Ore classification and particle size classification method and system based on deep learning network

Abstract

The present invention relates to the field of mining, and discloses an ore classification and particle size classification method and system based on a deep learning network, including an ore information collection terminal collecting ore information of an ore to be classified, an ore identification module obtaining a response state according to an identification feature, and identifying the ore type according to the ore identification information in the ore information. If the ore type is identified as a known ore type, the first classification information associated with the ore identification information is obtained, the ore identification feature in the first classification information is compared with the ore feature in the ore identification information, and the ore identification module outputs the classification and particle size classification results of the ore. The present invention can efficiently and accurately realize ore classification and particle size classification.

Description

Technical Field

[0001] The present invention relates to the field of mining, and in particular to a method and system for ore classification and particle size grading based on a deep learning network. Background Art

[0002] In the mining industry, ore classification and particle size classification are the basic links in ore processing, utilization and value assessment. Traditional ore classification and particle size classification mainly rely on manual observation and simple instrument measurement. This method is not only time-consuming and labor-intensive, but also greatly affected by human factors, and its accuracy is difficult to guarantee. With the advancement of science and technology, especially the rapid development of information technology and intelligent technology, new solutions have been provided for ore classification and particle size classification.

[0003] In the early days, some automated and semi-automated ore classification technologies began to emerge, such as methods based on spectral analysis, X-ray diffraction and other technologies, but these methods often have high requirements for equipment and can only classify specific types of ores, with poor universality. At the same time, these technologies are often unable to classify the particle size of ores, or can only perform simple coarse and fine classification, which cannot meet the needs of refined management.

[0004] In recent years, deep learning networks have achieved remarkable results in the fields of image recognition and speech recognition. Their powerful feature extraction and pattern recognition capabilities provide new ideas for ore classification and particle size classification. By training the deep learning model, it can automatically learn and extract the texture, color, shape and other features of the ore, thereby achieving accurate classification. At the same time, combined with the physical characteristics and density information of the ore, particle size classification can be further achieved.

[0005] However, the existing ore classification and particle size classification technology based on deep learning networks still has some problems. On the one hand, due to the wide variety of ores and complex characteristics, a single model is often difficult to cover all types of ores, resulting in limited classification accuracy. On the other hand, the particle size classification of ores requires comprehensive consideration of multiple factors, such as shape, size, density, etc. The existing technology can often only perform simple division and cannot meet the needs of high-precision classification.

[0006] In addition, existing ore classification and particle size classification systems often use a single local processing or cloud processing method, which has problems such as slow processing speed and large resource usage. Especially when processing a large amount of ore information, a single processing method often cannot meet the requirements of real-time and accuracy.

[0007] Therefore, developing a system and method that can efficiently and accurately achieve ore classification and particle size classification has become an urgent problem to be solved in the current mining field. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for ore classification and particle size classification based on a deep learning network, comprising the following steps:

[0009] Step 1: The ore information collection terminal collects the ore information of the ore to be classified and sends it to the ore identification module. The ore identification module extracts the corresponding identification feature from the received ore information according to the preset identification feature item. The ore identification module obtains the response status according to the identification feature. If it is an identification response, it goes to step 2; if it is no response, the received ore information is discarded;

[0010] Step 2: According to the ore identification information in the ore information, the ore identification module identifies the ore type. If it is a known ore type, then go to step 3; if it is an unknown ore type, then go to step 4;

[0011] Step 3, obtaining the first classification information associated with the ore identification information, and comparing the ore identification features in the first classification information with the ore features in the ore identification information. If the comparison is successful, proceed to step 6; if the comparison is unsuccessful, it is determined to be a classification error, and proceed to step 4;

[0012] Step 4, obtaining the second classification information associated with the ore identification information, and matching the second classification information with the classification information in the ore database of the cloud data analysis server. If a match is found, the cloud data analysis server sends the generated third classification information to the ore identification module, and then proceeds to step 5;

[0013] Step 5, the ore identification module generates ore classification and particle size classification information according to the third classification information received from the cloud data analysis server and the particle size classification information associated with the ore information;

[0014] Step six: the ore identification module outputs the ore classification and particle size classification results.

