Abstract:Wireless network planning and optimization for mine mobile communication,Internet of Things communication,and personnel and vehicle positioning systems all require the calculation of wireless transmission path loss.However,no statistical calculation method is currently available that is specifically designed for mine wireless transmission path loss under the special conditions of underground confined spaces.Existing general-purpose,open-space,and indoor wireless transmission path loss statistical calculation methods do not consider underground-specific environmental factors such as roadway cross-sectional area,making them difficult to apply directly to complex underground conditions.To address these problems,this study analyzed the main factors affecting mine wireless transmission path loss and proposed a mine wireless transmission path loss calculation method based on roadway cross-sectional area.The method considered both line-of-sight and non-line-of-sight transmission,and was related not only to frequency and distance but also to roadway cross-sectional area,curved roadways,branch roadways,belt conveyors,and roadway wall smoothness.Based on measured data,the characteristics of mine wireless transmission were revealed as follows:① Mine wireless transmission had a significant frequency band-pass characteristic,with a frequency turning point.When the operating frequency was lower than the frequency turning point,the roadway strongly affected wireless transmission,path loss was high,and the lower the operating frequency,the greater the influence of the roadway on wireless transmission;when the operating frequency was higher than the turning point,the influence of the roadway on wireless transmission was small,but wireless transmission path loss increased as the operating frequency increased.For straight roadways,the frequency turning point occurred around 700/800 MHz.② Wireless transmission path loss in mine roadways was greatly affected by roadway cross-sectional area.Smaller roadway cross-sections resulted in higher wireless transmission path loss,and the influence became even greater when the operating frequency was lower than the frequency turning point.③ Wireless transmission in mine curved and branch roadways was affected both by roadway cross-sectional area and by the frequency of non-line-of-sight wireless transmission,and the influence of frequency on wireless transmission in curved and branch roadways was greater than the influence of roadway cross-sectional area.Curved roadways and branch roadways increased wireless transmission path loss in mines.The proposed method was applied to calculate wireless transmission path loss in different mine scenarios.The results showed that it achieved an average absolute error of 3.8 dB,reducing the error by 7.0,3.3,2.7,5.2,2.9,3.5,and 5.1 dB compared with the FSPL,CIF,ABG,WINNER Ⅱ,ITU-R M.2412,3GPP InH-Office,and ITU-R P.1238 calculation methods,respectively.

Abstract:To address the problems of existing hydraulic intensifiers used in underground coal mines-such as large size,inability to achieve pressure boosting while drilling,and the need for auxiliary driving devices-a passive hydraulic intensifier for while-drilling operations was developed based on the piston differential pressurization principle,employing a multi-stage multi-chamber modular design and a coordinated control technique using directional valves.This intensifier did not rely on an external high-pressure pump and converted low-pressure underground mine water into high-pressure water autonomously,enabling high-pressure permeability-enhancement operations.It achieved a pressurization ratio of 1∶10 with a minimum pressurization pressure of 20 MPa.The influences of high-pressure outlet diameter,effective piston acting area ratio,and input pressure on the performance of the intensifier were analyzed:the larger the high-pressure outlet diameter,effective piston acting area ratio,and input pressure,the faster the flow velocity of the high-pressure liquid,but the longer the time required for the high-pressure liquid to reach a steady-state velocity.Field industrial test results showed that,during hydraulic punching and cavity creation in a soft coal seam with a firmness coefficient of 0.15,the average coal output per meter of borehole reached 1.5 t/m,and the intensifier exhibited stable performance,achieving pump-free high-pressure hydraulic permeability enhancement.

Abstract:In the single-bucket-truck process system of open-pit coal mines,coordinated operations between trucks and electric shovels often experience phenomena such as"trucks waiting for shovels"or"shovels waiting for trucks",leading to decreased coordination efficiency of equipment at the working face.Current research on coordination evaluation mostly focuses on single time indicators(e.g.,average waiting time)or qualitative descriptions,which struggle to comprehensively cover multidimensional demands such as cost loss,capacity utilization,operational steadiness,and supply-demand balance.Moreover,issues such as fuzzy indicator boundaries and insufficient data support make it difficult to provide precise quantitative foundations for scheduling optimization.To address the above problems,a comprehensive evaluation method for truck-shovel coordination efficiency in the single-bucket-truck process system at stripping faces was proposed.A comprehensive evaluation system comprising 16 core indicators was constructed from four main control factors:waiting cost,equipment utilization,system stability,and capacity matching degree.A structural equation model was defined to analyze the above four main controlling factors as global latent variables.The standardized paths of the evaluation system were established through model fitting,and reliability and validity analyses were further conducted to verify the practical effectiveness and reliability of the structural equation model.Using matter-element extension theory,the correlation degree calculation method for indicators was determined,deriving weight relationships between the evaluation problem and various primary and secondary controlling factors.Combined with the standardized paths from the structural equation model,the global comprehensive correlation degree of the entire evaluation system was calculated.Through case analysis,the coordination status and efficiency were quantitatively evaluated.Results showed that the comprehensive weights of the four primary controlling factors were 0.286,0.258,0.231,and 0.255,respectively,with capacity matching degree having the highest weight,indicating it as the key factor affecting truck-shovel coordination efficiency.By removing data items with excessively long waiting times from the dataset to construct a comparative experiment,the correlation degree indicators for system stability and waiting cost were significantly improved,changing from-0.088 66 and-0.056 83 to-0.006 34 and-0.077 48,respectively,and the evaluation grade increased from 1 to 3.The research results demonstrate that combining the structural equation model with matter-element extension theory can effectively address the multidimensional coupling and boundary fuzziness problems in coordination evaluation,providing reliable quantitative decision support for equipment capacity release and transportation scheduling optimization in open-pit coal mine stripping processes.

