A method and system for identifying abnormalities in mine blasting monitoring images

Abstract

The present application proposes a method and system for identifying anomalies in mine blasting monitoring images, wherein the method includes: obtaining the rock moisture distribution topology and rock temperature distribution topology of the mine, combining the rock type distribution topology to perform hardness zoning to obtain the rock hardness distribution topology; using the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints, collecting blasting samples from the blasting location list and blasting series list to obtain a blasting sample set, performing high-frequency blast pile morphology mining, and obtaining a blast pile baseline morphology; when the blast pile monitoring morphology is inconsistent with the blast pile baseline morphology, the blast pile monitoring morphology is marked as uncontrollable and sent to the user end. The present application solves the technical problem in the prior art that it is difficult to accurately identify anomalies in mine blasting monitoring images, and achieves the technical effect of accurately identifying anomalies in mine blasting monitoring images and issuing timely warnings.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for identifying abnormalities in mine blasting monitoring images. Background Art

[0002] With the continuous expansion of mining scale, mine blasting has been widely used in the process of mineral resource mining as an efficient rock crushing method. Mine blasting monitoring data is of great significance for ensuring blasting safety and improving blasting efficiency. However, mine blasting monitoring data has many dimensions, including vibration, noise, temperature, humidity and other parameters, and there is a complex mutual influence relationship between these parameters; at the same time, the factors affecting the effect of mine blasting show strong nonlinear characteristics, such as rock type, crack distribution, hardness distribution, etc., all have a significant impact on the blasting effect. Existing technologies mostly analyze from a single dimension, such as only focusing on vibration parameters or only analyzing temperature changes, and cannot fully consider the overall state of mine blasting. Therefore, mine blasting monitoring in existing technologies cannot effectively identify abnormal conditions. Summary of the Invention

[0003] The present invention aims to solve the technical problem in the prior art that it is difficult to accurately identify anomalies in mine blasting monitoring images, and provides a method and system for identifying anomalies in mine blasting monitoring images.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In a first aspect, the present invention provides a method for identifying anomalies in mine blasting monitoring images, comprising: obtaining the rock moisture distribution topology and rock temperature distribution topology of the mine, performing hardness zoning in combination with the rock type distribution topology, and obtaining the rock hardness distribution topology; using the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints, collecting blasting samples from the blasting location list and the blasting level list to obtain a blasting sample set, performing high-frequency blast pile morphology mining, and obtaining a blast pile baseline morphology; collecting mine blast pile monitoring images through an image acquisition device, performing image feature extraction, and obtaining a blast pile monitoring morphology; when the blast pile monitoring morphology is inconsistent with the blast pile baseline morphology, marking the mine blast pile monitoring image as uncontrollable and sending it to the user end.

[0006] In the second aspect, the present invention provides a mine blasting monitoring image anomaly identification system, including: a rock property analysis module, used to obtain the rock moisture distribution topology and rock temperature distribution topology of the mine, and perform hardness zoning in combination with the rock type distribution topology to obtain the rock hardness distribution topology; a reference morphology acquisition module, used to collect blasting samples from the blasting location list and the blasting level list based on the rock type distribution topology, rock fracture distribution topology and rock hardness distribution topology as constraints, obtain a blasting sample set, perform high-frequency blast pile morphology mining, and obtain a blast pile reference morphology; a monitoring morphology acquisition module, used to collect mine blast pile monitoring images through an image acquisition device, perform image feature extraction, and obtain blast pile monitoring morphology; an image identification module, used to identify the mine blast pile monitoring image as uncontrollable and send it to the user end when the blast pile monitoring morphology is inconsistent with the blast pile reference morphology.

[0007] The beneficial effects of the present invention are:

[0008] The rock mass moisture distribution topology and rock mass temperature distribution topology of the mine are obtained, and hardness zoning is performed in combination with the rock mass type distribution topology to obtain the rock mass hardness distribution topology. A multidimensional characteristic data set of the mine rock mass is established, which provides a basic parameter framework for subsequent blasting monitoring, so that the influence of multidimensional factors such as rock mass moisture, temperature and hardness on the blasting effect can be fully considered; with the rock mass type distribution topology, rock mass crack distribution topology and rock mass hardness distribution topology as constraints, blasting samples are collected for the blasting location list and blasting series list to obtain a blasting sample set, and high-frequency blasting pile morphology mining is performed to obtain the blasting pile benchmark morphology. By comprehensively considering the multidimensional characteristics of the rock mass, a blasting sample database subject to multidimensional constraints is established, and the sample is analyzed through high-frequency blasting pile morphology mining technology. The common features in this set form the baseline morphology of blast piles, which provides a reference basis for judging whether the blasting is normal; through the image acquisition device, the mine blast pile monitoring image is collected, and the image feature extraction is performed to obtain the blast pile monitoring morphology, which realizes the automated collection and analysis of the blast pile monitoring morphology from the actual blasting site, and the efficient and accurate blast pile monitoring morphology provides support for judging whether the mine blast pile monitoring image is abnormal; when the blast pile monitoring morphology is inconsistent with the blast pile baseline morphology, the mine blast pile monitoring image is marked as uncontrollable and sent to the user end, realizing the intelligent marking and early warning of the abnormal state of the mine blast monitoring image, providing an intuitive basis and decision-making support for the safety assessment of the blasting operation, thereby realizing the automatic identification and early warning of blasting anomalies and avoiding the occurrence of safety accidents.

[0009] Through the above technical solution, intelligent identification and automatic warning of abnormal conditions in mine blasting monitoring images are realized, solving the technical problem of difficulty in accurately identifying abnormalities in mine blasting monitoring images in the existing technology, and improving the accuracy of mine blasting monitoring image analysis and the safety of blasting operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic diagram of a process for identifying abnormalities in mine blasting monitoring images provided by the present invention;

[0011] Figure 2 This is a structural schematic diagram of a mine blasting monitoring image anomaly identification system provided by the present invention.

[0012] In the accompanying drawings, the components represented by the reference numerals are as follows:

[0013] Rock mass characteristic analysis module 11, reference shape acquisition module 12, monitoring shape acquisition module 13, image identification module 14. DETAILED DESCRIPTION

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0015] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0016] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0017] Example 1, as Figure 1 As shown, an embodiment of the present invention provides a method for identifying anomalies in mine blasting monitoring images, comprising:

[0018] S100: Obtain the rock mass moisture distribution topology and rock mass temperature distribution topology of the mine, perform hardness zoning based on the rock mass type distribution topology, and obtain the rock mass hardness distribution topology.