[0015] Furthermore, the ore identification module extracts corresponding identification features from the received ore information according to preset identification feature items, and the ore identification module obtains a response state according to the identification features, including:

[0016] The ore information of the ore to be classified includes ore texture features, color features, physical features and density features; the preset identification feature items include density range; the physical features include ore size and weight;

[0017] If the ore identification module extracts the density feature within the density range from the received ore information of the ore to be classified, the ore identification module performs an identification response, otherwise, it does not respond.

[0018] Furthermore, the ore identification module identifies the ore type according to the ore identification information in the ore information, including:

[0019] If the ore identification module matches the corresponding density feature in the local data module, it is a known ore type, otherwise it is an unknown ore.

[0020] Furthermore, the obtaining of the first classification information associated with the ore identification information and comparing the ore identification features in the first classification information with the ore features in the ore identification information include:

[0021] The ore texture features and color features corresponding to the density features are obtained in the local data module, and compared with the ore texture features and color features of the ore to be classified. If they are consistent, the comparison passes, and the particle size of the ore is obtained according to the density features.

[0022] Furthermore, the step of obtaining the second classification information associated with the ore identification information and matching the second classification information with the classification information in the ore database of the cloud data analysis server includes:

[0023] The second classification information is positioning information. According to the positioning information, the ore database corresponding to the positioning information is obtained in the cloud data analysis server. According to the ore information of the ore to be classified, matching is performed in the ore database corresponding to the positioning information to obtain the classification information of the ore to be classified.

[0024] Furthermore, the cloud data analysis server sends the generated third classification information to the ore identification module, including: the third classification information includes the ore type, ore texture characteristics and color characteristics of the ore to be classified, and is sent to the local data storage module of the ore identification module for storage.

[0025] Furthermore, the ore identification module generates ore classification and particle size classification information according to the third classification information received from the cloud data analysis server and the particle size classification information associated with the ore information, including:

[0026] The ore identification module obtains the ore classification according to the third classification information sent by the received cloud data analysis server, and obtains the ore particle size classification according to the density information and physical characteristics.

[0027] An ore classification and particle size grading system based on a deep learning network, applying the ore classification and particle size grading method based on a deep learning network, including: a cloud data analysis server, an ore identification module, a data processing module, a communication module and an ore information collection device;

[0028] The ore identification module and the communication module are respectively connected to the data processing module; the cloud data analysis server and the ore information collection device are respectively connected to the communication module for communication.

[0029] The beneficial effects of the present invention are:

[0030] Combination of cloud and local: The combination of local rapid identification and cloud-based deep analysis ensures both real-time processing and accurate classification.

[0031] Comprehensive information collection: Multi-dimensional, high-precision information collection provides a solid foundation for accurate classification and particle size grading of ores.

[0032] Automated processing: From information collection to result output, the entire process is automated, which greatly reduces manual intervention and improves work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the principle of ore classification and particle size classification method based on deep learning network;

[0034] Figure 2 Schematic diagram of the principle of ore classification and particle size classification system based on deep learning network;

[0035] Figure 3 This is a schematic diagram of the principle of the ore identification module. DETAILED DESCRIPTION

[0036] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0037] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

[0038] like Figure 1 As shown, the ore classification and particle size classification method based on deep learning network includes the following steps:

[0039] Step 1: The ore information collection terminal collects the ore information of the ore to be classified and sends it to the ore identification module. The ore identification module extracts the corresponding identification feature from the received ore information according to the preset identification feature item. The ore identification module obtains the response status according to the identification feature. If it is an identification response, it goes to step 2; if it is no response, the received ore information is discarded;

[0040] Step 2: According to the ore identification information in the ore information, the ore identification module identifies the ore type. If it is a known ore type, proceed to step 3; if it is an unknown ore type, proceed to step 4.