Abstract:During the relocation of crushing stations in open-pit coal mines,the haulage routes of overburden-transport trucks(overburden trucks)and coal-transport trucks(coal trucks)may intersect,which can affect transportation efficiency and may even lead to accidents.At present,studies on intersection-scheduling technologies remain at the theoretical level and differ from actual mine-site conditions.Taking the relocation of three crushing stations in the Baorixile Open-pit Coal Mine as the engineering background,a microscopic traffic flow model was established to address the route-intersection problem between overburden trucks and coal trucks,describing the motion states of individual vehicles and the overall traffic characteristics of the transportation system,with emphasis on analyzing the car-following behavior influenced by adjacent vehicles.On this basis,an intersection traffic flow control scheme based on intelligent traffic signals was proposed.By analyzing the lane-orientation combinations of the intersection and utilizing vehicle-type,speed,and other information collected by the intelligent traffic signals,a traffic-signal control scheme was developed.The vehicle-passing capacity of the intersection with and without intelligent signal control was compared.The results showed that without intelligent traffic signals,the maximum flow of overburden trucks passing through the intersection was 74 vehicles/h,which did not meet the required capacity of 112 vehicles/h and would have required increasing the number of lanes.After installing intelligent traffic signals,the maximum flow of overburden trucks reached 292 vehicles/h,meeting the required capacity without changing lane numbers,providing additional surplus capacity,and mitigating potential safety hazards.

Abstract:Complex terrains such as deep pits and high slopes in open-pit mines cause physical signal blockage and multipath fading.Furthermore,existing path-planning algorithms tend to result in unstable vehicle trajectories or congestion due to local optima,leading to low efficiency in open-pit mine vehicle scheduling.To address these issues,an intelligent vehicle scheduling method for open-pit mines based on 5G antenna and improved Dijkstra algorithm was proposed.At the communication level,based on a biconical antenna model,an L-shaped radiating stub was loaded,and rectangular and L-shaped slots were etched on the radiating patch to optimize current distribution,thereby forming an onboard dual-band omnidirectional dipole antenna.This antenna achieved coverage in the 2.3-2.7 GHz and 4.8-4.9 GHz dual frequency bands,solving the problems of signal blockage and attenuation caused by deep pits and high slopes in mining areas.At the path-planning level,depth-first search and a"container array"mechanism were introduced into the traditional Dijkstra algorithm.By recording all potential predecessor information of nodes,global path backtracking and optimal selection were realized,improving the smoothness of planned paths.Experimental results showed that the dual-band omnidirectional dipole antenna achieved a signal coverage rate of 81.2％in areas with severe signal blockage such as deep pits and high slopes,with an average signal strength of-94 dBm,outperforming traditional commercial 5G antennas.Compared with the Dijkstra algorithm,A*algorithm,and Rapidly-exploring Random Tree(RRT)algorithm,the improved Dijkstra algorithm planned paths with shorter distance,fewer inflection points,and smoother trajectories.Moreover,in multi-vehicle cooperative transportation scenarios,it exhibited lower path conflict rate and shorter path-replanning response time.In actual open-pit mine vehicle scheduling,compared with the production-completion method,earliest-loading method,and traffic-flow planning method,the proposed method effectively reduced vehicle waiting time and shortened loaded travel distance,and performed optimally in terms of single-shift total output,empty-haul rate,and scheduling-command response delay.