[0019] Specifically, first, data on the mine's rock mass moisture and temperature distribution topology are acquired through sensor networks or on-site surveys. The rock mass moisture distribution topology refers to the spatial distribution of water content in the rock mass across various regions of the mine, which directly affects the rock mass's physical properties. The rock mass temperature distribution topology refers to the spatial distribution of rock mass temperature across various regions of the mine, as temperature changes can alter the rock mass's internal stress state.

[0020] The rock mass moisture and temperature distribution topologies are then combined with the known rock mass type distribution topology. By comprehensively considering the impact of these three factors on rock mass hardness, hardness zoning of the mine rock mass is achieved. The rock mass type distribution topology refers to the spatial distribution structure of different rock types (such as granite, limestone, and shale) within the mine. Different rock types have different basic hardness characteristics. This comprehensive analysis of multiple factors overcomes the hardness assessment bias caused by traditional methods that only consider rock mass type. The resulting accurate rock mass hardness distribution topology lays the foundation for subsequent anomaly identification in blasting monitoring data.

[0021] Rock hardness directly affects the blasting effect, such as the accumulation morphology and crushed stone particle size distribution after blasting. Therefore, the obtained rock hardness distribution topology provides a parameter basis for the subsequent determination of the blasting pile benchmark morphology and provides data support for identifying anomalies in blasting monitoring data.

[0022] S200: Based on the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a blasting sample set, and high-frequency blast pile morphology mining is performed to obtain a blast pile benchmark morphology.

[0023] Specifically, first, using the topology of rock mass type distribution, rock mass fracture distribution, and rock mass hardness distribution as constraints, targeted blasting sample collection is performed within a predetermined list of blasting locations and blasting series. Rock mass fracture distribution topology refers to the spatial distribution structure of discontinuities such as cracks, faults, and joints within the mining rock mass, which directly affects the efficiency of blasting energy transfer and the rock mass fragmentation after blasting. By using these three rock mass characteristic topologies as constraints, precise screening of blasting samples is achieved, ensuring the representativeness and effectiveness of the collected samples.

[0024] The blasting location list contains the coordinate information of multiple blasting points within the mine; the blasting series list contains the blasting intensity grade data corresponding to each blasting point. By collecting blasting samples based on the blasting location list and blasting series list of the current mine blasting, a blasting sample set containing various blasting conditions and results is obtained. After obtaining the blasting sample set, high-frequency blast pile morphology mining is further performed to extract the most representative blast pile characteristic parameter combination from a large number of blasting samples to form a blast pile benchmark morphology. This blast pile benchmark morphology describes the blast pile morphology characteristics under specific rock mass conditions and blasting parameters, including but not limited to key parameters such as blast pile height, width, length, and slope angle.

[0025] Through high-frequency morphological mining based on multi-dimensional constraints, the problem that traditional single-dimensional analysis cannot comprehensively consider the overall status of mine blasting is effectively solved, providing a reliable reference standard for subsequent blasting anomaly monitoring.

[0026] S300: using an image acquisition device to collect a monitoring image of a mine blast pile, performing image feature extraction, and obtaining a monitoring morphology of the blast pile;

[0027] Specifically, after establishing the baseline morphology of the blast pile, the blast pile formed after the mine blasting is monitored in real time using on-site image acquisition devices. These devices include, but are not limited to, high-definition cameras, optical sensors mounted on drones, infrared thermal imagers, and other multi-source image acquisition equipment. By deploying these devices, multi-angle and all-round monitoring of the mine blast pile is achieved, generating monitoring images of the mine blast pile.

[0028] After preprocessing, collected images of mine blast pile monitoring systems enter the image feature extraction stage. This feature extraction process begins with image segmentation, precisely separating the blast pile area from the background. Edge detection algorithms are then used to identify the geometric outline of the blast pile. Deep learning models or traditional computer vision algorithms are then used to extract key morphological features of the blast pile, including parameters such as blast pile height, width, length, slope angle, and symmetry. Texture analysis can also be used to extract physical parameters such as surface looseness and particle size distribution. The image feature extraction process employs a multi-scale analysis approach, focusing on both the overall morphological characteristics of the blast pile and capturing local variations in detail. By constructing a feature vector space, a comprehensive and accurate description of the blast pile monitoring morphology is achieved. The extracted blast pile monitoring morphology is a digital representation of the blast pile's geometric characteristics and physical state, laying the foundation for subsequent comparative analysis with a baseline blast pile morphology.

[0029] By combining image acquisition with feature extraction, the subjectivity and inefficiency of traditional manual observation methods are overcome, and objective, accurate and efficient monitoring of mine blast pile morphology is achieved, providing reliable data support for the identification of abnormalities in blasting monitoring images.

[0030] S400: When the monitoring form of the mine explosion pile is inconsistent with the reference form of the mine explosion pile, an uncontrollable mark is added to the monitoring image of the mine explosion pile and sent to the user end.

[0031] Specifically, the phased characteristics of the mining blasting process necessitate the establishment of a corresponding baseline blast pile morphology for each stage of the blasting operation. After a particular stage of blasting is completed, a sensor array or image acquisition device deployed on-site captures the resulting blast pile monitoring morphology in real time and compares it with the baseline blast pile morphology of the corresponding stage. The blast pile monitoring morphology includes the geometric characteristics and physical state parameters of the rock mass after blasting, such as blast pile height, blast pile width, blast pile length, blast pile slope angle, loose coefficient, and large block ratio.

[0032] If the monitored blast pile morphology at a particular stage is detected to be inconsistent with the baseline blast pile morphology, the system determines that the blasting at that stage is abnormal, potentially due to uncontrollable factors such as abnormal rock mass structure, improper blasting parameter settings, or uneven distribution of blasting charges. Upon identifying a blasting anomaly, an uncontrollable indicator is immediately generated for the acquired mine blast pile monitoring image and transmitted to the user in real time via a network communication module. The uncontrollable indicator can include information such as the anomaly type, degree, and location, providing mine managers with timely and accurate warnings of blasting anomalies.

[0033] By marking the mine blast pile monitoring images that do not conform to the blast pile benchmark form as uncontrollable and sending them to the user end, mine workers can effectively identify anomalies in the mine blasting process, thereby effectively avoiding disasters caused by blasting anomalies and dealing with potential risks in the blasting process, improving the safety and controllability of mine blasting operations, and achieving accurate identification of anomalies in mine blasting monitoring images.