[0041] Step 3, obtaining the first classification information associated with the ore identification information, and comparing the ore identification features in the first classification information with the ore features in the ore identification information. If the comparison is successful, proceed to step 6; if the comparison is unsuccessful, it is determined to be a classification error, and proceed to step 4;

[0042] Step 4, obtaining the second classification information associated with the ore identification information, and matching the second classification information with the classification information in the ore database of the cloud data analysis server. If a match is found, the cloud data analysis server sends the generated third classification information to the ore identification module, and then proceeds to step 5;

[0043] Step 5, the ore identification module generates ore classification and particle size classification information according to the third classification information received from the cloud data analysis server and the particle size classification information associated with the ore information;

[0044] Step six: the ore identification module outputs the ore classification and particle size classification results.

[0045] The ore identification module extracts the corresponding identification feature from the received ore information according to the preset identification feature item, and the ore identification module obtains the response state according to the identification feature, including:

[0046] The ore information of the ore to be classified includes ore texture features, color features, physical features and density features; the preset identification feature items include density range; the physical features include ore size and weight;

[0047] If the ore identification module extracts the density feature within the density range from the received ore information of the ore to be classified, the ore identification module performs an identification response, otherwise, it does not respond.

[0048] The ore identification module identifies the ore type according to the ore identification information in the ore information, including:

[0049] If the ore identification module matches the corresponding density feature in the local data module, it is a known ore type, otherwise it is an unknown ore.

[0050] The step of obtaining the first classification information associated with the ore identification information and comparing the ore identification feature in the first classification information with the ore feature in the ore identification information includes:

[0051] The ore texture features and color features corresponding to the density features are obtained in the local data module, and compared with the ore texture features and color features of the ore to be classified. If they are consistent, the comparison passes, and the particle size of the ore is obtained according to the density features.

[0052] The step of obtaining the second classification information associated with the ore identification information and matching the second classification information with the classification information in the ore database of the cloud data analysis server includes:

[0053] The second classification information is positioning information. According to the positioning information, the ore database corresponding to the positioning information is obtained in the cloud data analysis server. According to the ore information of the ore to be classified, matching is performed in the ore database corresponding to the positioning information to obtain the classification information of the ore to be classified.

[0054] The cloud data analysis server sends the generated third classification information to the ore identification module, including: the third classification information includes the ore type, ore texture characteristics and color characteristics of the ore to be classified, and is sent to the local data storage module of the ore identification module for storage.

[0055] The ore identification module generates ore classification and particle size classification information according to the third classification information received from the cloud data analysis server and the particle size classification information associated with the ore information, including:

[0056] The ore identification module obtains the ore classification according to the third classification information sent by the received cloud data analysis server, and obtains the ore particle size classification according to the density information and physical characteristics.

[0057] like Figure 2 As shown, a deep learning network-based ore classification and particle size classification system applies a deep learning network-based ore classification and particle size classification method, including: a cloud data analysis server, an ore identification module, a data processing module, a communication module and an ore information collection device;

[0058] The ore identification module and the communication module are respectively connected to the data processing module; the cloud data analysis server and the ore information collection device are respectively connected to the communication module for communication.

[0059] like Figure 3 As shown, the ore identification module includes a local data storage module and an ore identification device; the local data storage module and the ore identification device are respectively connected to the data processing module.

[0060] Specifically, this method mainly includes the following core steps: ore information collection, preliminary identification and screening, local classification comparison, cloud matching and analysis, classification and granularity classification information generation, and final result output. Each step is closely linked and together constitutes a complete process of ore classification and granularity classification.

[0061] Step 1: Ore information collection and preliminary identification

[0062] The ore information collection terminal is the starting point of the entire process, responsible for collecting all-round information of the ore to be classified, including but not limited to the texture characteristics, color characteristics, physical characteristics (such as size, weight) and density characteristics of the ore. This information is obtained through high-precision sensors and imaging equipment to ensure the accuracy and reliability of the data.

