Hillside blasting half-hole rate intelligent prediction system and method based on deep learning

Abstract

The invention provides a hillside blasting half-hole rate intelligent prediction system and method based on deep learning. The method comprises the following steps: collecting geological information of a blasted hillside and blasting information of a blasting material, and constructing a plurality of field data of the blasting operation, a plurality of estimated blasting characteristics of the blasting operation are generated through training to establish a blasting decision scheme, a field model of a blasted hillside is established to analyze blasting material distribution information and blasting risk information corresponding to the blasting decision scheme, and the blasting material distribution information is optimized according to the blasting risk information. According to the method, the blasting semi-hole rate corresponding to the blasting operation is obtained and displayed, the semi-hole rate in the hillside blasting process is intelligently predicted through deep learning, the relation between the blasting condition and the semi-hole rate is accurately recognized, a real-time prediction result is provided for operators, the prediction precision and efficiency are remarkably improved, blasting parameters can be optimized, and the blasting efficiency is improved. And the safety risk is reduced, and powerful technical support is provided for slope blasting operation.

Description

Technical Field

[0001] The present invention relates to the technical field of blasting construction, and in particular to a system and method for intelligently predicting half-porosity of hillside blasting based on deep learning. Background Art

[0002] Construction blasting is an engineering technology that uses the energy generated by the explosion of explosives under controlled conditions to complete specific engineering tasks. This construction method is often used for tunnel excavation or rock crushing. Blasting plays a very important role in the excavation of key parts. Among them, the semi-porosity is an important indicator for evaluating the blasting effect. The semi-porosity is also called the blasting hole trace rate, which is the percentage of the total length of the blasting hole traces on the excavation wall to the blasting hole length. Generally, in order to ensure that the blasting results are the same as expected, it is necessary to predict the semi-porosity before blasting. However, when predicting the semi-porosity, it is often affected by many factors, such as:

[0003] Physical properties of rocks: such as rock hardness, density, degree of crack development, etc. The structure of rocks directly affects the effect of drilling and blasting performance.

[0004] Drilling parameters: hole diameter, hole spacing, hole depth, etc. will affect the effect after blasting. A larger hole diameter and appropriate hole spacing usually help increase the blasting effect, but may also affect the semi-porosity.

[0005] Selection and dosage of explosives: Different types of explosives and their dosage have a direct impact on the blasting effect. Reasonable selection of explosives and dosage configuration can improve the uniformity of blasting and affect the utilization rate of the hole.

[0006] Blasting design (such as hole diameter and hole spacing): The arrangement of drilling holes directly affects the blasting effect. Too large or too small hole spacing may lead to incomplete blasting and affect the utilization rate of the holes.

[0007] Geological conditions: Geological conditions such as cracks, joints, and rock bedding in the rock strata will affect the rock fragmentation after blasting, thereby affecting the prediction of semi-porosity.

[0008] Currently, most manufacturers make simple predictions based on historical data and engineering experience, which leads to large errors between the blasting results and the expected results. A small number of manufacturers are trying to perform scientific calculations, but the calculation errors are still large, the calculation process is long, and the cost of investment is too high, which is not suitable for the current situation.

[0009] Therefore, the present invention provides a system and method for intelligent prediction of hillside blasting semi-porosity based on deep learning. Summary of the invention

[0010] The present invention provides a system and method for intelligent prediction of semi-porosity of hillside blasting based on deep learning. Deep learning is used to intelligently predict the semi-porosity in the hillside blasting process, accurately identify the relationship between blasting conditions and semi-porosity, and provide real-time prediction results for operators, which significantly improves the prediction accuracy and efficiency, helps to optimize blasting parameters, reduce safety risks, and provide strong technical support for hillside blasting operations.

[0011] The present invention provides a deep learning-based intelligent prediction system for hillside blasting semi-porosity, comprising:

[0012] The data acquisition module is used to collect geological information of the blasted hillside and blasting information of the blasting materials, and construct several on-site data of this blasting operation;

[0013] A deep analysis module, used to train the field data, generate a number of estimated blasting features of the current blasting operation, and generate a blasting decision plan for the current blasting operation;

[0014] A prediction and analysis module, used to establish a field model of the blasted hillside, and use the field model to analyze the blasting material distribution information and blasting risk information corresponding to the blasting decision plan;

[0015] The optimization and adjustment module is used to optimize the blasting material distribution information according to the blasting risk information, obtain the blasting half-hole rate corresponding to the current blasting operation and display it respectively.

[0016] In one practicable manner,

[0017] The data acquisition module comprises:

[0018] A geological sampling unit is used to search the geographical location of the blasted hillside in the big data, determine the basic information of the blasted hillside, determine the high-probability geological information of the blasted hillside using the basic information, and perform on-site sampling on the blasted hillside using a first geological sampling method corresponding to the high-probability geological information to obtain a plurality of first sampling information;

[0019] A compensating sampling unit is used to obtain unknown information generated during the on-site sampling, establish known geological features of the blasted hillside according to the sampling information, deduce a number of associated geological features of the blasted hillside according to the known geological features, search the big data for a second geological sampling method corresponding to each of the associated geological features, perform compensating sampling on the blasted hillside using the second geological sampling method, obtain second sampling information corresponding to each of the unknown information, and establish geological information of the blasted hillside according to the first sampling information and the second sampling information;

[0020] A material collection unit, used to obtain several kinds of blasting materials contained in the blasted hillside, and respectively collect material parameters corresponding to each of the blasting materials, establish blasting information corresponding to each of the blasting materials, and respectively establish corresponding blasting marks for each of the blasting information;

[0021] The data generating unit is used to establish the geological data of the blasted hillside based on the geological information, establish the equipment data of the blasted hillside based on the blasting information, determine the equipment range of the corresponding equipment data according to the blasting mark corresponding to the blasting information, and generate a number of field data of the blasting operation.

[0022] In one practicable manner,

[0023] The in-depth analysis module comprises:

[0024] A data training unit, used for inputting each of the field data into a convolution prediction model for convolution training, obtaining a plurality of feature blocks presented by each of the field data in different convolution layers, clustering the corresponding feature blocks in each of the convolution layers, and generating a cluster set corresponding to each of the convolution layers;

[0025] A deep training unit, used for analyzing the plurality of cluster sets corresponding to each of the field data, generating data traceability information corresponding to each of the field data according to the convolution features corresponding to each of the convolution layers and the set features corresponding to each of the cluster sets, and performing error optimization on the data traceability information using an error function model to generate a plurality of data features;

[0026] A feature reorganization unit is used to obtain the blasting purpose of this blasting operation, substitute each of the data features into the blasting purpose, obtain the operation-related information corresponding to each of the field data, perform semantic analysis on the operation-related information, reorganize the data features according to the semantic analysis results, and generate a number of estimated blasting features of this blasting operation;

[0027] A plan generating unit is used to divide the estimated blasting characteristics into estimated geological characteristics and estimated material characteristics, establish a blasting plan for the blasted hillside according to the estimated geological characteristics, establish a blastable combination of the blasting materials according to the estimated material characteristics, and establish a blasting decision plan for the current blasting operation according to the blasting plan and the blastable combination.