[0034] Furthermore, taking the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a blasting sample set, including:

[0035] S210: Collecting blasting samples from the blasting location list and the blasting series list based on the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology to obtain a first-level blasting sample set, wherein the first-level blasting sample set includes a first-level blasting location label list and a first-level blasting series label list;

[0036] S220: Collect blasting samples from the first-level blasting position label list and the first-level blasting series label list based on the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology to obtain a second-level blasting sample set;

[0037] S230: until blasting samples are collected from the N-1 level blasting position label list and the N-1 level blasting series label list based on the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology, to obtain an N-level blasting sample set, where N is an integer and 5≥N≥2;

[0038] S240: Add the first-level blasting sample set, the second-level blasting sample set, and finally the N-level blasting sample set into the blasting sample set.

[0039] In an optional embodiment, a statistically significant blasting sample set is constructed through a progressive acquisition strategy to provide sufficient data support for the subsequent construction of the blast pile benchmark morphology.

[0040] First, using the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints, blasting samples are collected from a predetermined blasting location list and blasting series list to obtain a first-level blasting sample set. The first-level blasting sample set consists of multiple first-level blasting samples, each of which has a corresponding first-level blasting location label list and first-level blasting series label list. The first-level blasting location label list records the blasting location information in the first-level blasting sample, while the first-level blasting series label list records the corresponding blasting intensity level information.

[0041] Then, a progressive collection method is used, using the first-level blasting location label list and the first-level blasting series label list generated by the first-level blasting sample set as new parameters. A second round of blasting sample collection is continued, using the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints to obtain a second-level blasting sample set. By using the sample-to-sample strategy, the number and diversity of blasting samples can be effectively expanded, making the blasting sample set more statistically representative, thereby improving the accuracy and reliability of subsequent blast pile morphology analysis. This progressive collection process can flexibly adjust the collection depth according to actual needs and supports continued multi-level collection. This means that blasting samples can be collected for the N-1 level blasting location label list and the N-1 level blasting series label list to obtain an N-level blasting sample set. Where N is an integer that satisfies the condition 5 ≥ N ≥ 2. When N = 2, it means that only one round of expansion is performed on the basis of the first-level blasting sample set, forming a second-level sample structure. When the sample size requirement is large, a deeper level of collection can be selected, such as a five-level sample expansion with N = 5, to ensure that the sample size meets the significance requirements of statistical analysis. Afterwards, the first-level blasting sample set, the second-level blasting sample set, and finally the N-level blasting sample set are all added to the overall blasting sample set.

[0042] Through multi-level sample collection and integration, the problem of statistical insignificance caused by insufficient sample size in traditional blasting monitoring is effectively solved, providing reliable data guarantee for high-frequency blast pile morphology mining, thereby improving the accuracy and reliability of mine blasting monitoring.

[0043] Furthermore, taking the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a first-level blasting sample set, including:

[0044] S211: extracting the sample rock type distribution topology, sample rock fracture distribution topology, sample rock hardness distribution topology, sample blasting position list and sample blasting series list of the blasting samples to be analyzed;

[0045] S212: Counting the ratio of the number of rock mass types that are consistent at the same location in the sample rock mass type distribution topology and the rock mass type distribution topology to the number of rock mass types in the union, and setting it as the rock mass type topological similarity;

[0046] S213: Calculating the similarity of the fracture structure between the sample rock mass fracture distribution topology and the rock mass fracture distribution topology, and setting it as the rock mass fracture topology similarity;

[0047] S214: Counting the ratio of the number of hardness consistent in the same region of the sample rock mass hardness distribution topology and the rock mass hardness distribution topology to the total number of hardness partitions, and setting it as the rock mass hardness topology similarity;

[0048] S215: Based on the blasting position list and the blasting series list, combined with the sample blasting position list and the sample blasting series list, a ratio of the number of positions with the same blasting series to the number of blasting positions in a union is calculated, and the ratio is set as the blasting parameter similarity;

[0049] S216: When the rock type topological similarity is greater than or equal to the first similarity threshold, and the rock fracture topological similarity is greater than or equal to the second similarity threshold, and the rock hardness topological similarity is greater than or equal to the third similarity threshold, and the blasting parameter similarity is greater than or equal to the fourth similarity threshold, the blasting sample to be analyzed is added to the first-level blasting sample set.

[0050] In a preferred embodiment, by calculating the multi-dimensional similarity and threshold judgment between the blasting samples to be analyzed and the benchmark data, blasting sample screening based on multi-dimensional similarity is achieved, providing a basis for constructing a first-level blasting sample set and ensuring the accuracy and representativeness of blasting sample collection.

[0051] When collecting blasting samples, each blasting sample in the blasting sample database is considered as a blasting sample to be analyzed. The blasting sample is then analyzed against the current rock type distribution topology, rock fracture distribution topology, rock hardness distribution topology, blasting location list, and blasting series list to determine whether the blasting sample to be analyzed can be included in the first-level blasting sample set. First, the key feature data of the blasting sample to be analyzed is extracted, including the sample rock type distribution topology, sample rock fracture distribution topology, sample rock hardness distribution topology, sample blasting location list, and sample blasting series list. This feature data comprehensively reflects the rock mass conditions and blasting parameter information of the blasting sample to be analyzed, providing a data foundation for subsequent similarity calculations.

[0052] Then, the degree of similarity between the sample rock type distribution topology and the rock type distribution topology is calculated. Specifically, the ratio of the number of identical rock types at the same location to the number of rock type unions is used as an evaluation metric, designated as the rock type topology similarity, reflecting the degree of matching between the two topological structures in the spatial distribution of rock types. Simultaneously, the degree of similarity between the sample rock fracture distribution topology and the rock fracture distribution topology is calculated, designated as the rock fracture topology similarity. For example, when performing fracture structure similarity statistics on the sample rock fracture distribution topology and the rock fracture distribution topology, multi-dimensional characteristics such as fracture density distribution similarity, fracture orientation consistency, and matching fracture connectivity characteristics are considered. A comprehensive fracture structure similarity value is derived through a comprehensive evaluation and used as the rock fracture topology similarity. Simultaneously, the degree of similarity between the sample rock hardness distribution topology and the rock hardness distribution topology is calculated. Specifically, the ratio of the number of identical hardness distributions in the same region to the total number of hardness zones is used as an evaluation metric, designated as the rock hardness topology similarity, effectively reflecting the degree of consistency between the two topological structures in terms of rock hardness distribution characteristics. Furthermore, the degree of similarity between the blasting sample to be analyzed and the current blasting parameters is calculated based on the blasting location list, blasting series list, sample blasting location list, and sample blasting series list. Specifically, based on the blasting location list and blasting series list, as well as the sample blasting location list and sample blasting series list, the ratio of the number of locations with the same blasting series to the union of the number of blasting locations is calculated to obtain the blasting parameter similarity.