[0063] The collected ore information is then sent to the ore identification module. In this module, the received ore information is feature extracted according to the preset identification feature items (such as a specific density range). If the density feature of the ore falls within the preset density range, the ore identification module will generate an identification response and enter the next step of processing; otherwise, the information is considered to be unqualified for the current processing conditions and is discarded, thereby effectively filtering out irrelevant or erroneous data.

[0064] Step 2: Ore type identification

[0065] After confirming that the ore information is valid, the ore identification module further attempts to identify the type of ore based on the identification information in the ore information, such as texture, color, etc. This step relies on the known ore type data stored in the local data module. If a match can be found in the local data, it is identified as a known ore type and the process continues; if no match is found, it is considered an unknown ore type and requires more in-depth analysis.

[0066] Step 3: Compare local classification information

[0067] For a known ore type, the ore identification module obtains the first classification information associated with it, which usually includes detailed identification features of the ore type, such as finer texture, color description, etc. These features are then compared one by one with the actual features of the ore to be classified. If all features are consistent, the classification is correct and the process proceeds to step six; if there is any inconsistency, the classification is determined to be wrong and the process proceeds to step four for further processing.

[0068] Step 4: Cloud matching and third-category information generation

[0069] In the case of unknown ores or misclassification, the ore identification module will call the second classification information, which is usually a positioning information based on geographic location or other clues. Using this information, the ore identification module requests the cloud data analysis server for deep matching. The cloud server has a huge ore database, which can quickly locate the ore data in the relevant area based on the positioning information, and conduct a detailed comparison with the information of the ore to be classified, thereby generating the third classification information containing detailed information such as ore type, texture characteristics, color characteristics, etc.

[0070] Step 5: Classification and granularity classification information generation

[0071] The cloud data analysis server sends the generated third classification information back to the ore identification module, which combines the received information with the physical characteristics of the ore (such as size, weight) and density information to perform particle size classification. This process not only determines the type of ore, but also accurately divides its particle size grade, providing an important basis for subsequent ore processing and utilization.

[0072] Step 6: Output results

[0073] Finally, the ore identification module outputs the ore classification information and particle size classification results. These results can be directly used in ore mining, processing, sales and other links, improving the utilization efficiency and value of ore resources.

[0074] In order to implement the above method, a deep learning network-based ore classification and particle size classification system was constructed. The system consists of the following key components:

[0075] Cloud data analysis server: As the core of data processing and analysis, it stores a large amount of mineral data and runs complex deep learning models for high-level data analysis and decision support.

[0076] Ore identification module: It integrates the local data storage module and the ore identification device, and is responsible for the initial processing and identification of ore information, as well as the interaction with the cloud server.

[0077] Data processing module: coordinates the data flow between modules to ensure accurate transmission and processing of information.

[0078] Communication module: realize the communication connection between the system and external equipment (such as ore information collection equipment, cloud server) to ensure the real-time transmission of data.

[0079] Ore information collection equipment: responsible for collecting on-site ore information and is the key entry point for the system to obtain raw data.

[0080] Example 1: Implementation of iron ore classification and particle size classification based on deep learning network

[0081] As the basic raw material of the steel industry, accurate classification of the quality and particle size of iron ore is of great significance for improving production efficiency and optimizing resource allocation. Traditional manual classification methods are not only time-consuming and laborious, but also greatly affected by subjective factors, and it is difficult to meet the needs of modern industrial production. Therefore, it is particularly important to develop a method for iron ore classification and particle size classification based on a deep learning network. This embodiment will introduce the specific implementation process of the method in detail, including key links such as system construction, data collection, model training, classification and grading, and result evaluation.

[0082] Hardware platform: Build a hardware platform that includes high-performance computing servers, high-definition cameras, precision electronic scales, densitometers, etc. The server is used to run deep learning algorithms, the camera is used to collect image information of iron ore, and the electronic scale and densitometer are used to measure the weight and density of iron ore.