[0028] In one practicable manner,

[0029] The scheme generating unit comprises:

[0030] The feature classification subunit is used to perform regression analysis on the estimated blasting features, obtain the correlation relationship between different estimated blasting features, obtain a plurality of correlation feature chains of the current blasting operation, extract key information from each of the correlation feature chains, and according to the extraction results, regard a plurality of estimated blasting features included in the first correlation feature chain with geological description as estimated geological features, and regard a plurality of estimated blasting features included in the second correlation feature chain with material description as estimated material features;

[0031] A plan construction subunit is used to establish a geological model of the blasted hillside according to the estimated geological characteristics, determine a number of blasting locations of the blasting plan and the blasting intensity corresponding to each blasting location in combination with the corresponding blasting purpose, locate each blasting location in the geological model and mark the intensity accordingly, run the geological model to determine the blasting sequence of the blasting operation, and generate a blasting plan for the blasted hillside;

[0032] A combination construction subunit is used to obtain the geological hardness corresponding to each blasting position in the geological model, determine the optional material combination corresponding to each blasting position in combination with the corresponding blasting strength, determine the combination integrity and combination cost corresponding to each optional material combination according to the estimated material characteristics, and select the target optional material combination with qualified combination integrity and lowest combination cost corresponding to each blasting position as the blastable combination for the blasting position;

[0033] The scheme generating subunit is used to improve the blasting plan by using the blastable combination corresponding to each blasting position, determine the operation preparation guide, blasting sequence guide and blasting safety guide of the blasting operation, and generate a blasting decision scheme for the blasting operation.

[0034] In one practicable manner,

[0035] The prediction analysis module comprises:

[0036] A model building unit is used to build a field model of the blasted hillside according to the field data, determine a number of field blasting locations according to the blasting decision plan, and locate each of the field blasting locations in the field model as a key location of the model;

[0037] A material adjustment unit is used to establish the screening of the corresponding required blasting materials for each key position of the model according to the blasting decision plan, run the field model to obtain the required blasting material influence information between the key positions of different models, adjust the parameters of the corresponding required blasting materials according to the required blasting material influence information, and obtain the blasting material distribution information of the blasted hillside;

[0038] A key supervision unit is used to reset the material of the field model according to the blasting material distribution information, control the field model after the material reset to execute the blasting decision plan, collect the blasting information corresponding to each key position of the model, randomly sample the blasting information to obtain a plurality of training information, and establish a decision tree according to the relationship between each training information and different blasting information;

[0039] The risk identification unit is used to use random forest to perform risk prediction on each decision tree to generate corresponding risk results, and to count the risk level and risk range corresponding to each risk result to generate blasting risk information of the blasting decision plan.

[0040] In one practicable manner,

[0041] The optimization and adjustment module comprises:

[0042] A risk interpretation unit, configured to perform semantic analysis on the blasting risk information, generate a plurality of blasting risk descriptions of the blasting decision plan, respectively identify risk keywords corresponding to each of the blasting risk descriptions, and generate a keyword set of the blasting decision plan;

[0043] A method construction unit is used to establish a risk elimination method set for the blasting decision plan based on searching for a plurality of risk elimination methods corresponding to each of the risk keywords and inputting them into the corresponding word positions of the keyword set, and obtaining method exclusion features and method support features between the risk elimination methods corresponding to different word positions in the risk elimination method set;

[0044] A mode reorganization unit is used to generate a plurality of non-combination modes based on the mode exclusion characteristics, eliminate the non-combination modes in the risk elimination mode set, generate a plurality of target combination modes, and screen a plurality of optimal combination modes included in the target combination modes according to the mode support characteristics;

[0045] The optimization execution unit is used to optimize the blasting material distribution information by using the corresponding optimization combination in order from high to low according to the optimization degree corresponding to each optimization combination, and obtain the corresponding preliminary optimization result. When each of the preliminary optimization results is abnormal, the blasting material distribution information is optimized by using each target combination to obtain a second optimization result, and the effective target combination with the lowest blasting risk is selected to optimize the blasting material distribution information;

[0046] The continuous analysis unit is used to simulate the optimization result corresponding to the blasting material distribution information, obtain several safe blasting positions of the current blasting operation, identify several blasting forces corresponding to each of the safe blasting positions, obtain and display the blasting half-hole rate corresponding to each of the safe blasting positions.

[0047] In one practicable manner,

[0048] Also includes:

[0049] When the preliminary optimization result is normal, the blasting material distribution information is optimized using a corresponding effective optimal combination method.

[0050] In one practicable manner,

[0051] Also includes:

[0052] The blasting supervision module is used to estimate the regional blasting range corresponding to the safe blasting position and establish the blasting danger warning area of ​​the blasted hillside.

[0053] The present invention provides a method for intelligently predicting half-porosity of hillside blasting based on deep learning, comprising:

[0054] Step 1: Collect geological information of the blasted hillside and blasting information of the blasting materials, and construct several field data of this blasting operation;

[0055] Step 2: training the field data to generate a number of estimated blasting features of the current blasting operation, and generating a blasting decision plan for the current blasting operation;

[0056] Step 3: Establish a site model of the blasted hillside, and use the site model to analyze the blasting material distribution information and blasting risk information corresponding to the blasting decision plan;

[0057] Step 4: Optimize the blasting material distribution information according to the blasting risk information, obtain the blasting half-hole rate corresponding to the current blasting operation and display them respectively.