[0053] Afterwards, a multiple threshold judgment mechanism is set up. When the topological similarity of the rock type is greater than or equal to the first similarity threshold, the topological similarity of the rock fracture is greater than or equal to the second similarity threshold, the topological similarity of the rock hardness is greater than or equal to the third similarity threshold, and the similarity of the blasting parameters is greater than or equal to the fourth similarity threshold, the blasting sample to be analyzed is judged to have sufficient similarity and is added to the first-level blasting sample set. Among them, the first similarity threshold, the second similarity threshold, the third similarity threshold, and the fourth similarity threshold can be flexibly set by the mine blasting expert group according to the actual situation of the mine, taking into account both the strict requirements of blasting sample screening and the data fluctuation characteristics in the actual mining environment, to ensure the high-quality construction of the first-level blasting sample set.

[0054] Through blasting sample screening based on multi-dimensional similarity, accurate screening of blasting samples is achieved, ensuring that the samples in the first-level blasting sample set are sufficiently representative, laying a solid data foundation for subsequent anomaly identification of blasting monitoring data.

[0055] Furthermore, a burst sample set is obtained, and high-frequency burst pile morphology mining is performed to obtain the burst pile benchmark morphology, including:

[0056] S250: extracting the first-level blasting sample set, the second-level blasting sample set, and so on to the N-level blasting sample set from the blasting sample set;

[0057] S260: Perform high-frequency blast pile morphology mining on the N-level blast sample set to obtain the high-frequency morphology of the N-level blast pile;

[0058] S270: Perform high-frequency explosion pile morphology mining on the N-level explosion pile high-frequency morphology and the N-1-level explosion sample set to obtain the N-1-level explosion pile high-frequency morphology;

[0059] S280: until the high-frequency morphology of the secondary explosion pile and the primary explosion sample set are mined, and the explosion pile benchmark morphology is obtained.

[0060] In a preferred embodiment, high-frequency blast pile morphology mining is performed based on multi-level samples, the characteristic distribution of each level of blasting sample set is recursively analyzed, and the blast pile morphology model is constructed level by level, thereby extracting representative ones from the multi-level blasting sample set.

[0061] First, blasting sample subsets at each level are separated from the overall blasting sample set, resulting in the extraction of first-level, second-level, and finally N-level blasting sample sets, laying the data foundation for subsequent hierarchical analysis. Then, high-frequency blast pile morphology mining is performed on the N-level blasting sample set. High-frequency blast pile morphology mining involves statistically analyzing the distribution characteristics of blast pile morphology parameters within the blasting sample set to extract the most frequently occurring parameter combinations and form statistically significant blast pile morphology descriptions. By analyzing the N-level blasting sample set, high-frequency morphologies of the N-level blast piles are obtained, representing the most representative blast pile characteristics within the N-level blasting sample set. Subsequently, the high-frequency morphologies of the N-level blast piles are used as prior knowledge and combined with the N-1-level blasting sample set to perform high-frequency morphology mining, resulting in the N-1-level high-frequency morphologies. This stepwise fusion of the expanded samples into the base sample effectively leverages the diversity and statistical significance of the expanded samples, improving the accuracy and reliability of blast pile morphology feature extraction.

[0062] This recursive analysis process continues until the high-frequency morphology of the secondary and primary blast samples is mined, resulting in a baseline morphology. This baseline morphology synthesizes the characteristic distribution patterns of blast samples at all levels, reflecting the common characteristics of a large number of samples while also taking into account the specific characteristics of samples at different levels. This makes the baseline morphology highly representative and applicable.

[0063] Through the step-by-step fusion of high-frequency blast pile morphology mining method, the accurate blast pile benchmark morphology is obtained, which provides a reliable reference standard for subsequent blasting anomaly monitoring, thereby improving the accuracy and reliability of mine blasting monitoring.

[0064] Furthermore, high-frequency burst pile morphology mining is performed on the N-level burst sample set to obtain the high-frequency morphology of the N-level burst pile, including:

[0065] S261: Grouping the N-level blasting sample set according to the N-1-level blasting sample set to obtain multiple groups of N-level blasting samples;

[0066] S262: extracting blast pile morphological parameters from a first group of N-level blasting samples from the plurality of groups of N-level blasting samples, performing mode statistics on the same attribute parameters, and obtaining a first group of high-frequency morphologies of blast piles;

[0067] S263: until the Mth group of N-level blasting samples of the plurality of N-level blasting samples, extracting the blast pile morphological parameters and performing mode statistics of the same attribute parameters to obtain the Mth group of blast pile high-frequency morphology;

[0068] S264: Add the first group of explosion pile high frequency forms to the Mth group of explosion pile high frequency forms into the Nth level explosion pile high frequency forms.

[0069] In a preferred embodiment, by grouping and performing parameter statistics on a set of N-level blasting samples, accurate extraction of blast pile morphological features is achieved, providing a basis for constructing a blast pile benchmark morphology.

[0070] First, the N-level blasting sample set is grouped according to the feature distribution of the N-1-level blasting sample set to obtain multiple groups of N-level blasting samples. This grouping strategy based on the characteristics of the previous level samples can classify blasting samples with similar attributes into one category, improving the accuracy and pertinence of subsequent morphological mining.

[0071] Then, the first group of multiple N-level blasting samples was processed, and the blast pile morphological parameters were extracted and the mode statistics of the same attribute parameters were performed to obtain the first group of high-frequency morphologies of blast piles. Among them, the blast pile morphological parameters include but are not limited to blast pile height, blast pile width, blast pile length, blast pile slope angle, loose coefficient, settlement rate, large block rate, block uniformity, forward distance, accumulation morphological symmetry, foundation and umbrella rock, and other multidimensional characteristic quantities. The mode statistics of the same attribute parameters refers to the statistics of the most frequently occurring values of each type of parameter to form the most representative parameter combination of the group of samples. The above parameter statistical process was applied to multiple groups of N-level blasting samples in sequence, from the first group to the Mth group, to obtain the high-frequency morphology of the blast piles of each group. Through group processing, the differences in blast pile morphology under different conditions can be fully considered, avoiding the problem of feature ambiguity caused by the mixed statistics of all samples.

[0072] Afterwards, the high-frequency patterns from the first to the Mth group of burst piles were added to the N-level burst pile high-frequency patterns, forming a complete description system that encompasses high-frequency patterns under various conditions. By integrating multiple groups of high-frequency patterns, the characteristic differences between the samples in each group were preserved while constructing a unified morphological system, providing a multi-dimensional reference standard for subsequent morphological matching and anomaly identification.

[0073] Through high-frequency blast pile morphology mining based on group statistics, representative blast pile morphological features under various conditions can be accurately extracted from N-level blasting sample sets, effectively improving the accuracy and applicability of blast pile benchmark morphology, and providing reliable technical support for anomaly identification in mine blasting monitoring data.

