Automated Mining Backfilling System and Control Method Based on Machine Vision and Internet of Things

Abstract

The present invention discloses an automated mining filling system and a control method based on machine vision and the Internet of Things. The system includes a machine vision module, an Internet of Things monitoring module, a data fusion and analysis module, an autonomous optimization module, and a multi-level early warning and maintenance module. The data fusion and analysis module processes the recognition data, equipment operation parameters, and environmental parameters, generates a comparison report, and transmits it to the autonomous optimization module; the autonomous optimization module adjusts the working parameters of the mining equipment and the filling equipment based on the comparison report. The present invention performs fusion analysis on the information data obtained by the machine vision module and the Internet of Things monitoring module, then optimizes the control strategy, and adjusts the working parameters of each device, thereby reducing the dependence on manual experience, improving efficiency, and reducing resource waste and potential safety hazards.

Description

Technical Field

[0001] The present invention relates to the technical field of mining and filling, in particular to an automated mining and filling system, and also to a mining and filling control method. Background Art

[0002] Amidst increasing global demand for resources, the mining industry faces multiple challenges, including improving production efficiency, reducing operating costs, ensuring operational safety, and protecting the environment. Traditional mining operations, heavily reliant on manual judgment and control, suffer from low efficiency, frequent safety hazards, and significant resource waste. With the rapid development of machine vision and IoT technologies, their application in industrial automation is becoming increasingly widespread, opening up new ideas and solutions for the transformation and upgrading of the mining industry.

[0003] Existing technical approaches suffer from the following shortcomings: strong manual dependency, widespread data silos, delayed fault warnings, and a lack of effective basis for optimized decision-making. In traditional mining operations, the determination of ore type and quality, the adjustment of mining and filling strategies, and the adjustment of equipment operating parameters rely primarily on manual experience. This not only leads to low efficiency and difficulty in achieving precise resource utilization and efficient output, but is also susceptible to human influence, leading to resource waste and safety hazards. Secondly, multi-source data such as equipment operating data and environmental parameters in the mining process is often stored in a decentralized manner, lacking effective integration and analysis, making it difficult to generate valuable decision-making support information. Furthermore, traditional early warning systems rely primarily on post-analysis, making it difficult to provide early warning of equipment failures and environmental risks, resulting in prolonged downtime, high maintenance costs, and increased safety hazards. Summary of the Invention

[0004] This paper proposes an automated mining and filling system and control method based on machine vision and the Internet of Things. Its objectives are: 1. To reduce reliance on manual experience; 2. To achieve effective data integration and analysis to support control decisions; and 3. To provide early warning of equipment failures and environmental risks.

[0005] The technical solutions of the present invention are as follows:

[0006] An automated mining and filling system based on machine vision and the Internet of Things, including a machine vision module, an Internet of Things monitoring module, a data fusion analysis module, and an autonomous optimization module;

[0007] The machine vision module is used to identify the type, distribution and quality of the ore and transmit the identification data to the data fusion analysis module;

[0008] The IoT monitoring module is used to collect the operating parameters and environmental parameters of each device during the mining and filling process, and transmit the operating parameters and environmental parameters to the data fusion analysis module;

[0009] The data fusion analysis module processes the recognition data, equipment operation parameters, and environmental parameters, generates a comparison report, and transmits it to the autonomous optimization module;

[0010] The autonomous optimization module adjusts the working parameters of the mining equipment and filling equipment based on the comparison report.

[0011] As a further improvement of the described automated mining and filling system based on machine vision and the Internet of Things: It further includes a multi-level early warning and maintenance module;

[0012] The comparison report generated by the data fusion analysis module is also transmitted to the multi-level early warning and maintenance module; The multi-level early warning and maintenance module determines whether there is a fault risk based on the comparison report and issues a warning.

[0013] As a further improvement of the described automated mining and filling system based on machine vision and the Internet of Things: The machine vision module collects the spectral data of the mining working face through a spectrometer, performs operations such as noise removal, endmember extraction, spectral unmixing, and abundance inversion on the spectral data to obtain spectral analysis data; The spectral analysis data is the recognition data, including ore types, distribution, and quality.