Abstract:At present,commonly used open-pit mine slope stability analysis methods,such as the limit equilibrium method and the finite-element strength reduction method,rely on the safety factor as the sole criterion,which presents certain limitations.This study proposed the use of a fuzzy measure method,which effectively handled fuzziness and uncertainty,to conduct slope stability analysis for open-pit mines.To address the difficulty of accurately determining fuzzy parameters in this method,a Genetic Algorithm(GA)-optimized BP neural network(GABP)model was designed to predict the fuzzy parameters.The experimental results showed that the mean relative errors of the GABP-predicted fuzzy parameters ξ and η were 3.66％and 3.25％,respectively,both lower than those of the BP neural network.By substituting the fuzzy parameters predicted by the GABP model into the fuzzy measure method,the stability of the slope in the Gaocun Iron Mine was analyzed.The calculated slope failure probability was 0.155 4,indicating overall stability with local instability,which was highly consistent with field monitoring results.The finite-element strength reduction method and the Bishop method were further used to validate the accuracy of the GABP-optimized fuzzy measure method.The results showed that the three methods yielded consistent outcomes,whereas the GABP-optimized fuzzy measure method required lower computational cost and provided higher efficiency.It also accurately characterized the progressive evolution of the slope from stability to failure,which better matched practical engineering conditions.

Abstract:High-bench throwing blasting dust in open-pit mines exhibits essential differences in diffusion dynamics compared with conventional blasting dust.However,existing numerical models are mostly established under conventional blasting conditions and have not been adjusted or accurately characterized to capture the characteristics of throwing blasting dust.In addition,key input parameters for numerical simulations lack sufficient field-measured data support,which restricts the prediction accuracy and engineering applicability of simulation results.To address these problems,field monitoring was conducted using high-speed photographic observation technology.The results showed that the dust migration process could be divided into an impact motion stage,a"mushroom cloud"formation stage,and a diffusion motion stage.After blasting,dust concentration first rose rapidly,then fluctuated continuously,and finally decayed.Dust was dominated by medium-sized particles(20-100 μm),accounting for 56.2％of the cumulative proportion;fine particles(＜20 μm)accounted for 34.4％;and coarse particles(＞100 μm)accounted for only 9.4％.A three-dimensional geometric model of dust migration from high-bench throwing blasting in open-pit mines was established,and transient numerical simulations of dust migration were carried out using Fluent.The results indicated that ① during the initial rapid-release stage,dust exhibited highly aggregated distribution characteristics under the dominant effects of gravity and inertial forces;subsequently,during the sustained transport stage,dust was significantly lifted and expanded laterally under the effects of buoyancy and ambient wind fields;and finally,in the diffusion-settling stage,overall dust concentration decreased markedly due to gravitational settling and dilution by turbulent diffusion.② Coarse dust particles settled rapidly,medium dust particles migrated with the airflow,and fine dust particles remained suspended for a long period and were transported over long distances.

Abstract:Generating panoramic borehole wall images and"virtual cores"through image unwrapping and stitching is the basis for quantitative analysis of geological structures and is also an urgent need in borehole investigation.To address the problems of blurred image contours and discontinuous stitching in existing borehole image unwrapping and stitching methods,an image unwrapping method based on the Random Sample Consensus(RANSAC)algorithm and an image stitching method based on the Speeded Up Robust Features(SURF)plus Maximum Likelihood Estimation Sample Consensus(MLESAC)algorithm were proposed.The RANSAC algorithm was used to perform circle fitting on borehole edge data to obtain the borehole center and borehole inner diameter,thereby determining the effective annular image region.By combining the coordinate transformation algorithm and bilinear interpolation method,the effective annular region was then unwrapped into a rectangular image,which effectively compensated for geometric distortion caused by probe jitter and achieved high-precision unwrapping of borehole images.The SURF algorithm was used to rapidly extract feature points of adjacent images and perform coarse matching while maintaining feature stability under image rotation and scale transformation.The MLESAC algorithm was then used to perform precise matching of the extracted feature points,remove mismatched points,and select the optimal matching pairs,thereby estimating horizontal and vertical offset parameters and achieving global high-precision stitching of video images.Comparative analysis results showed that,compared with the image unwrapping and stitching methods used in existing devices,the proposed method produced panoramic borehole wall images that were continuous and uninterrupted,with high-resolution detail,and significantly improved both stitching quality and visual quality.

Abstract:Existing studies on the stability of internal dump slopes underlain by gently dipping weak layers have primarily focused on the effects of dump spatial morphological parameters or inclined basal surfaces,while the influence of the spatial geometric relationship between the internal dump and the basal surface on slope stability has rarely been examined.Under different horizontal intersection angles between the inclination of the internal dump and that of the basal weak layer,significant differences exist in slope stress distribution,potential landslide modes,and potential sliding-body geometries.Taking an internal dump underlain by a gently dipping weak layer in an open-pit coal mine in Xinjiang as the research background,this study investigated the stability of these internal dump slopes during mine turning in the open-pit mine.The deformation characteristics of slopes under internal dump turning angles of 15°,30°,45°,60°,75°,and 90° were analyzed,and the potential landslide modes of internal dump slopes underlain by gently dipping weak layers were identified.Using the strength reduction method,the variation pattern of slope stability factor with internal dump turning angle was examined.The results showed that the internal dump turning angle had a highly significant influence on slope stability:the slope stability factor decreased logarithmically with increasing turning angle.Moreover,a larger turning angle led to a stronger influence of the weak layer on the potential sliding body,and the potential landslide mode evolved from a cutting-layer-dominated mode to a bedding-plane-dominated composite sliding mode.With increasing turning angle,the basal dip angle of the potential main slip surface increased linearly,while the internal dump space volume increased quadratically.Engineering practice recommends an internal dump turning angle of 45 °-60°;if a larger angle is required,slope-stability enhancement measures must be adopted.