[0074] Furthermore, high-frequency explosion pile morphology mining is performed on the N-level explosion pile high-frequency morphology and the N-1-level explosion sample set to obtain the N-1-level explosion pile high-frequency morphology, including:

[0075] S271: Based on the first group of high-frequency morphologies of explosion piles, the first associated blasting sample of the N-1 level blasting sample set is expanded with respect to the explosion pile morphology parameters, until the Mth associated blasting sample of the N-1 level blasting sample set is expanded with respect to the explosion pile morphology parameters based on the Mth group of high-frequency morphologies of explosion piles, thereby obtaining an updated result of the explosion pile morphology parameters of the N-1 level blasting samples;

[0076] S272: Grouping the N-1 level blasting sample set according to the N-2 level blasting sample set to obtain multiple groups of N-1 level blasting samples;

[0077] S273: Based on the multiple groups of N-1 level blasting samples, the explosion pile morphology parameters of the N-1 level blasting samples are updated, and high-frequency explosion pile morphology mining is performed to obtain the high-frequency morphology of the N-1 level explosion pile.

[0078] In a preferred embodiment, by combining parameter expansion with high-frequency morphology mining, the high-frequency morphology of the N-level explosive pile and the N-1-level blasting sample set are effectively integrated to achieve the transfer and fusion of morphological features between blasting samples of different levels, providing support for the construction of the explosive pile benchmark morphology.

[0079] First, based on the N-level high-frequency morphology, the N-1-level blasting sample set is expanded with its morphological parameters. This expansion involves adding the characteristic parameters extracted from the N-level high-frequency morphology as supplementary information to the morphological description of the N-1-level blasting samples, enriching their features. Specifically, the first set of high-frequency morphologies from the N-level blasting is added to the blasting sample with matching features (i.e., the first associated blasting sample) in the N-1-level blasting sample set. This continues until the M-th set of high-frequency morphological parameters is added to the corresponding M-th associated blasting sample, ultimately obtaining the updated morphological parameters for the N-1-level blasting samples.

[0080] Then, based on the characteristic distribution of the N-2 level blasting sample set, the N-1 level blasting sample set is grouped to obtain multiple groups of N-1 level blasting samples. This grouping strategy based on the characteristics of the previous level samples can classify blasting samples with similar attributes, improving the accuracy and specificity of subsequent morphological mining. Subsequently, based on multiple groups of N-1 level blasting samples, high-frequency morphological mining is performed on the updated results of the morphological parameters of the N-1 level blasting samples after parameter expansion to obtain the high-frequency morphology of the N-1 level blasting pile. Through group mining, it is possible to integrate the characteristic information of the high-frequency morphology of the N-1 level blasting pile while retaining the original characteristics of the N-1 level blasting samples, forming a more comprehensive and accurate description of the high-frequency morphology of the N-1 level blasting pile.

[0081] By combining parameter expansion with high-frequency morphological mining, the morphological feature transfer and fusion between blasting samples of different levels are effectively realized, and the effective transfer of morphological features of different levels is achieved, providing support for the construction of a complete blast pile benchmark morphology, thereby improving the accuracy and reliability of anomaly identification in mine blasting monitoring data.

[0082] Furthermore, the rock mass moisture distribution topology and rock mass temperature distribution topology of the mine are obtained, and hardness zoning is performed in combination with the rock mass type distribution topology to obtain the rock mass hardness distribution topology, including:

[0083] S110: Configuring rock mass humidity recording data, rock mass temperature recording data, and rock mass type recording data;

[0084] S120: Collecting a centralized value of a rock mass hardness record data set that satisfies the rock mass humidity record data, the rock mass temperature record data, and the rock mass type record data, and setting the centralized value as a true value of the rock mass hardness;

[0085] S130: Using the rock mass hardness true value as supervision, and the rock mass humidity record data, the rock mass temperature record data, and the rock mass type record data, to train a random forest to obtain a rock mass hardness predictor;

[0086] S140: Processing the rock mass moisture distribution topology, the rock mass temperature distribution topology, and the rock mass type distribution topology by the rock mass hardness predictor to obtain rock mass hardness distribution information;

[0087] S150: Partitioning the rock mass hardness distribution information based on a rock mass hardness threshold to obtain the rock mass hardness distribution topology.

[0088] In a preferred embodiment, by collecting multi-dimensional characteristic data of the mine rock mass and using random forest technology to build a prediction relationship, an accurate mapping from the basic characteristics of the rock mass to the hardness distribution is achieved.

[0089] First, rock mass moisture, temperature, and type data are collected. Rock mass moisture data refers to the water content of rock at various points in the mine, obtained through borehole sampling or field testing; rock mass temperature data refers to the actual rock temperature measurements at each point in the mine; and rock mass type data refers to the rock type classification information at each point in the mine, such as granite, limestone, shale, etc. Next, a rock mass hardness data set that meets a specific combination of rock mass moisture, temperature, and type data is collected, and the central value is extracted as the true rock mass hardness value. Specifically, for each specific combination of rock mass moisture, temperature, and type, multiple hardness measurements are collected, and their central tendency value (such as the mean or median) is extracted using statistical methods to eliminate the influence of measurement error. This central value is the true rock mass hardness value. Then, using the true rock mass hardness value as supervision, a random forest is trained using the rock mass moisture, temperature, and type data as features to obtain a rock mass hardness predictor. Random forest is an ensemble learning method that effectively handles nonlinear relationships in rock hardness prediction by constructing multiple decision trees and averaging their predictions. During training, cross-validation is used to optimize random forest hyperparameters, such as the number of trees and maximum depth, to improve prediction accuracy.

[0090] The trained rock hardness predictor was then applied to the rock moisture distribution topology, temperature distribution topology, and rock type distribution topology across the entire mining area to predict continuous rock hardness distribution information. This rock hardness distribution information refers to the predicted hardness values at each spatial location within the mine, forming a three-dimensional hardness distribution field. The rock hardness distribution information was then partitioned based on rock hardness thresholds to obtain a rock hardness distribution topology. These thresholds are pre-set hardness cutoffs based on engineering experience and blasting requirements. These thresholds divide the continuous hardness distribution information into several discrete hardness intervals, such as the very soft zone, soft zone, medium-hard zone, hard zone, and very hard zone. This creates a clear rock hardness zoning structure, facilitating the targeted setting of subsequent blasting parameters and anomaly monitoring.