[0058] In one practicable manner,

[0059] The step 4 comprises:

[0060] Step 41: performing semantic analysis on the blasting risk information to generate a plurality of blasting risk descriptions of the blasting decision plan, respectively identifying risk keywords corresponding to each of the blasting risk descriptions, and generating a keyword set of the blasting decision plan;

[0061] Step 42: Based on searching for a plurality of risk elimination methods corresponding to each of the risk keywords and inputting them into the corresponding word positions of the keyword set, a risk elimination method set of the blasting decision plan is established, and method exclusion features and method support features between risk elimination methods corresponding to different word positions are obtained in the risk elimination method set;

[0062] Step 43: generating a plurality of non-combination methods based on the method exclusion feature, eliminating the non-combination methods in the risk elimination method set, generating a plurality of target combination methods, and screening a plurality of optimal combination methods included in the target combination methods according to the method support feature;

[0063] Step 44: optimizing the blasting material distribution information by using the corresponding optimization combinations in descending order of the optimization degree corresponding to each optimization combination mode, and obtaining corresponding preliminary optimization results; when each of the preliminary optimization results is abnormal, optimizing the blasting material distribution information by using each of the target combination modes respectively, and obtaining a second optimization result; selecting the effective target combination mode with the lowest blasting risk to optimize the blasting material distribution information;

[0064] Step 45: Simulate the optimization result corresponding to the blasting material distribution information, obtain several safe blasting positions for the current blasting operation, identify several blasting forces corresponding to each of the safe blasting positions, obtain the blasting half-hole rate corresponding to each of the safe blasting positions and display it.

[0065] The achievable beneficial effects of the above technical solution are as follows: in order to help operators optimize blasting operations and reduce operational risks, geological information of the blasted hillside and blasting information of the blasting materials are first collected before the blasting operation is carried out, so as to construct the field data of this blasting operation, and further train the field data to determine several estimated blasting characteristics of this report operation, and further construct the corresponding blasting decision plan based on the relationship between each estimated blasting characteristic, and then determine the blasting material distribution information and blasting decision plan of the blasting decision plan by constructing a field model and using the model to simulate the plan. At this time, a blasting decision plan that can complete this blasting has been obtained. However, in order to avoid spontaneous combustion or blasting interruption caused by uneven distribution of blasting materials, the blasting risk information is used to optimize the blasting material distribution information. After the optimization is completed, the blasting half-hole rate of this blasting operation is determined. In this way, not only can the blasting operation of the blasted hillside be guided by decision-making, an effective decision-making plan can be generated, thereby determining the blasting half-hole rate suitable for this blasting operation. This can not only provide technical references for operators, but also help operators to blast according to the blasting decision plan, greatly reducing the original risks, and provide operators with accurate and effective half-hole rate data, which is convenient for operators to carry out the next step of work.

[0066] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0067] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0069] Figure 1 It is a schematic diagram of the composition of a deep learning-based intelligent prediction system for half-porosity of hillside blasting in an embodiment of the present invention;

[0070] Figure 2 The present invention is a schematic diagram of the workflow of a method for intelligently predicting half-porosity of hillside blasting based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0072] Example 1

[0073] This embodiment provides a deep learning-based intelligent prediction system for hillside blasting semi-porosity. Figure 1 As shown, including:

[0074] The data acquisition module is used to collect geological information of the blasted hillside and blasting information of the blasting materials, and construct several on-site data of this blasting operation;

[0075] A deep analysis module, used to train the field data, generate a number of estimated blasting features of the current blasting operation, and generate a blasting decision plan for the current blasting operation;

[0076] A prediction and analysis module, used to establish a field model of the blasted hillside, and use the field model to analyze the blasting material distribution information and blasting risk information corresponding to the blasting decision plan;

[0077] The optimization and adjustment module is used to optimize the blasting material distribution information according to the blasting risk information, obtain the blasting half-hole rate corresponding to the current blasting operation and display it respectively.

[0078] In this example, the geological information represents the internal structure and material composition of the blasted hillside;

[0079] In this example, blasting materials refer to materials used in blasting operations, such as explosives;

[0080] In this example, the blasting information indicates the quantity and distribution of a blasting material;

[0081] In this example, the field data represent the data present on the blasted hillside;

[0082] In this example, the estimated blasting characteristics represent the characteristics of the blasted hillside before the blasting operation is performed;

[0083] In this example, the blasting decision plan represents the plan used to carry out this blasting operation;

[0084] In this example, the scene model means a model showing the structure of each area of ​​the scene in a virtual space;

[0085] In this example, the blasting material distribution information indicates the distribution information formed after various blasting materials need to be arranged in advance when executing the blasting decision plan;

[0086] In this example, the blasting risk information indicates information about potential dangers generated when performing blasting operations;

[0087] In this example, the purpose of optimizing the blasting material distribution information using the blasting risk information is to reasonably arrange the locations of the blasting materials to avoid continuous blasting or blasting interruption.

[0088] The working principle and beneficial effects of the above technical solution are as follows: In order to help operators optimize blasting operations and reduce operational risks, geological information of the blasted hillside and blasting information of the blasting materials are collected before the blasting operation, so as to construct the field data of this blasting operation, and further train the field data to determine several estimated blasting characteristics of this report operation, and further construct the corresponding blasting decision plan based on the relationship between each estimated blasting characteristic, and then construct a field model and use the model to simulate the plan to determine the blasting material distribution information and blasting decision plan of the blasting decision plan. At this time, a method that can complete this blasting has been obtained. However, in order to avoid spontaneous combustion or blasting interruption caused by uneven distribution of blasting materials, the blasting risk information is used to optimize the blasting material distribution information. After the optimization is completed, the blasting half-hole rate of this blasting operation is determined. In this way, not only can the blasting operation of the blasted hillside be guided by decision-making, an effective decision-making plan can be generated, thereby determining the blasting half-hole rate suitable for this blasting operation. This can not only provide technical references for operators, but also help operators to blast according to the blasting decision plan, greatly reducing the original risks, and provide operators with accurate and effective half-hole rate data, which is convenient for operators to carry out the next step of work.

[0089] Example 2

[0090] On the basis of Example 1, the data acquisition module of the intelligent prediction system for hillside blasting semi-porosity based on deep learning includes:

[0091] A geological sampling unit is used to search the geographical location of the blasted hillside in the big data, determine the basic information of the blasted hillside, determine the high-probability geological information of the blasted hillside using the basic information, and perform on-site sampling on the blasted hillside using a first geological sampling method corresponding to the high-probability geological information to obtain a plurality of first sampling information;

[0092] A compensating sampling unit is used to obtain unknown information generated during the on-site sampling, establish known geological features of the blasted hillside according to the sampling information, deduce a number of associated geological features of the blasted hillside according to the known geological features, search the big data for a second geological sampling method corresponding to each of the associated geological features, perform compensating sampling on the blasted hillside using the second geological sampling method, obtain second sampling information corresponding to each of the unknown information, and establish geological information of the blasted hillside according to the first sampling information and the second sampling information;

[0093] A material collection unit, used to obtain several kinds of blasting materials contained in the blasted hillside, and respectively collect material parameters corresponding to each of the blasting materials, establish blasting information corresponding to each of the blasting materials, and respectively establish corresponding blasting marks for each of the blasting information;

[0094] The data generating unit is used to establish the geological data of the blasted hillside based on the geological information, establish the equipment data of the blasted hillside based on the blasting information, determine the equipment range of the corresponding equipment data according to the blasting mark corresponding to the blasting information, and generate a number of field data of the blasting operation.

