

[0013] Pre-process the survey site photos, including image noise reduction and contrast enhancement;

[0014] The preprocessed on-site photos are predicted through an optimized and trained multi-layer convolutional neural network recognition model to automatically identify and mark the boundaries and features of different strata.

[0015] Furthermore, the prediction and classification of the exploration data to generate stratigraphic naming suggestions specifically include:

[0016] Establishing a stratigraphic database, which includes stratigraphic characteristics, naming rules, and related geological information and survey data, and performing data cleaning and organization;

[0017] Use machine learning algorithms to train the stratigraphic database;

[0018] Process and extract features from the survey data obtained at the survey site;

[0019] Based on the trained stratigraphic database, the processed and feature-extracted exploration data are predicted and classified to generate stratigraphic naming suggestions.

[0020] Furthermore, the machine learning algorithm includes a decision tree, a random forest or a deep learning model.

[0021] A mine resource exploration field information collection device based on big data and image recognition technology, comprising:

[0022] A data receiving module is used to receive the exploration site photos of relevant mining resources uploaded by the client;

[0023] Image recognition module, used to perform image recognition and analysis on uploaded survey site photos, and automatically identify stratum boundaries and features;

[0024] A big data analysis module, which is used to generate stratigraphic naming suggestions by predicting and classifying exploration data, wherein the exploration data includes exploration site photos, geographic locations, and photo-taking times;

[0025] The data delivery module returns the generated boundaries and features of different strata and stratum naming suggestions to the client.

[0026] Furthermore, the client is a mobile phone or a tablet computer.

[0027] Furthermore, the image recognition module is specifically used for:

[0028] All previous exploration color maps are used as input data. The input data is uniformly taken as the center point of the image 300*300 pixels, with a total of three channels of R, G, and B, and the pixel values ​​of each channel are normalized. Color images have 3 channels, so all color map data are stored in a four-dimensional array, and its form is (300,100,100,3). Recorded as X; the output data is the geotechnical type, the type number is 0, 1, 2, 3, 4...n, recorded as Y; the input data X and output data Y are randomly divided into 80% training set and 20% test set; the stacking model under the Keras module in TensorFlow2 is used to build a multi-layer convolutional neural network recognition model;

[0029] Use image datasets with annotated stratum boundaries and features for supervised learning and optimize the training of multi-layer convolutional neural network recognition models;

[0030] Pre-process the survey site photos, including image noise reduction and contrast enhancement;

[0031] The preprocessed on-site photos are predicted through an optimized and trained multi-layer convolutional neural network recognition model to automatically identify and mark the boundaries and features of different strata.

[0032] Furthermore, the big data analysis module is specifically used for:

[0033] Establishing a stratigraphic database, which includes stratigraphic characteristics, naming rules, and related geological information and survey data, and performing data cleaning and organization;

[0034] Use machine learning algorithms to train the stratigraphic database;

[0035] Process and extract features from the survey data obtained at the survey site;

[0036] Based on the trained stratigraphic database, the processed and feature-extracted exploration data are predicted and classified to generate stratigraphic naming suggestions.

[0037] Furthermore, the machine learning algorithm includes a decision tree, a random forest or a deep learning model.

[0038] The present invention realizes automatic stratum boundary and feature recognition through image recognition, provides accurate stratum naming suggestions by using big data analysis, and greatly improves exploration efficiency and information management convenience through rapid acquisition function and data processing and storage function. The popularization and application of the present invention will bring revolutionary changes to mine exploration, improve exploration efficiency, reduce costs, and further promote the scientific and rational use of mine resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the method for collecting field information for mining resource exploration based on big data and image recognition technology of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Figure 1 The figure shows a flow chart of a method for collecting field information of mining resource exploration based on big data and image recognition technology according to an embodiment of the present invention, which can be presented in the form of an APP during implementation. The method may include the following steps:

[0042] S1. Users register an account on a client (such as a mobile phone or other mobile communication device) and can use the functions of the APP after logging in and authenticating.

[0043] S2. When working in the field, users use mobile phones or other devices to take photos of the exploration site of relevant mining resources and upload them to the server. At the same time, users can also add additional information such as geographic location and photo time.

[0044] S3. The server performs image recognition and analysis on the uploaded survey site photos, automatically identifies the boundaries and features of the strata, and generates suggestions for stratum naming by predicting and classifying the survey data. Surveyors only need to take photos of the site and upload them, and the APP can quickly identify and return the stratum information, greatly reducing the time and labor of on-site work.

[0045] Step S3 specifically includes the following implementation steps:

[0046] S3.1 Image Recognition

[0047] A multi-layer convolutional neural network recognition model is used to perform image recognition on survey site photos.

[0048] First, a multi-layer convolutional neural network recognition model is constructed using previous exploration data: all previous exploration color maps are used as input data. The input data is uniformly taken as the center point of the image 300*300 pixels, with a total of three channels of R, G, and B, and the pixel values ​​of each channel are normalized. The color image has 3 channels, so all color image data can be stored in a four-dimensional array, and its form is (300,100,100,3). Recorded as X. The output data is the geotechnical type, and the type number is 0, 1, 2, 3, 4...n, recorded as Y. The input data X and output data Y are randomly divided into 80% training set and 20% test set; the stacking model under the Keras module in TensorFlow2 is used to construct a multi-layer convolutional neural network recognition model.