[0014] As a further improvement of the described automated mining and filling system based on machine vision and the Internet of Things: The Internet of Things monitoring module uses Internet of Things sensors for data collection.

[0015] As a further improvement of the described automated mining and filling system based on machine vision and the Internet of Things: The data fusion analysis module includes a real-time database, a historical database, a reference data storage unit, and a data analysis unit;

[0016] The input end of the real-time database is respectively connected to the output ends of the machine vision module and the Internet of Things monitoring module;

[0017] The input end of the historical database is connected to the output end of the real-time database;

[0018] The input ends of the data analysis unit are respectively connected to the output ends of the real-time database, the historical database, and the reference data storage unit.

[0019] As a further improvement of the described automated mining and filling system based on machine vision and the Internet of Things:

[0020] The real-time database receives the recognition data from the machine vision module and the device operation parameters and environmental parameters from the Internet of Things monitoring module, and then performs data standardization processing on the received data according to the data sources respectively to generate a real-time data set; a real-time data set contains the values of each data item at the current time point, and the data items refer to multiple data types including recognition data, operation parameters, and environmental parameters; the real-time database transmits the obtained real-time data set to the historical database and the data analysis unit respectively;

[0021] The historical database stores the real-time data set from the real-time database as a historical data set;

[0022] The reference data storage unit stores multiple reference data sets corresponding to different working conditions; a reference data set contains the standard values of each data item; the reference data storage unit also stores the standard solutions corresponding to each reference data set one by one, and each standard solution records the recommended values of the working parameters of the mining equipment and filling equipment under the corresponding working conditions;

[0023] The data analysis unit reads all the reference data sets and makes a horizontal comparison with the real-time data set, and at the same time makes a vertical comparison of the real-time data set with the historical data set corresponding to a previous time point stored in the historical database to obtain a comparison report.

[0024] As a further improvement of the above-mentioned automated mining and filling system based on machine vision and the Internet of Things: during the comparison, the real-time data set in the real-time data set is compared with the historical data set or reference data set used as the comparison data set in the following way:

[0025] Let the sequence formed by the values of each data item in the real-time data set be , be the number of data items, and the sequence formed by the values of each data item in the comparison data set be , calculate the Pearson correlation coefficient:

[0026] ;

[0027] Among them, and are the means of the sequences and the sequence respectively;

[0028] Denote the Pearson correlation coefficient between the real-time data set and the historical data set as the historical trend degree , and denote the Pearson correlation coefficient between the real-time data set and the th reference data set as the working condition tendency degree .

[0029] As a further improvement of the automated mining filling system based on machine vision and the Internet of Things: The way to adjust the working parameters of the mining equipment and the filling equipment in the autonomous optimization module is as follows:

[0030] Regard the working condition corresponding to the maximum value of the working condition tendency as the current working condition, and then obtain the recommended values of the working parameters of the mining equipment and the filling equipment in the standard solution corresponding to this working condition;

[0031] Calculate the difference between the vector formed by the obtained recommended values of the working parameters and the vector formed by the set values of the current working parameters of the mining equipment and the filling equipment, and record it as the parameter adjustment benchmark ;

[0032] According to the historical trend degree in the comparison report calculate the parameter adjustment amplitude coefficient ;

[0033] Take each element in the vector as the adjustment amount of the corresponding working parameter and send it to all equipment for real-time parameter adjustment.

[0034] As a further improvement of the automated mining filling system based on machine vision and the Internet of Things, in the multi-level early warning maintenance module, early warning is carried out in the following way:

[0035] The threshold setting and identification unit compares all data items in the current real-time data set element with the thresholds of each data item, finds out the data items that exceed the threshold range, and determines the potential failure risk types according to these data items;

[0036] For each data item that exceeds the threshold range, according to the size of the excess beyond the threshold range and the size of the corresponding threshold range calculate the over-limit ratio , and then based on the over-limit ratio and the historical trend degree in the current comparison report calculate the early warning level of the failure risk type corresponding to this data item:

[0037] <, ;

[0038] In the formula, is the floor function;

[0039] After determining the failure risk type and the early warning level, transmit the early warning level and the early warning content as an early warning instruction to the early warning maintenance disposal unit (52) corresponding to the failure risk type;

[0040] After receiving the early warning instruction, the early warning maintenance and disposal unit (52) immediately reminds the construction personnel by at least one of voice, short message, and alarm, and guides them to take intervention measures. At the same time, the early warning maintenance and disposal unit (52) records the detailed information and early warning handling situation in each early warning instruction for subsequent analysis and improvement.