[0095] In this example, the basic information represents the location, height, slope, hardness, etc. of the blasted hillside;

[0096] In this example, high-probability geological information refers to the geological information that may be presented by the blasted hillside, that is, the geological information with a high probability that appears on the blasted hillside under the influence of its specific basic information;

[0097] In this example, the first geological sampling method refers to the sampling method used when the blasted hillside is sampled for the first time, that is, data verifying high-probability geological information can be collected through this sampling method, and the first geological sampling methods corresponding to different high-probability geological information are not necessarily the same;

[0098] In this example, the first sampling information represents sampling information that may correspond to high-probability geological information;

[0099] In this example, unknown information refers to information whose specific geological information has not been determined after sampling and verification using the first geological sampling method;

[0100] In this example, the known geological feature represents the geological feature of the blasted hillside constructed by the first sampling information, and the known geological feature corresponds to the high-probability geological information;

[0101] In this example, the associated geological feature refers to a geological feature that is associated with a known geological feature;

[0102] In this example, the second geological sampling method refers to the method used for the second sampling of the blasted hillside, that is, data for verifying the associated geological features can be collected through this sampling method, and the second geological sampling methods corresponding to different associated geological features are not necessarily the same;

[0103] In this example, compensatory sampling refers to the process of performing a second sampling according to a second geological sampling method;

[0104] In this example, the material parameters represent the parameters presented by the blasting material, including: functional parameters, appearance parameters, position parameters, etc.;

[0105] In this example, the equipment range represents the range of a blasting material involved in the work.

[0106] The working principle and beneficial effects of the above technical solution are as follows: before carrying out the blasting operation, the geographical location of the hillside to be blasted is first ascertained, and then sampling is carried out to determine the geological information of the hillside to be blasted, and the blasting materials on the hillside to be blasted are counted to determine the equipment information on the hillside to be blasted. Based on the two types of information obtained, the on-site data around the blasting is established, and in the process of analyzing the geological information, deep learning technology is used to analyze the high-frequency geological information of the hillside to be blasted, and its related geological information is derived. In this way, not only the geological information of the hillside to be blasted can be quickly determined, but also the efficiency and accuracy of the information analysis are guaranteed, providing strong technical support for subsequent blasting work.

[0107] Example 3

[0108] On the basis of Example 1, the deep learning-based intelligent prediction system for hillside blasting semi-porosity, the deep analysis module includes:

[0109] A data training unit, used for inputting each of the field data into a convolution prediction model for convolution training, obtaining a plurality of feature blocks presented by each of the field data in different convolution layers, clustering the corresponding feature blocks in each of the convolution layers, and generating a cluster set corresponding to each of the convolution layers;

[0110] A deep training unit, used for analyzing the plurality of cluster sets corresponding to each of the field data, generating data traceability information corresponding to each of the field data according to the convolution features corresponding to each of the convolution layers and the set features corresponding to each of the cluster sets, and performing error optimization on the data traceability information using an error function model to generate a plurality of data features;

[0111] A feature reorganization unit is used to obtain the blasting purpose of this blasting operation, substitute each of the data features into the blasting purpose, obtain the operation-related information corresponding to each of the field data, perform semantic analysis on the operation-related information, reorganize the data features according to the semantic analysis results, and generate a number of estimated blasting features of this blasting operation;

[0112] A plan generating unit is used to divide the estimated blasting characteristics into estimated geological characteristics and estimated material characteristics, establish a blasting plan for the blasted hillside according to the estimated geological characteristics, establish a blastable combination of the blasting materials according to the estimated material characteristics, and establish a blasting decision plan for the current blasting operation according to the blasting plan and the blastable combination.

[0113] In this example, the convolutional prediction model refers to a model that performs prediction tasks using a convolutional neural network;

[0114] In this example, the number of convolutional layers is 3;

[0115] In this example, the feature block represents the features presented by a field data in a convolutional layer;

[0116] In this example, the cluster set represents the result of grouping feature blocks with the same features in the same convolutional layer into one category;

[0117] In this example, one convolutional layer corresponds to one convolutional feature, and the convolutional feature indicates the degree of convolution of a convolutional layer;

[0118] In this example, a cluster set corresponds to a set feature, and the set feature represents the common features of all feature blocks in a cluster set;

[0119] In this example, the data traceability information indicates the data source of a field data;

[0120] In this example, the error function model represents a model used to measure the gap between the predicted value and the actual value of the data source information;

[0121] In this example, the operation association information indicates the relationship between a field data and the current blasting operation when the current blasting operation is performed;

[0122] In this example, the estimated blasting characteristics represent the characteristics generated when the current blasting operation is performed after the data characteristics are reorganized. The estimated blasting characteristics can be divided into estimated geological characteristics and estimated material characteristics.

[0123] In this example, the estimated geological characteristics represent the characteristics of the geological changes on the blasted hillside during the blasting operation, and the estimated material characteristics represent the characteristics of the changes on the blasting materials on the blasted hillside during the blasting operation;

[0124] In this example, the exploded plan represents the plan that needs to be executed in order to achieve the explosion purpose;

[0125] In this example, the blastable combination refers to a combination of materials that need to be detonated in order to achieve the purpose of blasting.

[0126] The working principle and beneficial effects of the above technical solution: In order to achieve the purpose of blasting, it is necessary to establish an effective blasting decision-making plan. First, the field data is input into the convolution prediction model for training, and its feature blocks in different convolution layers are obtained. Then, the feature blocks in the same convolution layer are clustered to construct a cluster set. The convolution features of the convolution layer and the set features of the cluster set are further used to trace the field data. Then, the error function model is used to optimize the error of the data tracing information, and the data characteristics of the field data are determined. The operation association information between each field data and the blasting operation is further analyzed in combination with the blasting purpose. The data features are further reorganized through semantic analysis to determine the estimated blasting features of this blasting operation. Then, the estimated blasting features are classified and the corresponding blasting plan and blastable combination are constructed. Finally, the two are integrated to obtain the blasting decision plan for this blasting operation. Through the above means, not only can the blasting decision plan for the blasting operation of this board be constructed, but also the most effective blasting decision plan can be built according to the characteristics of the blasted hillside, thereby improving the intelligence of the system.