[0091] Furthermore, the image acquisition device is used to collect the mine blast pile monitoring image, perform image feature extraction, and obtain the blast pile monitoring morphology, including:

[0092] S310: Collecting multiple sets of data, wherein any set of the multiple sets of data includes burst pile image data and labels identifying burst pile morphological parameters;

[0093] S320: Training a burst pile morphology parameter extraction model using a convolutional neural network based on the multiple sets of data;

[0094] S330: Processing the mine explosion pile monitoring image according to the explosion pile morphology parameter extraction model to obtain the explosion pile monitoring morphology.

[0095] In a preferred embodiment, a deep learning-based method is used to construct an explosion pile morphological parameter extraction model to achieve automatic mapping from explosion pile monitoring images to morphological parameters.

[0096] First, multiple sets of data are collected as the basic datasets for model training. Each set of data consists of two core components: blast pile image data and corresponding identified blast pile morphological parameters. Explosion pile image data refers to high-resolution images of the rock mass accumulation after blasting, acquired through an image acquisition device; identified blast pile morphological parameters refer to key parameters of the blast pile morphology, such as quantitative indicators such as blast pile height, width, length, slope angle, loose coefficient, and large block rate, annotated by blasting experts or automated measurement systems. By constructing a large-scale annotated dataset encompassing multiple blasting conditions and rock mass environments, i.e., multiple sets of data, a rich and diverse sample base is provided for subsequent deep learning model training.

[0097] Then, based on the multiple sets of collected data, a model for extracting burst pile morphological parameters was constructed and trained. This model utilizes a convolutional neural network (CNN) architecture, which excels at processing image data with spatial correlation and can automatically learn to extract hierarchical features from burst pile images. The model receives burst pile image data at its input and generates burst pile morphological parameter predictions at its output. During training, the network parameters are continuously optimized using a backpropagation algorithm to minimize the error between the model's output morphological parameter predictions and the labels. To improve the model's generalization and robustness, data augmentation techniques such as rotation, scaling, and brightness adjustment are employed during training, and regularization methods are applied to prevent overfitting.

[0098] The trained blast pile morphological parameter extraction model was then applied to actual mine blast pile monitoring images. The model was able to quickly and accurately extract key morphological parameters from these newly acquired mine blast pile monitoring images, forming a complete blast pile monitoring morphological description. Specifically, the mine blast pile monitoring images were input into the blast pile morphological parameter extraction model. After undergoing multi-level feature extraction and transformation, including convolutional layers, pooling layers, and fully connected layers, the final output was a comprehensive description of the blast pile monitoring morphology, including both geometric features and physical state parameters.

[0099] This deep learning-based method for extracting blast pile morphological parameters overcomes the limitations of traditional manual extraction or rule-based algorithms in complex environments, improving the accuracy and efficiency of morphological acquisition in blast pile monitoring. This method is adaptable to analyzing blast pile images under varying lighting conditions and viewing angles, exhibiting strong environmental adaptability and robust parameter extraction, laying a foundation for subsequent blast anomaly monitoring.

[0100] Example 2, as Figure 2 As shown, based on the same inventive concept as the method for identifying anomalies in mine blasting monitoring images provided in Example 1, an embodiment of the present invention further provides a system for identifying anomalies in mine blasting monitoring images, comprising:

[0101] The rock mass characteristic analysis module 11 is used to obtain the rock mass moisture distribution topology and rock mass temperature distribution topology of the mine, and to perform hardness zoning based on the rock mass type distribution topology to obtain the rock mass hardness distribution topology;

[0102] A reference morphology acquisition module 12 is configured to collect blasting samples from a blasting location list and a blasting series list based on the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology, obtain a blasting sample set, perform high-frequency blast pile morphology mining, and obtain a blast pile reference morphology.

[0103] The monitoring morphology acquisition module 13 is used to collect the mine blast pile monitoring image through the image acquisition device, perform image feature extraction, and obtain the blast pile monitoring morphology;

[0104] The image identification module 14 is configured to, when the monitoring form of the mine explosion pile is inconsistent with the reference form of the mine explosion pile, mark the mine explosion pile monitoring image as uncontrollable and send it to the user end.

[0105] Furthermore, the reference form acquisition module 12 includes the following execution steps:

[0106] Taking the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a first-level blasting sample set, wherein the first-level blasting sample set has a first-level blasting location label list and a first-level blasting series label list;

[0107] Taking the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints, blasting samples are collected from the first-level blasting position label list and the first-level blasting series label list to obtain a second-level blasting sample set;

[0108] Until the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology are constrained, blasting samples are collected for the N-1 level blasting position label list and the N-1 level blasting series label list to obtain an N-level blasting sample set, where N is an integer, 5 ≥ N ≥ 2;

[0109] The first-level blasting sample set, the second-level blasting sample set, and finally the N-level blasting sample set are added into the blasting sample set.

[0110] Furthermore, the reference form acquisition module 12 further includes the following execution steps:

[0111] Extract the sample rock type distribution topology, sample rock fracture distribution topology, sample rock hardness distribution topology, sample blasting location list and sample blasting series list of the blasting samples to be analyzed;

[0112] Counting the ratio of the number of rock mass types that are consistent at the same location in the sample rock mass type distribution topology and the rock mass type distribution topology to the number of rock mass types in the union, and setting it as the rock mass type topological similarity;

[0113] Counting the similarity of the fracture structure between the sample rock mass fracture distribution topology and the rock mass fracture distribution topology, and setting it as the rock mass fracture topology similarity;

[0114] Counting the ratio of the number of hardness consistent in the same region of the sample rock mass hardness distribution topology and the rock mass hardness distribution topology to the total number of hardness partitions, and setting it as the rock mass hardness topology similarity;

[0115] Based on the blasting position list and the blasting series list, combined with the sample blasting position list and the sample blasting series list, the ratio of the number of positions with the same blasting series to the number of blasting positions in the union is counted, and the ratio is set as the blasting parameter similarity;

[0116] When the rock type topological similarity is greater than or equal to the first similarity threshold, the rock fracture topological similarity is greater than or equal to the second similarity threshold, the rock hardness topological similarity is greater than or equal to the third similarity threshold, and the blasting parameter similarity is greater than or equal to the fourth similarity threshold, the blasting sample to be analyzed is added to the first-level blasting sample set.

[0117] Furthermore, the reference form acquisition module 12 further includes the following execution steps:

[0118] Extracting the first-level blasting sample set, the second-level blasting sample set, and finally the N-level blasting sample set from the blasting sample set;

[0119] Perform high-frequency blast pile morphology mining on the N-level blast sample set to obtain the high-frequency morphology of the N-level blast pile;

[0120] Perform high-frequency morphology mining on the N-level explosion pile high-frequency morphology and the N-1-level explosion sample set to obtain the N-1-level explosion pile high-frequency morphology;

[0121] Until the high-frequency morphology of the secondary explosion pile and the first-level blasting sample set are mined, the explosion pile benchmark morphology is obtained.