[0041] The present invention also discloses an automated mining filling control method based on machine vision and the Internet of Things. The steps include:

[0042] S1. Collect the spectral data of the mining working face through the spectrometer in the machine vision module, perform noise removal, endmember extraction, spectral unmixing, and abundance inversion on the collected spectral data to identify the types, distributions, and qualities of ores;

[0043] S2. The Internet of Things monitoring module collects the operating parameters and environmental parameters of each device during the mining and filling processes through sensors;

[0044] S3. The real-time database receives the identification data from the machine vision module and the device operating parameters and environmental parameters from the Internet of Things monitoring module, generates a real-time data set, and synchronizes it to the historical database for storage;

[0045] S4. The data analysis unit reads the real-time data set, historical data set, and reference data set in the reference data storage unit, performs horizontal and vertical comparisons using the Pearson correlation coefficient, and generates a comparison report;

[0046] S5. The autonomous optimization module adjusts the working parameters of the mining equipment and filling equipment according to the comparison report;

[0047] S6. The threshold setting and identification unit determines whether there is a risk of failure based on the comparison report and the comparison between the real-time data set and the preset early warning threshold. If a risk of failure is found, an early warning instruction is generated and transmitted to the corresponding early warning maintenance and disposal unit.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. The present invention performs fusion analysis on the information data obtained from the machine vision module and the Internet of Things monitoring module, then optimizes the control strategy, and adjusts the working parameters of each device, thereby reducing the dependence on manual experience, improving efficiency, and reducing resource waste and potential safety hazards.

[0050] 2. The data fusion analysis module realizes the effective analysis and integration of multi-source data, and generates a comparison report based on historical data, providing a scientific basis for optimization decisions.

[0051] 3. The autonomous optimization module can automatically adjust the mining and filling strategies and equipment operating parameters according to the comparison report, achieving precise utilization of resources and high-efficiency output, and reducing operating costs.

[0052] 4. The multi-level early warning and maintenance module realizes early warning of equipment failures and environmental risks by presetting early warning thresholds, and reminds construction personnel to intervene in a variety of ways in a timely manner, effectively reducing downtime and potential safety hazards.

[0053] 5. The machine vision module can accurately identify the types, distributions, and qualities of ores through spectral analysis technology, providing a reliable data basis for mining and filling, significantly improving the recognition accuracy and operation efficiency, and further reducing the dependence on manual experience. Description of the Drawings

[0054] Figure 1 It is a schematic diagram of the overall structure of the automated mining and filling system of the present invention;

[0055] Figure 2 It is a schematic diagram of the structure of the data fusion and analysis module in the present invention;

[0056] Figure 3 It is a schematic diagram of the structure of the multi-level early warning and maintenance module in the present invention;

[0057] Figure 4 It is a schematic diagram of the specific structure provided by the embodiment of the present invention;

[0058] Figure 5 It is a flowchart of the automated mining and filling control method of the present invention.

[0059] Description of the Reference Numerals:

[0060] 1. Machine vision module; 2. Internet of Things monitoring module; 3. Data fusion and analysis module; 31. Real-time database; 32. Historical database; 33. Reference data storage unit; 34. Data analysis unit; 4. Autonomous optimization module; 5. Multi-level early warning and maintenance module; 51. Threshold setting and identification unit; 52. Early warning and maintenance handling unit. Detailed Embodiment

[0061] The technical solution of the present invention will be described in detail below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0062] As Figure 1 , an automated mining and filling system based on machine vision and the Internet of Things includes a machine vision module 1, an Internet of Things monitoring module 2, a data fusion and analysis module 3, and an autonomous optimization module 4, and further includes a multi-level early warning and maintenance module 5.

[0063] The machine vision module 1 is used to identify the types, distributions, and qualities of ores and transmit the identified data to the data fusion and analysis module 3.