[0127] Example 4

[0128] On the basis of Example 3, the intelligent prediction system for hillside blasting semi-porosity based on deep learning, the scheme generation unit comprises:

[0129] The feature classification subunit is used to perform regression analysis on the estimated blasting features, obtain the correlation relationship between different estimated blasting features, obtain a plurality of correlation feature chains of the current blasting operation, extract key information from each of the correlation feature chains, and according to the extraction results, regard a plurality of estimated blasting features included in the first correlation feature chain with geological description as estimated geological features, and regard a plurality of estimated blasting features included in the second correlation feature chain with material description as estimated material features;

[0130] A plan construction subunit is used to establish a geological model of the blasted hillside according to the estimated geological characteristics, determine a number of blasting locations of the blasting plan and the blasting intensity corresponding to each blasting location in combination with the corresponding blasting purpose, locate each blasting location in the geological model and mark the intensity accordingly, run the geological model to determine the blasting sequence of the blasting operation, and generate a blasting plan for the blasted hillside;

[0131] A combination construction subunit is used to obtain the geological hardness corresponding to each blasting position in the geological model, determine the optional material combination corresponding to each blasting position in combination with the corresponding blasting strength, determine the combination integrity and combination cost corresponding to each optional material combination according to the estimated material characteristics, and select the target optional material combination with qualified combination integrity and lowest combination cost corresponding to each blasting position as the blastable combination for the blasting position;

[0132] The scheme generating subunit is used to improve the blasting plan by using the blastable combination corresponding to each blasting position, determine the operation preparation guide, blasting sequence guide and blasting safety guide of the blasting operation, and generate a blasting decision scheme for the blasting operation.

[0133] In this example, regression analysis represents the process of analyzing the relationship between different estimated burst characteristics;

[0134] In this example, the association relationship represents the relationship between different estimated burst features, including: logical relationship, position relationship, embedded relationship and joint relationship;

[0135] In this example, the associated feature chain represents the result of connecting the estimated blasting features with associated relationships according to logical conditions;

[0136] In this example, the first associated feature chain represents a feature chain belonging to geological information, and the second associated feature chain represents a feature chain belonging to blasting material information;

[0137] In this example, the geological model refers to a model that displays geological information of a blasted hillside in a virtual space;

[0138] In this example, the optional material combination represents the blasting materials required to blast a blasting location;

[0139] In this example, the combination completeness indicates the completeness of the optional material combination composed of the existing materials on the blasted hillside, that is, analyzing whether the existing blasting materials on the blasted hillside can meet the corresponding optional material combination;

[0140] In this example, the combination cost represents the total cost of an optional material combination;

[0141] In this example, the operation preparation guide indicates the contents that need to be prepared before the blasting operation, the blasting sequence guide indicates the blasting sequence when the blasting operation is performed, and the blasting safety guide indicates the safety operations that need to be followed when blasting at each blasting location.

[0142] The working principle and beneficial effects of the above technical scheme: In order to further improve the blasting decision-making plan and provide an effective work reference for the operator, the estimated blasting features are subjected to regression analysis to determine the correlation between them, thereby generating a corresponding correlation feature chain, and the features corresponding to each key feature chain are determined through semantic description, and then the blasting intensity of each blasting position is analyzed in combination with the corresponding blasting purpose, and the blasting sequence of this blasting operation is determined by running the geological model, and the corresponding blasting plan is generated, and further optional material combinations are constructed according to geological hardness and blasting intensity, and the appropriate blastable combination is selected through the total integrity and total cost, and then a number of guidelines related to this blasting operation are generated, and a blasting decision plan is obtained. In this way, not only the estimated blasting features can be analyzed and combined in detail, but also a blasting decision plan can be constructed by conducting a detailed analysis of each blasting position, thereby improving the effectiveness and practicality of the blasting decision plan.

[0143] Example 5

[0144] On the basis of Example 1, the intelligent prediction system for hillside blasting semi-porosity based on deep learning, the prediction and analysis module includes:

[0145] A model building unit is used to build a field model of the blasted hillside according to the field data, determine a number of field blasting locations according to the blasting decision plan, and locate each of the field blasting locations in the field model as a key location of the model;

[0146] A material adjustment unit is used to establish the screening of the corresponding required blasting materials for each key position of the model according to the blasting decision plan, run the field model to obtain the required blasting material influence information between the key positions of different models, adjust the parameters of the corresponding required blasting materials according to the required blasting material influence information, and obtain the blasting material distribution information of the blasted hillside;

[0147] A key supervision unit is used to reset the material of the field model according to the blasting material distribution information, control the field model after the material reset to execute the blasting decision plan, collect the blasting information corresponding to each key position of the model, randomly sample the blasting information to obtain a plurality of training information, and establish a decision tree according to the relationship between each training information and different blasting information;

[0148] The risk identification unit is used to use random forest to perform risk prediction on each decision tree to generate corresponding risk results, and to count the risk level and risk range corresponding to each risk result to generate blasting risk information of the blasting decision plan.

[0149] In this example, the key locations of the model correspond one-to-one with the blasting locations on site;

[0150] In this example, the required blasting material indicates the material required for blasting the key positions of the model;

[0151] In this example, the blasting material impact information indicates that when an blasting material performs an operation, the operations that the related blasting materials will perform under its influence;

[0152] In this example, parameter adjustment includes: position adjustment, material quantity adjustment, material replacement and other operations related to adjusting the required blasting materials;

[0153] In this example, the decision tree represents a cross tree used to present the relationship between a training information and different burst information;

[0154] In this example, the risk level indicates the level of danger that occurs;

[0155] In this example, the risk scope refers to the scope affected when a hazard occurs.

[0156] The working principle and beneficial effects of the above technical solution: Since blasting operations are dangerous, in order to reduce the probability of accidents, a site model is first built before the blasting operation is carried out, in which the key positions of the model are located, and the required blasting materials corresponding to each key position of the model are determined. Then, the blasting material impact information is used to adjust the parameters of the required blasting materials, and the distribution information of the blasting materials on the blasted hillside is determined, and then the site model is synchronously reset. Finally, the blasting decision plan is simulated by running the site model, and the blasting information corresponding to each key position of the model is collected. The blasting information is then sampled and trained, and random forests are used for risk estimation to determine the blasting risk information of the blasting decision plan. By using random forests for risk estimation, each blasting position can be taken into account, thereby improving the authenticity of the blasting risk information.