[0083] Software platform: Deploy a deep learning framework (such as TensorFlow or PyTorch) on the server and develop corresponding iron ore classification and particle size classification software. The software includes data acquisition module, preprocessing module, feature extraction module, classification and grading module and result display module.

[0084] Data collection and preprocessing

[0085] Data collection: Use high-definition cameras to capture iron ore samples of different types and sizes, and record their weight and density information. To ensure the diversity and representativeness of the data, iron ore samples from different origins and mining areas should be collected.

[0086] Data preprocessing: Perform preprocessing operations such as cropping, scaling, and normalization on the collected image data to eliminate noise and interference and improve image quality. At the same time, convert the weight and density information into numerical format for subsequent processing.

[0087] Model training and feature extraction

[0088] Model selection: According to the requirements of iron ore classification and particle size classification, select a suitable deep learning model. This embodiment uses convolutional neural network (CNN) as the classification model because it has powerful image feature extraction capabilities.

[0089] Feature extraction: The CNN model is used to extract features from the preprocessed image data. Through multi-layer convolution and pooling operations, key features such as texture, color, and shape of the iron ore are extracted. At the same time, the weight and density information are combined to form a complete feature vector.

[0090] Model training: Use the extracted feature vectors and corresponding label data (i.e., the type and particle size of the iron ore) to train the CNN model. By continuously adjusting the model parameters and optimizing the algorithm, the classification accuracy and generalization ability of the model can be improved.

[0091] Classification and grading

[0092] Classification: Input the iron ore sample to be classified into the trained CNN model, and the model will output the probability distribution of the sample belonging to each type of iron ore. According to the probability distribution and the preset threshold, the category to which the sample belongs is determined.

[0093] Classification: For the classified iron ore samples, the particle size grade of the samples is further determined based on their weight and density information combined with the preset particle size classification standard. This embodiment adopts the internationally accepted particle size classification standard to classify iron ore into four grades: coarse particles, medium particles, fine particles and ultra-fine particles.

[0094] Results evaluation and presentation

[0095] Result evaluation: In order to verify the accuracy and reliability of the method, the classification and grading results were evaluated by cross-validation and comparative experiments. By comparing with other classification methods (such as support vector machines, decision trees, etc.), it is proved that the method based on deep learning network has advantages in classification accuracy and granularity classification accuracy.

[0096] Result display: The classification and grading results are displayed to users in an intuitive way, such as through charts, reports, etc. At the same time, data export and printing functions are provided to facilitate users' subsequent processing and analysis.

[0097] The iron ore classification and particle size classification method based on deep learning network described in this embodiment has achieved remarkable results in practical applications. Through this method, the classification accuracy and particle size classification accuracy of iron ore are improved, providing strong support for the production decision-making of steel enterprises. At the same time, the method also has the advantages of high degree of automation, fast processing speed, strong adaptability, etc., which helps to improve the utilization efficiency and value of iron ore resources.

[0098] Example 2: Implementation of coal gangue classification and particle size classification system based on deep learning network

[0099] Gangue is solid waste generated during coal mining and washing. Its large accumulation not only occupies land resources, but also may pollute the environment. However, gangue also contains certain useful components, such as aluminum, silicon, etc. Through reasonable classification and utilization, resources can be recycled and reused. Therefore, it is of great significance to develop an efficient and accurate method for gangue classification and particle size classification. This embodiment will introduce the implementation process of a gangue classification and particle size classification system based on a deep learning network.

[0100] System architecture: The system adopts a distributed architecture, including data collection layer, data processing layer, deep learning layer, application layer, etc. The data collection layer is responsible for the collection and preliminary processing of coal gangue samples; the data processing layer cleans, converts and stores the collected data; the deep learning layer runs the deep learning model to realize the classification and particle size classification of coal gangue; the application layer provides user interface and result display functions.