[0064] In this embodiment, the machine vision module 1 collects spectral data of the mining working face through a spectrometer, and performs operations such as noise removal, endmember extraction, spectral unmixing, and abundance inversion on the spectral data to obtain spectral analysis data. The spectral analysis data is the above-mentioned identified data, including the types, distributions, and qualities of ores.

[0065] Specifically, first, the spectrometer is used to collect spectral data of the mining working face. The spectrometer can measure the wavelength distribution of the light reflected or emitted by the object surface, and these spectral information contains key information such as the chemical composition and physical state of the ore.

[0066] Since the collection process may be interfered by various noises including ambient light and instrument noise, it is necessary to perform noise removal processing on the collected spectral data, which can be achieved through a low-pass filtering algorithm to reduce the influence of noise on subsequent analysis. The low-pass filtering algorithm specifically includes the following steps:

[0067] The low-pass filter allows low-frequency signals to pass through while attenuating high-frequency signals. For a one-dimensional signal , the output of the low-pass filter can be calculated by the following convolution formula:

[0068] ;

[0069] where is the impulse response of the filter, which determines the characteristics of the filter. For a moving average filter, the impulse response is of equal weight, specifically:

[0070] , for ;

[0071] where is the length or window size of the filter.

[0072] An endmember refers to the basic component unit in the spectral data, usually corresponding to different mineral components in the ore. The purpose of endmember extraction is to separate these basic component units from the complex spectral data to provide a basis for subsequent spectral unmixing. The VCA algorithm can be used for endmember extraction. The VCA algorithm specifically includes the following steps:

[0073] Objective: Find a set of endmember spectra , where is 's vector ( is the number of bands), such that the mixed spectral data can be approximately expressed as , where is the abundance matrix.

[0074] Constraint: Each column of the abundance matrix (corresponding to the abundance of a mixed pixel) is non - negative and the sum is 1 (i.e., , where is a column of and

[0075]

[0076] Method: The VCA algorithm approximates the endmembers by iteratively finding the vertices of the convex polyhedron.In the mining working face, the ore is often not composed of a single mineral, but a mixture of multiple minerals. The purpose of spectral unmixing is to decompose the mixed spectrum into the spectra of each endmember, so as to obtain the abundance of each endmember in the mixture. This process can be realized by the N - FINDR algorithm. The N - FINDR algorithm specifically includes the following steps:

[0077] Initialization: Select any spectral vector in the dataset as the first endmember , and initialize the endmember set ;

[0078] Iterative process (for , where is the number of endmembers to be extracted):

[0079] Projection calculation: For each spectral vector ( , where is the total number of spectral vectors) in the dataset, calculate its projection length on the subspace spanned by the current endmember set . This is achieved by calculating the projection on the orthogonal complement space of the subspace formed by the column vectors of and . Use an approximate method, that is, calculate the residuals of and each endmember in , and select the direction with the smallest residual as the projection direction, that is, find the spectral vector that maximizes in the subspace spanned by the column vectors of .

[0080] Select a new endmember: Select the spectral vector with the maximum projection length as the new endmember , and add it to the endmember set: .

[0081] Termination condition: When the number of extracted endmembers reaches the preset When this occurs, the algorithm terminates.

[0082] After obtaining the relative abundances of each endmember through spectral unmixing, it is necessary to further perform abundance inversion to obtain the specific distribution and quality information of various minerals in the ore. The results of abundance inversion are usually presented in the form of images or data tables, showing the detailed information of ore types, distribution, and quality. It can be achieved using the least squares algorithm, which includes the following steps:

[0083] When the endmember matrix is known, the abundances are solved by minimizing the sum of the squared errors between the observed spectra and the linear combination of endmember spectra, which can be expressed as solving the following optimization problem:

[0084] ;

[0085] where, is the observed mixed spectral data matrix, is the endmember matrix, is the abundance matrix to be solved, is the all - ones vector, represents the Euclidean norm. This optimization problem is subject to two constraints: the non - negativity constraint ( ), and the "sum - to - one" constraint ( ). The former ensures that the abundances are non - negative, and the latter ensures that the sum of the abundances of all endmembers in each mixed pixel is 1.