[0157] Example 6

[0158] On the basis of Example 1, the intelligent prediction system for hillside blasting semi-porosity based on deep learning, the optimization and adjustment module includes:

[0159] A risk interpretation unit, configured to perform semantic analysis on the blasting risk information, generate a plurality of blasting risk descriptions of the blasting decision plan, respectively identify risk keywords corresponding to each of the blasting risk descriptions, and generate a keyword set of the blasting decision plan;

[0160] A method construction unit is used to establish a risk elimination method set for the blasting decision plan based on searching for a plurality of risk elimination methods corresponding to each of the risk keywords and inputting them into the corresponding word positions of the keyword set, and obtaining method exclusion features and method support features between the risk elimination methods corresponding to different word positions in the risk elimination method set;

[0161] A mode reorganization unit is used to generate a plurality of non-combination modes based on the mode exclusion characteristics, eliminate the non-combination modes in the risk elimination mode set, generate a plurality of target combination modes, and screen a plurality of optimal combination modes included in the target combination modes according to the mode support characteristics;

[0162] The optimization execution unit is used to optimize the blasting material distribution information by using the corresponding optimization combination in order from high to low according to the optimization degree corresponding to each optimization combination, and obtain the corresponding preliminary optimization result. When each of the preliminary optimization results is abnormal, the blasting material distribution information is optimized by using each target combination to obtain a second optimization result, and the effective target combination with the lowest blasting risk is selected to optimize the blasting material distribution information;

[0163] The continuous analysis unit is used to simulate the optimization result corresponding to the blasting material distribution information, obtain several safe blasting positions of the current blasting operation, identify several blasting forces corresponding to each of the safe blasting positions, obtain and display the blasting half-hole rate corresponding to each of the safe blasting positions.

[0164] In this example, the method exclusion feature represents the contradictory content between the two risk elimination methods, and the method expenditure feature represents the identical content between the two risk elimination methods;

[0165] In this example, non-combination means two or more risk elimination methods that cannot occur at the same time;

[0166] In this example, the target combination method represents the result of arranging the risk elimination methods in the risk elimination method set;

[0167] In this example, the optimal combination mode represents the target combination mode with mode support characteristics;

[0168] In this example, the degree of optimization is related to the support characteristics of the method. The larger the support characteristics, the higher the degree of optimization;

[0169] In this example, a safe blasting location is subjected to at least one blasting force.

[0170] The working principle and beneficial effects of the above technical solution: In order to reduce the risk of this blasting work, the blasting risk information is semantically analyzed to determine the risk keywords, and then the risk elimination method corresponding to each risk keyword is searched, and the method exclusion characteristics and method support characteristics between the risk elimination methods are used to select a suitable combination method to optimize the blasting material distribution information, determine the blasting force corresponding to each safe blasting position, and determine its corresponding blasting semi-hole rate. In this way, the semi-hole rate corresponding to each safe blasting position can be calculated, providing technical references for operators to facilitate the next step of work.

[0171] Example 7

[0172] On the basis of Example 6, the deep learning-based intelligent prediction system for hillside blasting semi-porosity further includes:

[0173] When the preliminary optimization result is normal, the blasting material distribution information is optimized using a corresponding effective optimal combination method.

[0174] The working principle and beneficial effects of the above technical solution are as follows: when the preliminary optimization result is normal, the blasting material distribution information is directly optimized, thereby improving the efficiency and speed of the optimization.

[0175] Example 8

[0176] On the basis of Example 6, the deep learning-based intelligent prediction system for hillside blasting semi-porosity further includes:

[0177] The blasting supervision module is used to estimate the regional blasting range corresponding to the safe blasting position and establish the blasting danger warning area of ​​the blasted hillside.

[0178] The working principle and beneficial effects of the above technical solution: In order to further ensure the safety of blasting work, the blasting range of each safe blasting position is analyzed and corresponding warnings are carried out.

[0179] Example 9

[0180] This example provides an intelligent prediction method for hillside blasting semi-porosity based on deep learning. Figure 2 As shown, including:

[0181] Step 1: Collect geological information of the blasted hillside and blasting information of the blasting materials, and construct several field data of this blasting operation;

[0182] Step 2: training the field data to generate a number of estimated blasting features of the current blasting operation, and generating a blasting decision plan for the current blasting operation;

[0183] Step 3: Establish a site model of the blasted hillside, and use the site model to analyze the blasting material distribution information and blasting risk information corresponding to the blasting decision plan;

[0184] Step 4: Optimize the blasting material distribution information according to the blasting risk information, obtain the blasting half-hole rate corresponding to the current blasting operation and display them respectively.

[0185] In this example, the geological information represents the internal structure and material composition of the blasted hillside;

[0186] In this example, blasting materials refer to materials used in blasting operations, such as explosives;

[0187] In this example, the blasting information indicates the quantity and distribution of a blasting material;

[0188] In this example, the field data represent the data present on the blasted hillside;

[0189] In this example, the estimated blasting characteristics represent the characteristics of the blasted hillside before the blasting operation is performed;

[0190] In this example, the blasting decision plan represents the plan used to carry out this blasting operation;

[0191] In this example, the scene model means a model showing the structure of each area of ​​the scene in a virtual space;

[0192] In this example, the blasting material distribution information indicates the distribution information formed after various blasting materials need to be arranged in advance when executing the blasting decision plan;

[0193] In this example, the blasting risk information indicates information about potential dangers generated when performing blasting operations;

[0194] In this example, the purpose of optimizing the blasting material distribution information using the blasting risk information is to reasonably arrange the locations of the blasting materials to avoid continuous blasting or blasting interruption.

[0195] The working principle and beneficial effects of the above technical solution are as follows: In order to help operators optimize blasting operations and reduce operational risks, geological information of the blasted hillside and blasting information of the blasting materials are collected before the blasting operation, so as to construct the field data of this blasting operation, and further train the field data to determine several estimated blasting characteristics of this report operation, and further construct the corresponding blasting decision plan based on the relationship between each estimated blasting characteristic, and then construct a field model and use the model to simulate the plan to determine the blasting material distribution information and blasting decision plan of the blasting decision plan. At this time, a method that can complete this blasting has been obtained. However, in order to avoid spontaneous combustion or blasting interruption caused by uneven distribution of blasting materials, the blasting risk information is used to optimize the blasting material distribution information. After the optimization is completed, the blasting half-hole rate of this blasting operation is determined. In this way, not only can the blasting operation of the blasted hillside be guided by decision-making, an effective decision-making plan can be generated, thereby determining the blasting half-hole rate suitable for this blasting operation. This can not only provide technical references for operators, but also help operators to blast according to the blasting decision plan, greatly reducing the original risks, and provide operators with accurate and effective half-hole rate data, which is convenient for operators to carry out the next step of work.