[0122] Furthermore, the reference form acquisition module 12 further includes the following execution steps:

[0123] According to the N-1 level blasting sample set, the N level blasting sample set is grouped to obtain multiple groups of N level blasting samples;

[0124] Extracting the blast pile morphological parameters of the first group of N-level blasting samples from the plurality of groups of N-level blasting samples, performing mode statistics of the same attribute parameters, and obtaining the first group of high-frequency morphologies of the blast piles;

[0125] Until the Mth group of N-level blasting samples of the multiple groups of N-level blasting samples are extracted, the blast pile morphological parameters are subjected to mode statistics of the same attribute parameters, and the high-frequency morphology of the Mth group of blast piles is obtained;

[0126] The first group of explosion pile high frequency forms up to the Mth group of explosion pile high frequency forms are added to the Nth level explosion pile high frequency forms.

[0127] Furthermore, the reference form acquisition module 12 further includes the following execution steps:

[0128] Based on the first group of high-frequency morphologies of explosion piles, the first associated blasting sample of the N-1 level blasting sample set is expanded with the explosion pile morphology parameters, until the Mth associated blasting sample of the N-1 level blasting sample set is expanded with the explosion pile morphology parameters based on the Mth group of high-frequency morphologies of explosion piles, thereby obtaining an updated result of the explosion pile morphology parameters of the N-1 level blasting samples;

[0129] According to the N-2 level blasting sample set, the N-1 level blasting sample set is grouped to obtain multiple groups of N-1 level blasting samples;

[0130] Based on the multiple groups of N-1 level blasting samples, the explosion pile morphology parameters of the N-1 level blasting samples are updated, and high-frequency explosion pile morphology mining is performed to obtain the high-frequency morphology of the N-1 level explosion pile.

[0131] Furthermore, the rock mass property analysis module 11 includes the following execution steps:

[0132] Configure rock mass humidity recording data, rock mass temperature recording data, and rock mass type recording data;

[0133] Collecting a centralized value of a rock mass hardness record data set that satisfies the rock mass humidity record data, the rock mass temperature record data, and the rock mass type record data, and setting the centralized value as a true value of the rock mass hardness;

[0134] Using the true value of the rock mass hardness as supervision, and the rock mass humidity record data, the rock mass temperature record data, and the rock mass type record data, a random forest is trained to obtain a rock mass hardness predictor;

[0135] Processing the rock mass moisture distribution topology, the rock mass temperature distribution topology, and the rock mass type distribution topology by the rock mass hardness predictor to obtain rock mass hardness distribution information;

[0136] Based on the rock mass hardness threshold, the rock mass hardness distribution information is partitioned to obtain the rock mass hardness distribution topology.

[0137] Furthermore, the monitoring morphology acquisition module includes the following execution steps:

[0138] Collecting multiple sets of data, wherein any set of the multiple sets of data includes burst pile image data and labels identifying burst pile morphological parameters;

[0139] Based on the multiple sets of data, a model for extracting the morphological parameters of the burst pile is trained by a convolutional neural network;

[0140] The mine explosion pile monitoring image is processed according to the explosion pile morphology parameter extraction model to obtain the explosion pile monitoring morphology.

[0141] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0142] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0143] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0144] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0145] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0146] Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts become known.

[0147] 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 present invention and its equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for identifying abnormalities in mine blasting monitoring images, characterized in that: include: Obtain the rock mass moisture distribution topology and rock mass temperature distribution topology of the mine, combine the rock mass type distribution topology to perform hardness zoning, and obtain the rock mass hardness distribution topology, where the distribution topology represents the spatial distribution structure; Taking the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a blasting sample set, and high-frequency blast pile morphology mining is performed to obtain a blast pile benchmark morphology, wherein the blast pile benchmark morphology represents a blast pile morphology feature description under specific rock conditions and blasting parameters; The image acquisition device is used to collect monitoring images of mine blast piles, and image features are extracted to obtain monitoring morphologies of the blast piles, wherein the monitoring morphologies of the blast piles represent a comprehensive description of geometric features and physical state parameters of the blast piles; When the monitoring form of the mine explosion pile is inconsistent with the reference form of the mine explosion pile, the monitoring image of the mine explosion pile is marked as uncontrollable and sent to the user end; The blasting sample collection is performed on the blasting location list and the blasting series list with the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints to obtain a blasting sample set, including: Extract the sample rock type distribution topology, sample rock fracture distribution topology, sample rock hardness distribution topology, sample blasting location list and sample blasting series list of the blasting samples to be analyzed; Counting the ratio of the number of rock mass types that are consistent at the same location in the sample rock mass type distribution topology and the rock mass type distribution topology to the number of rock mass types in the union, and setting it as the rock mass type topological similarity; Counting the similarity of the fracture structure between the sample rock mass fracture distribution topology and the rock mass fracture distribution topology, and setting it as the rock mass fracture topology similarity; Counting the ratio of the number of hardness consistent in the same region of the sample rock mass hardness distribution topology and the rock mass hardness distribution topology to the total number of hardness partitions, and setting it as the rock mass hardness topology similarity; Based on the blasting position list and the blasting series list, combined with the sample blasting position list and the sample blasting series list, the ratio of the number of positions with the same blasting series to the number of blasting positions in the union is counted, and the ratio is set as the blasting parameter similarity; When the rock type topological similarity is greater than or equal to the first similarity threshold, the rock fracture topological similarity is greater than or equal to the second similarity threshold, the rock hardness topological similarity is greater than or equal to the third similarity threshold, and the blasting parameter similarity is greater than or equal to the fourth similarity threshold, the blasting sample to be analyzed is added to the blasting sample set.

2. The method according to claim 1, wherein Taking the rock type distribution topology, rock fracture distribution topology and rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a blasting sample set, including: Taking the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints, blasting samples are collected from the blasting location list and the blasting series list to obtain a first-level blasting sample set, wherein the first-level blasting sample set has a first-level blasting location label list and a first-level blasting series label list; Taking the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints, blasting samples are collected from the first-level blasting position label list and the first-level blasting series label list to obtain a second-level blasting sample set; Until the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology are constrained, blasting samples are collected for the N-1 level blasting position label list and the N-1 level blasting series label list to obtain an N-level blasting sample set, where N is an integer, 5 ≥ N ≥ 2; The first-level blasting sample set, the second-level blasting sample set, and finally the N-level blasting sample set are added into the blasting sample set.