[0086] The spectral analysis data (including ore types, distribution, and quality) obtained through the above - mentioned processing will be transmitted to the data fusion and analysis module 3 for further processing and analysis. These data are important bases for subsequent optimization of mining and filling strategies, prediction of equipment failures, and environmental accidents.

[0087] The Internet of Things monitoring module 2 is used to collect the operating parameters and environmental parameters of each device during the mining and filling processes, and transmit the operating parameters and environmental parameters to the data fusion and analysis module 3. Specifically, the Internet of Things monitoring module 2 uses Internet of Things sensors to collect data.

[0088] Specifically, the Internet of Things monitoring module 2 relies on various types of sensors to collect data. These sensors are usually installed in different parts of mining equipment and key positions in the mining area. Sensor types include but are not limited to:

[0089] Temperature sensors: used to monitor the internal temperature of the equipment, environmental temperature, etc., to ensure that the equipment operates within an appropriate temperature range and prevent failures caused by overheating;

[0090] Pressure sensors: installed on equipment such as pumps and compressors to monitor the working pressure and prevent overpressure;

[0091] Vibration sensor: used to monitor the vibration of equipment. By analyzing vibration data, the operating status of the equipment and potential mechanical failures can be judged;

[0092] Flow sensor: installed in the filling pipeline to monitor the flow rate of the filling material and ensure the accuracy of the filling operation;

[0093] Humidity sensor: monitors the environmental humidity in the mining area, which is crucial for preventing safety issues such as dust explosions;

[0094] Gas sensor: monitors the concentration of harmful gases (such as methane, oxygen, carbon monoxide, etc.) underground in the mine to ensure the safety of personnel;

[0095] Each sensor converts the sensed physical quantity into an electrical signal according to its physical or chemical properties. These electrical signals are then processed by the Internet of Things monitoring module 2 and transmitted to the data fusion analysis module 3 via wired or wireless means. In the data fusion analysis module 3, the data is stored in the real-time database 31 for subsequent analysis and synchronized to the historical database 32 for long-term storage to facilitate historical data analysis and trend prediction.

[0096] The data fusion analysis module 3 processes the identification data, equipment operating parameters, and environmental parameters, generates a comparison report, and transmits it to the autonomous optimization module 4.

[0097] Such as Figure 2 and 4 As described above, the data fusion analysis module 3 includes a real-time database 31, a historical database 32, a reference data storage unit 33, and a data analysis unit 34.

[0098] The input end of the real-time database 31 is respectively connected to the output end of the machine vision module 1 and the output end of the Internet of Things monitoring module 2.

[0099] The input end of the historical database 32 is connected to the output end of the real-time database 31.

[0100] The input end of the data analysis unit 34 is respectively connected to the output end of the real-time database 31, the output end of the historical database 32, and the output end of the reference data storage unit 33.

[0101] The real-time database 31 receives the identification data from the machine vision module 1 and the equipment operating parameters and environmental parameters from the Internet of Things monitoring module 2, and then performs data standardization processing on the received data according to the data source to generate a real-time data set; a real-time data set contains the numerical values of each data item at the current time point, and the data item refers to multiple data types from the identification data, operating parameters, and environmental parameters; the real-time database 31 transmits the obtained real-time data set to the historical database 32 and the data analysis unit 34 respectively;

[0102] The historical database 32 stores the real-time data set from the real-time database 31 as a historical data set.

[0103] The reference data storage unit 33 stores multiple reference data sets corresponding to different working conditions; a reference data set contains the standard values of each data item; the reference data storage unit 33 also stores standard solutions corresponding one-to-one to each reference data set, and the recommended working parameter values of the mining equipment and the filling equipment under the corresponding working conditions are recorded in each standard solution.