[0049] During the training process, supervised learning is performed using an image dataset with annotated layer boundaries and features to optimize the accuracy and robustness of the model.

[0050] Preprocess the survey site photos, including image noise reduction, contrast enhancement and other operations, to improve image quality.

[0051] The preprocessed on-site photos are predicted through an optimized and trained multi-layer convolutional neural network recognition model to automatically identify and mark the boundaries and features of different strata.

[0052] S3.2 Big Data Analysis

[0053] Establish a stratigraphic database, including stratigraphic characteristics, naming rules and related geological information and survey data (such as stratigraphic location depth information, rock type, mineral composition, geological structure), and perform effective data cleaning and organization.

[0054] The stratigraphic database is trained using machine learning algorithms such as decision trees, random forests, or deep learning models (such as multi-layer perceptrons, recurrent neural networks, etc.). The trained database not only improves the accuracy of stratigraphic naming, but also reduces subjective differences between different surveyors.

[0055] The survey data obtained at the survey site are processed and features extracted, wherein the survey data includes survey site photos, geographic location, photo taking time, etc.

[0056] Based on the trained stratigraphic database, the processed and feature-extracted exploration data are predicted and classified to generate accurate stratigraphic naming suggestions.

[0057] Image recognition algorithm formula:

[0058] Input: Investigation site photo (Image)

[0059] Output: Boundaries, Features

[0060] The output of the image recognition algorithm can be expressed as: Boundaries, Features = ImageRecognitionAlgorithm(Image)

[0061] Big data analysis algorithm formula:

[0062] Input: Survey data (Data)

[0063] Output: Naming Suggestions

[0064] The output of the big data analysis algorithm can be expressed as: Naming_Suggestions = DataAnalysisAlgorithm(Data).

[0065] Through deep learning and big data analysis algorithms, the present invention enables the APP to automatically process uploaded photos and data, reducing the need for manual intervention. This not only improves the processing speed, but also improves the consistency and reliability of data processing; through machine learning algorithms, the APP can continuously learn and optimize its analysis model from new data to ensure the provision of the latest and most accurate stratigraphic information and naming suggestions. This real-time update and optimization mechanism ensures the continuous improvement and efficient operation of the system; the present invention can integrate and analyze data from multiple sources such as photos, geographic location, time, etc., to provide more comprehensive and accurate stratigraphic information. This multi-source data fusion processing method can significantly improve the data utilization value and decision-making support capabilities.

[0066] S4. Return the results of the big data analysis in S3 (stratum boundaries, characteristics, and naming suggestions for strata) to the user APP for display, highlighting the judgment points of the data analysis system and displaying detailed information of the stratum.

[0067] The APP of the present invention provides an intuitive and easy-to-use interface design. Surveyors only need to perform simple operations to complete data collection and upload, which greatly reduces the threshold for use and training costs. According to user needs, the APP can provide personalized data display methods, including stratum information, location data, time records, etc., to help users better understand and use the collected data.

[0068] S5. Users can provide feedback after seeing the judgment given by the analysis system on the APP.

[0069] S5.1. If the user believes that the analysis system's judgment is correct, the judgment will be entered into the geological big data database to enrich the geological big data.

[0070] S5.2. If the user believes that the analysis system has made an error, the system will prompt the user to mark the data and then take multi-angle photos of the stratum again. The analysis system will compare the information marked by the user with the geological database again and perform correction calculations.

[0071] Surveyors can instantly view the system analysis results and provide feedback on the results. If the user believes that the system is correct, the results can be entered into the geological database; if it is considered incorrect, the system will prompt the user to take photos again and re-analyze. This two-way feedback mechanism not only enriches the database, but also continuously improves the accuracy of the system.

[0072] The present invention will automatically upload the collected data to the server for processing and storage. The server-side algorithm will further analyze and organize the data to generate corresponding reports and results. Surveyors can view and download these reports at any time through the APP. This automated data processing and storage function improves the convenience of data management.

[0073] The embodiment of the present invention further provides a mining resource exploration field information collection device based on big data and image recognition technology, comprising:

[0074] A data receiving module is used to receive the exploration site photos of relevant mining resources uploaded by the client;

[0075] Image recognition module, used to perform image recognition and analysis on uploaded survey site photos, and automatically identify stratum boundaries and features;

[0076] A big data analysis module, which is used to generate stratigraphic naming suggestions by predicting and classifying exploration data, wherein the exploration data includes exploration site photos, geographic locations, and photo-taking times;

[0077] The data delivery module returns the generated boundaries and features of different strata and stratum naming suggestions to the client.

[0078] The field information collection method for mining resource exploration based on big data and image recognition technology of the present invention integrates technologies such as image recognition, big data analysis and mobile applications. It aims to solve the problems of time-consuming, inefficient and inconsistent stratum naming in traditional resource exploration methods. Through the present invention, surveyors can quickly and accurately collect field data and obtain accuracy assessments of stratum naming, thereby improving exploration efficiency and the accuracy of geological surveys.