[0101] Hardware configuration: The system is equipped with high-definition cameras, laser scanners, electronic scales and other hardware equipment to collect information such as images, three-dimensional shapes, and weights of coal gangue. At the same time, it is also equipped with high-performance computing servers to run deep learning algorithms and store large amounts of data.

[0102] Data collection and processing

[0103] Data collection: High-definition cameras are used to capture images of gangue, laser scanners are used to obtain its three-dimensional shape data, and electronic scales are used to measure its weight. To ensure data diversity and accuracy, samples of gangue from different mining areas and with different particle sizes should be collected.

[0104] Data preprocessing: Clean and transform the collected data, such as removing noise, adjusting image size, normalizing, etc. At the same time, convert the 3D shape data into 2D images or feature vectors for subsequent processing.

[0105] Deep learning model construction and training

[0106] Model selection: According to the requirements of gangue classification and particle size classification, select a suitable deep learning model. This embodiment adopts a model combining convolutional neural network (CNN) and recurrent neural network (RNN) to make full use of the characteristics of image and sequence data.

[0107] Feature extraction: The CNN model is used to extract features from the preprocessed image data, and key features such as texture and color of the gangue are extracted. At the same time, the RNN model is used to extract features from the three-dimensional shape data, and features such as shape and volume of the gangue are extracted.

[0108] Model training: The extracted feature vectors and corresponding label data (i.e. the type and particle size of coal gangue) are input into the deep learning model for training. The classification accuracy and particle size classification precision of the model are improved by continuously adjusting the model parameters and optimizing the algorithm.

[0109] Classification and grading implementation

[0110] Classification: Input the gangue sample to be classified into the trained deep learning model, and the model will output the probability distribution of the sample belonging to each type of gangue. According to the probability distribution and the preset threshold, the category to which the sample belongs is determined. In this embodiment, the gangue is divided into several categories such as sandstone, mudstone, and carbonaceous rock.

[0111] Classification: For the classified gangue samples, the particle size grade of the samples is further determined based on their weight, volume, shape and other characteristics, combined with the preset particle size classification standard. This embodiment adopts a custom particle size classification standard to classify the gangue into four grades: large block, medium block, small block and powder.

[0112] System application and effect evaluation

[0113] System application: Deploy the trained deep learning model to the gangue classification and particle size classification system. Input the gangue sample information to be processed through the user interface, and the system can automatically output the classification and grading results. At the same time, the system also provides data query, statistics and analysis functions, which facilitates users to conduct subsequent processing and analysis of the processing results.

[0114] Effect evaluation: To verify the accuracy and reliability of the system, actual gangue samples were used for testing and evaluation. By comparing with other classification methods (such as traditional machine learning algorithms, manual classification, etc.), it is proved that the method based on deep learning network has obvious advantages in classification accuracy and particle size classification accuracy. At the same time, the system also has the advantages of fast processing speed, high degree of automation, and strong adaptability, which helps to improve the classification efficiency and utilization value of gangue.

[0115] The gangue classification and particle size classification system based on deep learning network described in this embodiment has achieved good results in practical applications. Through this method, the classification accuracy and particle size classification accuracy of gangue have been significantly improved, providing strong support for resource recovery and reuse of coal enterprises. At the same time, the system also has the advantages of automation and intelligence, which helps to improve the overall technical level and competitiveness of the coal industry.

[0116] With the continuous development and improvement of deep learning technology, the system is expected to be applied in more fields, such as coal exploration and coal quality assessment. In addition, by combining other advanced technologies (such as the Internet of Things and cloud computing), the application scope and scenarios of the system can be further expanded to contribute to the sustainable development of the coal industry. At the same time, in view of the actual needs of gangue classification and particle size classification, the system can be continuously optimized and upgraded to improve its performance and stability to better meet the actual application needs.