[0104] The working parameters of the mining equipment include but are not limited to: the size of the crusher feed inlet, the gap of the crushing chamber, the rotational speed, and the crushing force. The working parameters of the filling equipment include but are not limited to: the flow rate, mass, conveying speed, density, and pressure of the filling material. The specific explanations are as follows:

[0105] Size of the crusher feed inlet: Adjust the feed inlet size according to the ore type and distribution to optimize the crushing efficiency and reduce energy consumption;

[0106] Gap of the crushing chamber: Adjust the gap of the crushing chamber according to the hardness and particle size distribution of the ore to avoid over-crushing or blockage;

[0107] Rotational speed of the crusher: Adjust the rotational speed of the crusher according to the crushing difficulty of the ore to achieve the best crushing effect;

[0108] Crushing force: Control the crushing force by adjusting the power of the crusher or the pressure of the hydraulic system to ensure that the ore is fully crushed.

[0109] Flow rate of the filling material: Adjust the flow rate of the filling material according to the mining progress and filling requirements to maintain a stable filling speed.

[0110] Quality of the filling material: Ensure that the quality of the filling material meets the requirements by adjusting the raw material ratio or the screening process;

[0111] Conveying speed: Adjust the conveying speed according to the filling distance and the performance of the filling equipment to balance the filling efficiency and energy consumption.

[0112] Density and pressure: Control the density and pressure of the filling body by adjusting the compaction equipment or vibration equipment during the filling process to improve the stability and bearing capacity of the filling body.

[0113] The data analysis unit 34 reads all the reference data sets and makes a horizontal comparison with the real-time data set, and at the same time makes a vertical comparison between the real-time data set and the historical data set corresponding to a previous time point stored in the historical database 32, and obtains a comparison report.

[0114] When making a comparison, the real-time data set is compared with the historical data set or the reference data set used as the comparison data set in the following manner:

[0115] Let the sequence formed by the numerical values of each data item in the real-time data set be , be the number of data items, and the sequence formed by the numerical values of each data item in the comparison data set be , and calculate the Pearson correlation coefficient:

[0116] ;

[0117] where and are the means of the sequences and the sequence respectively.

[0118] Denote the Pearson correlation coefficient between the real-time data set and the historical data set as the historical trend degree , and denote the Pearson correlation coefficient between the real-time data set and the ith reference data set as the working condition tendency degree .

[0119] The autonomous optimization module 4 adjusts the working parameters of the mining equipment and the filling equipment based on the comparison report.

[0120] The way to adjust the working parameters of the mining equipment and the filling equipment in the autonomous optimization module 4 is as follows:

[0121] Regard the working condition corresponding to the maximum value of the working condition tendency degree as the current working condition, and then obtain the recommended values of the working parameters of the mining equipment and the filling equipment in the standard solution corresponding to this working condition.

[0122] Calculate the difference between the vector formed by the obtained recommended values of the working parameters and the vector formed by the set values of the current working parameters of the mining equipment and the filling equipment, and denote it as the parameter adjustment benchmark .

[0123] According to the historical trend degree in the comparison report, calculate the parameter adjustment amplitude coefficient .

[0124] Take each element in the vector as the adjustment amount of the corresponding working parameter and send it to all equipment for real-time adjustment of the parameters.

[0125] Adopting the above method, when the historical trend degree is low (that is, the current working condition is quite different from the historical situation), the parameter adjustment amplitude coefficient is large, so a large-amplitude adjustment is carried out; while when the historical trend degree is high, the parameter adjustment amplitude coefficient is small, and a small-amplitude adjustment is carried out.

[0126] The multi - level early warning maintenance module 5 determines whether there is a fault risk according to the comparison report and issues an early warning.

[0127] Specifically, as Figure 3 and 4 , in the multi - level early warning maintenance module 5, the early warning is carried out in the following way:

[0128] The threshold setting and identification unit 51 compares all data items in the current real - time data set element with the thresholds of each data item, finds out the data items that exceed the threshold range, and determines the potential fault risk types according to these data items. The thresholds include, but are not limited to, abnormal values of equipment operation parameters (such as temperature, pressure, vibration, etc.), critical values of environmental parameters (such as gas concentration, humidity, etc.), and abnormal ranges of ore identification data, etc.

[0129] For each data item that exceeds the threshold range, according to the size of the excess beyond the threshold range and the size of the corresponding threshold range calculate the over - limit ratio , and then based on the over - limit ratio and the historical trend degree in the current comparison report calculate the early warning level of the fault risk type corresponding to this data item:

[0130] ;

[0131] In the formula, is the floor function. The above formula divides the early warning into 5 levels based on the over - limit ratio and the historical trend degree.