[0196] Example 10

[0197] On the basis of Example 9, the method for intelligently predicting half-porosity of hillside blasting based on deep learning, step 4, comprises:

[0198] Step 41: performing semantic analysis on the blasting risk information to generate a plurality of blasting risk descriptions of the blasting decision plan, respectively identifying risk keywords corresponding to each of the blasting risk descriptions, and generating a keyword set of the blasting decision plan;

[0199] Step 42: Based on searching for a plurality of risk elimination methods corresponding to each of the risk keywords and inputting them into the corresponding word positions of the keyword set, a risk elimination method set of the blasting decision plan is established, and method exclusion features and method support features between risk elimination methods corresponding to different word positions are obtained in the risk elimination method set;

[0200] Step 43: generating a plurality of non-combination methods based on the method exclusion feature, eliminating the non-combination methods in the risk elimination method set, generating a plurality of target combination methods, and screening a plurality of optimal combination methods included in the target combination methods according to the method support feature;

[0201] Step 44: optimizing the blasting material distribution information by using the corresponding optimization combinations in descending order of the optimization degree corresponding to each optimization combination mode, and obtaining corresponding preliminary optimization results; when each of the preliminary optimization results is abnormal, optimizing the blasting material distribution information by using each of the target combination modes respectively, and obtaining a second optimization result; selecting the effective target combination mode with the lowest blasting risk to optimize the blasting material distribution information;

[0202] Step 45: Simulate the optimization result corresponding to the blasting material distribution information, obtain several safe blasting positions for the current blasting operation, identify several blasting forces corresponding to each of the safe blasting positions, obtain the blasting half-hole rate corresponding to each of the safe blasting positions and display it.

[0203] In this example, the method exclusion feature represents the contradictory content between the two risk elimination methods, and the method expenditure feature represents the identical content between the two risk elimination methods;

[0204] In this example, non-combination means two or more risk elimination methods that cannot occur at the same time;

[0205] In this example, the target combination method represents the result of arranging the risk elimination methods in the risk elimination method set;

[0206] In this example, the optimal combination mode represents the target combination mode with mode support characteristics;

[0207] In this example, the degree of optimization is related to the support characteristics of the method. The larger the support characteristics, the higher the degree of optimization;

[0208] In this example, a safe blasting location is subjected to at least one blasting force.

[0209] The working principle and beneficial effects of the above technical solution: In order to reduce the risk of this blasting work, the blasting risk information is semantically analyzed to determine the risk keywords, and then the risk elimination method corresponding to each risk keyword is searched, and the method exclusion characteristics and method support characteristics between the risk elimination methods are used to select a suitable combination method to optimize the blasting material distribution information, determine the blasting force corresponding to each safe blasting position, and determine its corresponding blasting semi-hole rate. In this way, the semi-hole rate corresponding to each safe blasting position can be calculated, providing technical references for operators to facilitate the next step of work.

[0210] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A deep learning-based intelligent prediction system for hillside blasting semi-hole rate, characterized in that: include: The data acquisition module is used to collect geological information of the blasted hillside and blasting information of the blasting materials, and construct several on-site data of this blasting operation; A deep analysis module, used to train the field data, generate a number of estimated blasting features of the current blasting operation, and generate a blasting decision plan for the current blasting operation; A prediction and analysis module, used to establish a field model of the blasted hillside, and use the field model to analyze the blasting material distribution information and blasting risk information corresponding to the blasting decision plan; The optimization and adjustment module is used to optimize the blasting material distribution information according to the blasting risk information, obtain the blasting half-hole rate corresponding to the current blasting operation and display it respectively.

2. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 1, characterized in that: The data acquisition module comprises: A geological sampling unit is used to search the geographical location of the blasted hillside in the big data, determine the basic information of the blasted hillside, determine the high-probability geological information of the blasted hillside using the basic information, and perform on-site sampling on the blasted hillside using a first geological sampling method corresponding to the high-probability geological information to obtain a plurality of first sampling information; A compensating sampling unit is used to obtain unknown information generated during the on-site sampling, establish known geological features of the blasted hillside according to the sampling information, deduce a number of associated geological features of the blasted hillside according to the known geological features, search the big data for a second geological sampling method corresponding to each of the associated geological features, perform compensating sampling on the blasted hillside using the second geological sampling method, obtain second sampling information corresponding to each of the unknown information, and establish geological information of the blasted hillside according to the first sampling information and the second sampling information; A material collection unit, used to obtain several kinds of blasting materials contained in the blasted hillside, and respectively collect material parameters corresponding to each of the blasting materials, establish blasting information corresponding to each of the blasting materials, and respectively establish corresponding blasting marks for each of the blasting information; The data generating unit is used to establish the geological data of the blasted hillside based on the geological information, establish the equipment data of the blasted hillside based on the blasting information, determine the equipment range of the corresponding equipment data according to the blasting mark corresponding to the blasting information, and generate a number of field data of the blasting operation.

3. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 1, characterized in that: The in-depth analysis module comprises: A data training unit, used for inputting each of the field data into a convolution prediction model for convolution training, obtaining a plurality of feature blocks presented by each of the field data in different convolution layers, clustering the corresponding feature blocks in each of the convolution layers, and generating a cluster set corresponding to each of the convolution layers; A deep training unit, used for analyzing the plurality of cluster sets corresponding to each of the field data, generating data traceability information corresponding to each of the field data according to the convolution features corresponding to each of the convolution layers and the set features corresponding to each of the cluster sets, and performing error optimization on the data traceability information using an error function model to generate a plurality of data features; A feature reorganization unit is used to obtain the blasting purpose of this blasting operation, substitute each of the data features into the blasting purpose, obtain the operation-related information corresponding to each of the field data, perform semantic analysis on the operation-related information, reorganize the data features according to the semantic analysis results, and generate a number of estimated blasting features of this blasting operation; A plan generating unit is used to divide the estimated blasting characteristics into estimated geological characteristics and estimated material characteristics, establish a blasting plan for the blasted hillside according to the estimated geological characteristics, establish a blastable combination of the blasting materials according to the estimated material characteristics, and establish a blasting decision plan for the current blasting operation according to the blasting plan and the blastable combination.

4. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 3, characterized in that: The scheme generating unit comprises: The feature classification subunit is used to perform regression analysis on the estimated blasting features, obtain the correlation relationship between different estimated blasting features, obtain a plurality of correlation feature chains of the current blasting operation, extract key information from each of the correlation feature chains, and according to the extraction results, regard a plurality of estimated blasting features included in the first correlation feature chain with geological description as estimated geological features, and regard a plurality of estimated blasting features included in the second correlation feature chain with material description as estimated material features; A plan construction subunit is used to establish a geological model of the blasted hillside according to the estimated geological characteristics, determine a number of blasting locations of the blasting plan and the blasting intensity corresponding to each blasting location in combination with the corresponding blasting purpose, locate each blasting location in the geological model and mark the intensity accordingly, run the geological model to determine the blasting sequence of the blasting operation, and generate a blasting plan for the blasted hillside; A combination construction subunit is used to obtain the geological hardness corresponding to each blasting position in the geological model, determine the optional material combination corresponding to each blasting position in combination with the corresponding blasting strength, determine the combination integrity and combination cost corresponding to each optional material combination according to the estimated material characteristics, and select the target optional material combination with qualified combination integrity and lowest combination cost corresponding to each blasting position as the blastable combination for the blasting position; The scheme generating subunit is used to improve the blasting plan by using the blastable combination corresponding to each blasting position, determine the operation preparation guide, blasting sequence guide and blasting safety guide of the blasting operation, and generate a blasting decision scheme for the blasting operation.

5. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 1, characterized in that: The prediction analysis module comprises: A model building unit is used to build a field model of the blasted hillside according to the field data, determine a number of field blasting locations according to the blasting decision plan, and locate each of the field blasting locations in the field model as a key location of the model; A material adjustment unit is used to establish the screening of the corresponding required blasting materials for each key position of the model according to the blasting decision plan, run the field model to obtain the required blasting material influence information between the key positions of different models, adjust the parameters of the corresponding required blasting materials according to the required blasting material influence information, and obtain the blasting material distribution information of the blasted hillside; A key supervision unit is used to reset the material of the field model according to the blasting material distribution information, control the field model after the material reset to execute the blasting decision plan, collect the blasting information corresponding to each key position of the model, randomly sample the blasting information to obtain a plurality of training information, and establish a decision tree according to the relationship between each training information and different blasting information; The risk identification unit is used to use random forest to perform risk prediction on each decision tree to generate corresponding risk results, and to count the risk level and risk range corresponding to each risk result to generate blasting risk information of the blasting decision plan.

6. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 1, characterized in that: The optimization and adjustment module comprises: A risk interpretation unit, configured to perform semantic analysis on the blasting risk information, generate a plurality of blasting risk descriptions of the blasting decision plan, respectively identify risk keywords corresponding to each of the blasting risk descriptions, and generate a keyword set of the blasting decision plan; A method construction unit is used to establish a risk elimination method set for the blasting decision plan based on searching for a plurality of risk elimination methods corresponding to each of the risk keywords and inputting them into the corresponding word positions of the keyword set, and obtaining method exclusion features and method support features between the risk elimination methods corresponding to different word positions in the risk elimination method set; A mode reorganization unit is used to generate a plurality of non-combination modes based on the mode exclusion characteristics, eliminate the non-combination modes in the risk elimination mode set, generate a plurality of target combination modes, and screen a plurality of optimal combination modes included in the target combination modes according to the mode support characteristics; The optimization execution unit is used to optimize the blasting material distribution information by using the corresponding optimization combination in order from high to low according to the optimization degree corresponding to each optimization combination, and obtain the corresponding preliminary optimization result. When each of the preliminary optimization results is abnormal, the blasting material distribution information is optimized by using each target combination to obtain a second optimization result, and the effective target combination with the lowest blasting risk is selected to optimize the blasting material distribution information; The continuous analysis unit is used to simulate the optimization result corresponding to the blasting material distribution information, obtain several safe blasting positions of the current blasting operation, identify several blasting forces corresponding to each of the safe blasting positions, obtain and display the blasting half-hole rate corresponding to each of the safe blasting positions.

7. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 6, characterized in that: Also includes: When the preliminary optimization result is normal, the blasting material distribution information is optimized using a corresponding effective optimal combination method.

8. The deep learning-based intelligent prediction system for hillside blasting semi-porosity as claimed in claim 6, characterized in that: Also includes: The blasting supervision module is used to estimate the regional blasting range corresponding to the safe blasting position and establish the blasting danger warning area of ​​the blasted hillside.

9. A method for intelligent prediction of hillside blasting semi-porosity based on deep learning, characterized in that: include: Step 1: Collect geological information of the blasted hillside and blasting information of the blasting materials, and construct several field data of this blasting operation; Step 2: training the field data to generate a number of estimated blasting features of the current blasting operation, and generating a blasting decision plan for the current blasting operation; Step 3: Establish a site model of the blasted hillside, and use the site model to analyze the blasting material distribution information and blasting risk information corresponding to the blasting decision plan; Step 4: Optimize the blasting material distribution information according to the blasting risk information, obtain the blasting half-hole rate corresponding to the current blasting operation and display them respectively.

10. The method for intelligent prediction of hillside blasting semi-porosity based on deep learning as claimed in claim 9, characterized in that: The step 4 comprises: Step 41: performing semantic analysis on the blasting risk information to generate a plurality of blasting risk descriptions of the blasting decision plan, respectively identifying risk keywords corresponding to each of the blasting risk descriptions, and generating a keyword set of the blasting decision plan; Step 42: Based on searching for a plurality of risk elimination methods corresponding to each of the risk keywords and inputting them into the corresponding word positions of the keyword set, a risk elimination method set of the blasting decision plan is established, and method exclusion features and method support features between risk elimination methods corresponding to different word positions are obtained in the risk elimination method set; Step 43: generating a plurality of non-combination methods based on the method exclusion feature, eliminating the non-combination methods in the risk elimination method set, generating a plurality of target combination methods, and screening a plurality of optimal combination methods included in the target combination methods according to the method support feature; Step 44: optimizing the blasting material distribution information by using the corresponding optimization combinations in descending order of the optimization degree corresponding to each optimization combination mode, and obtaining corresponding preliminary optimization results; when each of the preliminary optimization results is abnormal, optimizing the blasting material distribution information by using each of the target combination modes respectively, and obtaining a second optimization result; selecting the effective target combination mode with the lowest blasting risk to optimize the blasting material distribution information; Step 45: Simulate the optimization result corresponding to the blasting material distribution information, obtain several safe blasting positions for the current blasting operation, identify several blasting forces corresponding to each of the safe blasting positions, obtain the blasting half-hole rate corresponding to each of the safe blasting positions and display it.

Images