3. The method according to claim 2, wherein Obtain a set of blasting samples, perform high-frequency blast pile morphology mining, and obtain the blast pile benchmark morphology, including: Extracting the first-level blasting sample set, the second-level blasting sample set, and finally the N-level blasting sample set from the blasting sample set; Perform high-frequency blast pile morphology mining on the N-level blast sample set to obtain the high-frequency morphology of the N-level blast pile; Perform high-frequency morphology mining on the N-level explosion pile high-frequency morphology and the N-1-level explosion sample set to obtain the N-1-level explosion pile high-frequency morphology; Until the high-frequency morphology of the secondary explosion pile and the first-level blasting sample set are mined, the explosion pile benchmark morphology is obtained.

4. The method according to claim 3, wherein For the N-level blasting sample set, high-frequency blast pile morphology mining is performed to obtain the high-frequency morphology of the N-level blast pile, including: According to the N-1 level blasting sample set, the N level blasting sample set is grouped to obtain multiple groups of N level blasting samples; Extracting the blast pile morphological parameters of the first group of N-level blasting samples from the plurality of groups of N-level blasting samples, performing mode statistics of the same attribute parameters, and obtaining the first group of high-frequency morphologies of the blast piles; Until the Mth group of N-level blasting samples of the multiple groups of N-level blasting samples are extracted, the blast pile morphological parameters are subjected to mode statistics of the same attribute parameters, and the high-frequency morphology of the Mth group of blast piles is obtained; The first group of explosion pile high frequency forms up to the Mth group of explosion pile high frequency forms are added to the Nth level explosion pile high frequency forms.

5. The method according to claim 4, wherein For the N-level explosion pile high-frequency morphology and the N-1-level explosion sample set, high-frequency explosion pile morphology mining is performed to obtain the N-1-level explosion pile high-frequency morphology, including: Based on the first group of high-frequency morphologies of explosion piles, the first associated blasting sample of the N-1 level blasting sample set is expanded with the explosion pile morphology parameters, until the Mth associated blasting sample of the N-1 level blasting sample set is expanded with the explosion pile morphology parameters based on the Mth group of high-frequency morphologies of explosion piles, thereby obtaining an updated result of the explosion pile morphology parameters of the N-1 level blasting samples; According to the N-2 level blasting sample set, the N-1 level blasting sample set is grouped to obtain multiple groups of N-1 level blasting samples; Based on the multiple groups of N-1 level blasting samples, the explosion pile morphology parameters of the N-1 level blasting samples are updated, and high-frequency explosion pile morphology mining is performed to obtain the high-frequency morphology of the N-1 level explosion pile.

6. The method according to claim 1, wherein Obtain the rock mass moisture distribution topology and rock mass temperature distribution topology of the mine, combine the rock mass type distribution topology to perform hardness zoning, and obtain the rock mass hardness distribution topology, including: Configure rock mass humidity recording data, rock mass temperature recording data, and rock mass type recording data; Collecting a centralized value of a rock mass hardness record data set that satisfies the rock mass humidity record data, the rock mass temperature record data, and the rock mass type record data, and setting the centralized value as a true value of the rock mass hardness; Using the true value of the rock mass hardness as supervision, and the rock mass humidity record data, the rock mass temperature record data, and the rock mass type record data, a random forest is trained to obtain a rock mass hardness predictor; Processing the rock mass moisture distribution topology, the rock mass temperature distribution topology, and the rock mass type distribution topology by the rock mass hardness predictor to obtain rock mass hardness distribution information; Based on the rock mass hardness threshold, the rock mass hardness distribution information is partitioned to obtain the rock mass hardness distribution topology.

7. The method according to claim 1, wherein The image acquisition device is used to collect mine blast pile monitoring images, perform image feature extraction, and obtain the blast pile monitoring morphology, including: Collecting multiple sets of data, wherein any set of the multiple sets of data includes burst pile image data and labels identifying burst pile morphological parameters; Based on the multiple sets of data, a model for extracting the morphological parameters of the burst pile is trained by a convolutional neural network; The mine explosion pile monitoring image is processed according to the explosion pile morphology parameter extraction model to obtain the explosion pile monitoring morphology.

8. A mine blasting monitoring image abnormality identification system, characterized in that: For implementing the method according to any one of claims 1 to 7, the system comprises: The rock mass characteristic analysis module is used to obtain the rock mass moisture distribution topology and rock mass temperature distribution topology of the mine, and to perform hardness zoning based on the rock mass type distribution topology to obtain the rock mass hardness distribution topology. The distribution topology represents the spatial distribution structure. A reference morphology acquisition module is used to collect blasting samples from a blasting location list and a blasting series list based on the rock type distribution topology, rock fracture distribution topology, and rock hardness distribution topology, obtain a blasting sample set, perform high-frequency blast pile morphology mining, and obtain a blast pile reference morphology, wherein the blast pile reference morphology represents a description of the blast pile morphology characteristics under specific rock mass conditions and blasting parameters; A monitoring morphology acquisition module is used to collect mine blast pile monitoring images through an image acquisition device, perform image feature extraction, and obtain the blast pile monitoring morphology, wherein the blast pile monitoring morphology represents a comprehensive description of the blast pile geometric characteristics and physical state parameters; An image identification module is used to mark the mine blast pile monitoring image as uncontrollable and send it to the user end when the blast pile monitoring form is inconsistent with the blast pile reference form; The blasting sample collection is performed on the blasting location list and the blasting series list with the rock type distribution topology, the rock fracture distribution topology, and the rock hardness distribution topology as constraints to obtain a blasting sample set, including: Extract the sample rock type distribution topology, sample rock fracture distribution topology, sample rock hardness distribution topology, sample blasting location list and sample blasting series list of the blasting samples to be analyzed; Counting the ratio of the number of rock mass types that are consistent at the same location in the sample rock mass type distribution topology and the rock mass type distribution topology to the number of rock mass types in the union, and setting it as the rock mass type topological similarity; Counting the similarity of the fracture structure between the sample rock mass fracture distribution topology and the rock mass fracture distribution topology, and setting it as the rock mass fracture topology similarity; Counting the ratio of the number of hardness consistent in the same region of the sample rock mass hardness distribution topology and the rock mass hardness distribution topology to the total number of hardness partitions, and setting it as the rock mass hardness topology similarity; Based on the blasting position list and the blasting series list, combined with the sample blasting position list and the sample blasting series list, the ratio of the number of positions with the same blasting series to the number of blasting positions in the union is counted, and the ratio is set as the blasting parameter similarity; When the rock type topological similarity is greater than or equal to the first similarity threshold, the rock fracture topological similarity is greater than or equal to the second similarity threshold, the rock hardness topological similarity is greater than or equal to the third similarity threshold, and the blasting parameter similarity is greater than or equal to the fourth similarity threshold, the blasting sample to be analyzed is added to the blasting sample set.

Images