[0132] After determining the fault risk type and the early warning level, the early warning level and the early warning content are transmitted as an early warning instruction to the early warning maintenance and disposal unit 52 corresponding to the fault risk type.

[0133] After receiving the early warning instruction, the early warning maintenance and disposal unit 52 immediately reminds the construction personnel by at least one of voice, short message, and alarm.

[0134] In addition to issuing early warning information, the early warning maintenance and disposal unit 52 also provides specific intervention operation guidance. For example, in the case of equipment fault early warning, it guides the construction personnel on how to quickly locate the fault point, take temporary remedial measures, or prepare repair tools, etc. The early warning maintenance and disposal unit 52 is also responsible for recording the detailed information of each early warning and the subsequent processing situation, forming an early warning maintenance log, and these logs are of great significance for analyzing the reasons for early warnings, optimizing early warning thresholds, and improving maintenance strategies, etc.

[0135] As Figure 5 , this embodiment also provides an automated mining filling control method based on machine vision and the Internet of Things. The steps include:

[0136] S1. The spectrometer in the machine vision module 1 collects the spectral data of the mining working face, performs noise removal, endmember extraction, spectral unmixing, and abundance inversion on the collected spectral data to identify the types, distributions, and qualities of ores.

[0137] S2. The Internet of Things monitoring module 2 collects the operating parameters and environmental parameters of each device during the mining and filling processes through sensors.

[0138] S3. The real-time database 31 receives the identification data from the machine vision module 1, the device operating parameters, and environmental parameters from the Internet of Things monitoring module 2, generates a real-time data set, and synchronizes it to the historical database 32 for storage.

[0139] S4. The data analysis unit 34 reads the real-time data set, the historical data set, and the reference data set in the reference data storage unit 33, and performs horizontal and vertical comparisons using the Pearson correlation coefficient to generate a comparison report.

[0140] S5. The autonomous optimization module 4 adjusts the working parameters of the mining equipment and the filling equipment according to the comparison report.

[0141] S6. The threshold setting and identification unit 51 determines whether there is a risk of failure based on the comparison report and the comparison between the real-time data set and the preset warning threshold. If a risk of failure is found, a warning instruction is generated and transmitted to the corresponding warning and maintenance disposal unit 52.

[0142] It should be noted that for those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. The scope of the present invention is defined by the claims rather than the above description.