[0079] The above is only a specific embodiment of the present invention, but the protection scope of the present invention 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 by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for collecting field information for mining resource exploration based on big data and image recognition technology, characterized in that: The steps include: Receive on-site exploration photos of relevant mining resources uploaded by the client; Perform image recognition and analysis on uploaded survey site photos to automatically identify stratum boundaries and features; Generate stratigraphic naming suggestions by predicting and classifying survey data, wherein the survey data includes survey site photos, geographic locations, and photo taking times; The generated boundaries and features of different strata and strata naming suggestions are returned to the client.

2. The method for collecting field information for mining resource exploration based on big data and image recognition technology according to claim 1, characterized in that: The client is a mobile phone or a tablet computer.

3. The method for collecting field information for mining resource exploration based on big data and image recognition technology according to claim 1, characterized in that: The uploaded survey site photos are subjected to image recognition and analysis to automatically identify stratum boundaries and features, specifically including: All previous exploration color maps are used as input data. The input data is uniformly taken as the center point of the image 300*300 pixels, with a total of three channels of R, G, and B, and the pixel values ​​of each channel are normalized. Color images have 3 channels, so all color map data are stored in a four-dimensional array, and its form is (300,100,100,3). Recorded as X; the output data is the geotechnical type, the type number is 0, 1, 2, 3, 4...n, recorded as Y; the input data X and output data Y are randomly divided into 80% training set and 20% test set; the stacking model under the Keras module in TensorFlow2 is used to build a multi-layer convolutional neural network recognition model; Use image datasets with annotated stratum boundaries and features for supervised learning and optimize the training of multi-layer convolutional neural network recognition models; Pre-process the survey site photos, including image noise reduction and contrast enhancement; The preprocessed on-site photos are predicted through an optimized and trained multi-layer convolutional neural network recognition model to automatically identify and mark the boundaries and features of different strata.

4. The method for collecting field information for mining resource exploration based on big data and image recognition technology according to claim 1, characterized in that: The above-mentioned prediction and classification of the exploration data to generate stratigraphic naming suggestions specifically include: Establishing a stratigraphic database, which includes stratigraphic characteristics, naming rules, and related geological information and survey data, and performing data cleaning and organization; Use machine learning algorithms to train the stratigraphic database; Process and extract features from the survey data obtained at the survey site; Based on the trained stratigraphic database, the processed and feature-extracted exploration data are predicted and classified to generate stratigraphic naming suggestions.

5. The method for collecting field information for mining resource exploration based on big data and image recognition technology according to claim 4, characterized in that: The machine learning algorithm includes a decision tree, a random forest, or a deep learning model.

6. A mining resource exploration field information collection device based on big data and image recognition technology, characterized in that: include: A data receiving module is used to receive the exploration site photos of relevant mining resources uploaded by the client; Image recognition module, used to perform image recognition and analysis on uploaded survey site photos, and automatically identify stratum boundaries and features; A big data analysis module, which is used to generate stratigraphic naming suggestions by predicting and classifying exploration data, wherein the exploration data includes exploration site photos, geographic locations, and photo-taking times; The data delivery module returns the generated boundaries and features of different strata and stratum naming suggestions to the client.

7. The mining resource exploration field information collection device based on big data and image recognition technology as claimed in claim 6, characterized in that: The client is a mobile phone or a tablet computer.

8. The mining resource exploration field information collection device based on big data and image recognition technology as claimed in claim 6, characterized in that: The image recognition module is specifically used for: All previous exploration color maps are used as input data. The input data is uniformly taken as the center point of the image 300*300 pixels, with a total of three channels of R, G, and B, and the pixel values ​​of each channel are normalized. Color images have 3 channels, so all color map data are stored in a four-dimensional array, and its form is (300,100,100,3). Recorded as X; the output data is the geotechnical type, the type number is 0, 1, 2, 3, 4...n, recorded as Y; the input data X and output data Y are randomly divided into 80% training set and 20% test set; the stacking model under the Keras module in TensorFlow2 is used to build a multi-layer convolutional neural network recognition model; Use image datasets with annotated stratum boundaries and features for supervised learning and optimize the training of multi-layer convolutional neural network recognition models; Pre-process the survey site photos, including image noise reduction and contrast enhancement; The preprocessed on-site photos are predicted through an optimized and trained multi-layer convolutional neural network recognition model to automatically identify and mark the boundaries and features of different strata.

9. The mining resource exploration field information collection device based on big data and image recognition technology as claimed in claim 6, characterized in that: The big data analysis module is specifically used for: Establishing a stratigraphic database, which includes stratigraphic characteristics, naming rules, and related geological information and survey data, and performing data cleaning and organization; Use machine learning algorithms to train the stratigraphic database; Process and extract features from the survey data obtained at the survey site; Based on the trained stratigraphic database, the processed and feature-extracted exploration data are predicted and classified to generate stratigraphic naming suggestions.

10. The mining resource exploration field information collection device based on big data and image recognition technology as claimed in claim 9, characterized in that: The machine learning algorithm includes a decision tree, a random forest, or a deep learning model.

Images