Claims

1. A method for ore classification and particle size grading based on a deep learning network, characterized in that: The following steps are involved: Step 1: The ore information collection terminal collects the ore information of the ore to be classified and sends it to the ore identification module. The ore identification module extracts the corresponding identification feature from the received ore information according to the preset identification feature item. The ore identification module obtains the response status according to the identification feature. If it is an identification response, it goes to step 2; if it is no response, the received ore information is discarded; Step 2: According to the ore identification information in the ore information, the ore identification module identifies the ore type. If it is a known ore type, then go to step 3; if it is an unknown ore type, then go to step 4; Step 3, obtaining the first classification information associated with the ore identification information, and comparing the ore identification features in the first classification information with the ore features in the ore identification information. If the comparison is successful, proceed to step 6; if the comparison is unsuccessful, it is determined to be a classification error, and proceed to step 4; Step 4, obtaining the second classification information associated with the ore identification information, and matching the second classification information with the classification information in the ore database of the cloud data analysis server. If a match is found, the cloud data analysis server sends the generated third classification information to the ore identification module, and then proceeds to step 5; Step 5, the ore identification module generates ore classification and particle size classification information according to the third classification information received from the cloud data analysis server and the particle size classification information associated with the ore information; Step six: the ore identification module outputs the ore classification and particle size classification results.

2. The method for ore classification and particle size classification based on deep learning network according to claim 1, characterized in that: The ore identification module extracts the corresponding identification feature from the received ore information according to the preset identification feature item, and the ore identification module obtains the response state according to the identification feature, including: The ore information of the ore to be classified includes ore texture features, color features, physical features and density features; the preset identification feature items include density range; the physical features include ore size and weight; If the ore identification module extracts the density feature within the density range from the received ore information of the ore to be classified, the ore identification module performs an identification response, otherwise, it does not respond.

3. The method for ore classification and particle size classification based on deep learning network according to claim 2 is characterized in that: The ore identification module identifies the ore type according to the ore identification information in the ore information, including: If the ore identification module matches the corresponding density feature in the local data module, it is a known ore type, otherwise it is an unknown ore.

4. The method for ore classification and particle size classification based on deep learning network according to claim 3 is characterized in that: The step of obtaining the first classification information associated with the ore identification information and comparing the ore identification feature in the first classification information with the ore feature in the ore identification information includes: The ore texture features and color features corresponding to the density features are obtained in the local data module, and compared with the ore texture features and color features of the ore to be classified. If they are consistent, the comparison passes, and the particle size of the ore is obtained according to the density features.

5. The method for ore classification and particle size classification based on deep learning network according to claim 4 is characterized in that: The step of obtaining the second classification information associated with the ore identification information and matching the second classification information with the classification information in the ore database of the cloud data analysis server includes: The second classification information is positioning information. According to the positioning information, the ore database corresponding to the positioning information is obtained in the cloud data analysis server. According to the ore information of the ore to be classified, matching is performed in the ore database corresponding to the positioning information to obtain the classification information of the ore to be classified.

6. The method for ore classification and particle size classification based on deep learning network according to claim 5, characterized in that: The cloud data analysis server sends the generated third classification information to the ore identification module, including: the third classification information includes the ore type, ore texture characteristics and color characteristics of the ore to be classified, and is sent to the local data storage module of the ore identification module for storage.

7. The method for ore classification and particle size classification based on deep learning network according to claim 6, characterized in that: The ore identification module generates ore classification and particle size classification information according to the third classification information received from the cloud data analysis server and the particle size classification information associated with the ore information, including: The ore identification module obtains the ore classification according to the third classification information sent by the received cloud data analysis server, and obtains the ore particle size classification according to the density information and physical characteristics.

8. A system for ore classification and particle size classification based on a deep learning network, applying the method for ore classification and particle size classification based on a deep learning network according to any one of claims 1 to 7, comprising: Cloud data analysis server, ore identification module, data processing module, communication module and ore information collection equipment; The ore identification module and the communication module are respectively connected to the data processing module; the cloud data analysis server and the ore information collection device are respectively connected to the communication module for communication.

Images