Claims

1. An automated mining filling system based on machine vision and the Internet of Things, characterized in that: It includes a machine vision module (1), an Internet of Things monitoring module (2), a data fusion and analysis module (3), and an autonomous optimization module (4); The machine vision module (1) is used to identify the ore type, distribution, and quality, and transmit the identification data to the data fusion and analysis module (3); The Internet of Things monitoring module (2) is used to collect the operating parameters and environmental parameters of each device during the mining and filling processes, and transmit the operating parameters and environmental parameters to the data fusion and analysis module (3); The data fusion and analysis module (3) processes the identification data, device operating parameters, and environmental parameters, generates a comparison report, and transmits it to the autonomous optimization module (4); The autonomous optimization module (4) adjusts the working parameters of the mining equipment and filling equipment based on the comparison report; It also includes a multi-level warning and maintenance module (5); The comparison report generated by the data fusion and analysis module (3) is also transmitted to the multi-level warning and maintenance module (5); the multi-level warning and maintenance module (5) determines whether there is a fault risk based on the comparison report and issues a warning; The data fusion and analysis module (3) includes a real-time database (31), a historical database (32), a reference data storage unit (33), and a data analysis unit (34); The input end of the real-time database (31) is respectively connected to the output end of the machine vision module (1) and the output end of the Internet of Things monitoring module (2); The input end of the historical database (32) is connected to the output end of the real-time database (31); the input end of the data analysis unit (34) is respectively connected to the output end of the real-time database (31), the output end of the historical database (32), and the output end of the reference data storage unit (33); the real-time database (31) receives the identification data from the machine vision module (1) and the device operating parameters and environmental parameters from the Internet of Things monitoring module (2), and then performs data standardization processing on the received data according to the data sources respectively to generate a real-time data set; a real-time data set contains the numerical values of each data item at the current time point, and the data item refers to multiple data types from the identification data, operating parameters, and environmental parameters; the real-time database (31) transmits the obtained real-time data set to the historical database (32) and the data analysis unit (34) respectively; the historical database (32) stores the real-time data set from the real-time database (31) as a historical data set; The reference data storage unit (33) stores multiple reference data sets corresponding to different working conditions; A reference data set contains the standard numerical values of each data item; the reference data storage unit (33) also stores standard solutions corresponding to each reference data set one by one, and the recommended values of the working parameters of the mining equipment and filling equipment under the corresponding working conditions are recorded in each standard solution; The data analysis unit (34) reads all the reference data sets and makes a horizontal comparison with the real-time data set, and at the same time makes a vertical comparison of the real-time data set with the historical data set corresponding to a previous time point stored in the historical database (32) to obtain a comparison report; When making a comparison, the real-time dataset is compared with the historical dataset or reference dataset used as the comparison dataset in the following way: Let the sequence formed by the values of each data item in the real-time data set be \(X = \{x_1, x_2, \ldots, x\) n \}, where \(n\) is the number of data items, and the sequence formed by the values of each data item in the comparison data set be \(Y = \{y_1, y_2, \ldots, y\) n \}. Calculate the Pearson correlation coefficient: wherein, and are the means of sequence X and sequence Y, respectively; Denote the Pearson correlation coefficient between the real-time data set and the historical data set as the historical trend degree γ his , and denote the Pearson correlation coefficient between the real-time data set and the i-th reference data set as the working condition tendency degree The way to adjust the working parameters of the mining equipment and the filling equipment in the independent optimization module (4) is as follows: take the working condition corresponding to the maximum value of the working condition tendency degree as the current working condition, and then obtain the recommended values of the working parameters of the mining equipment and the filling equipment in the standard solution corresponding to this working condition; The difference between the vector formed by the recommended values of the obtained working parameters and the vector formed by the set values of the working parameters of the current mining equipment and filling equipment is denoted as the parameter adjustment benchmark Base; According to the historical trend degree γ in the comparison report his Calculate the parameter adjustment amplitude coefficient Each element in the vector α * Base is used as the adjustment amount of the corresponding working parameter and sent to all equipment for real-time parameter adjustment.

2. The automated mining filling system based on machine vision and the Internet of Things according to claim 1, characterized in that: The machine vision module (1) collects the spectral data of the mining working face through a spectrometer, performs noise removal, endmember extraction, spectral unmixing, and abundance inversion operations on the spectral data to obtain spectral analysis data; the spectral analysis data is the recognition data, including ore type, distribution, and quality.

3. The automated mining filling system based on machine vision and the Internet of Things according to claim 1, characterized in that: The Internet of Things monitoring module (2) uses Internet of Things sensors to collect data.

4. The automated mining filling system based on machine vision and the Internet of Things according to claim 1, characterized in that In the multi-level early warning and maintenance module (5), early warning is carried out in the following way: The threshold setting and recognition unit (51) compares all data items in the current real-time dataset element with the thresholds of each data item, finds out the data items that exceed the threshold range, and determines the potential failure risk types according to these data items; For each data item that exceeds the threshold range, calculate the over-limit ratio R = Exc / Th based on the size Exc of the excess beyond the threshold range and the size Th of the corresponding threshold range, and then based on the over-limit ratio R and the historical trend degree γ in the current comparison report his Calculate the warning level of the failure risk type corresponding to this data item: L = min(max(floor(R×10×(1 - γ his ) + 1), 1), 5); In the formula, floor(·) is the floor function; After determining the failure risk type and early warning level, the early warning level and early warning content are transmitted as early warning instructions to the early warning and maintenance disposal unit (52) corresponding to the failure risk type; After receiving the early warning instructions, the early warning and maintenance disposal unit (52) immediately reminds the construction personnel by at least one of voice, short message, and alarm, and guides the adoption of intervention measures. At the same time, the early warning and maintenance disposal unit (52) records the detailed information and early warning processing situation in each early warning instruction for subsequent analysis and improvement.

Images