Abstract:Slope engineering is a key factor affecting production safety in large open-pit coal mines.This study conducts a statistical analysis of the current status and future technical conditions of slope engineering in 20 typical large open-pit coal mines in China,revealing the development trend from medium-high slopes toward ultra-high slopes.It summarizes common issues in slope engineering,including unclear attribute variations and complicated yet independent monitoring methods.This study proposes slope engineering digitalization and integrated monitoring and early warning as the core technical pathway to effectively address these challenges.From the perspectives of digitalization of slope-engineering geological bodies and spatiotemporal attribute modeling in slope engineering,the study analyzes key technologies for slope-engineering digitalization and points out that future breakthroughs will be achieved in semantic consistency between production data and geological models,in digitalization and application of multi-attribute geological spatial data,and in refined reconstruction and dynamic updating of models.The analysis indicates that integrated early warning,advanced prediction,and intelligent decision-making are key directions for the effective utilization of multi-source monitoring data.Future developments in integrated monitoring and early warning technologies will include breakthroughs in autonomous landslide source tracing and causal inference,coordinated inspection and rapid response using unmanned aerial vehicle swarms,and expert systems for emergency decision-making based on large multimodal models.Ultimately,an intelligent disaster-prevention system for slope engineering in large open-pit coal mines will be formed,integrating monitoring,early warning,prediction,and decision-making,thereby enhancing the capability for forward-looking and precise prevention and control of landslide hazards.

Abstract:In high-altitude underground mining operations,the soil moisture content at mine outlets is unevenly distributed and easily affected by rainfall seepage.The moisture variation from shallow to deep layers is a key factor inducing debris flow or mud inrush disasters.Traditional contact measurement methods cannot effectively monitor this area.To address this problem,a soil moisture inversion method based on Stepped Frequency Continuous Wave(SFCW)ground-penetrating radar and Support Vector Machine(SVM)was proposed to measure the soil moisture content at mine outlets.The method set the detection depth at 0.3 m,and achieved rapid and non-contact measurement of soil moisture content at mine outlets through radar scanning,signal preprocessing,feature extraction(reflection coefficient and phase difference),and SVM modeling and inversion.Simulation results showed that SFCW radar effectively identified reflection features caused by moist regions,verifying the feasibility of this method in detecting layered and heterogeneous soils.Under laboratory conditions,SFCW radar was used to measure sandy loam samples collected from the Pulang Copper Mine outlet,and by comparing multiple modeling methods,the SVM model based on the combined features of the reflection coefficient and phase difference was found to produce the best inversion performance.Field measurements at the Pulang Copper Mine outlet showed that the relative error between predicted and measured values was within the range of 9.67％to 14.53％,indicating that the method provides relatively reliable soil moisture estimation within an effective detection range of 1 m horizontally and 0.3 m in depth.

Abstract:UAVs have significant advantages over traditional measurement and remote-sensing technologies in monitoring open-pit mining areas.At present,existing UAV-based detection methods for spontaneous combustion in open-pit coal seams mainly suffer from the lack of corresponding detection models capable of identifying high-temperature points,low recognition accuracy for small-size and multi-scale high-temperature points,and confusion between exhaust-pipe high temperatures of excavators and spontaneous combustion high-temperature points on coal seams.To address these problems,a YOLO-WRC-based detection method for spontaneous combustion in open-pit coal seams using UAV imagery was proposed.Wavelet Transform Convolution(WTConv)was integrated into the backbone network to focus on richer feature information;a Reparameterized Generalized Feature Pyramid Network(RepGFPN)was used to reconstruct the neck network,enhance the ability of feature extraction and fusion and the recognition accuracy of easily confused high temperature points;a Concentrated Layerwise Localization Attention Head(CLLAHead)was introduced to coordinate feature and semantic information across different levels,focusing on the identification of micro high-temperature points;and the PIoUv2 loss function was adopted to improve the model's regression performance for multi-scale abnormal high-temperature points.The experimental results showed that ① the accuracy,recall,and mAP@0.5 of YOLO-WRC reached 88.2％,90.1％,and 95.4％,respectively,which were 1.3％,2.2％,and 3.2％higher than those of the original YOLOv8n model.② The recall and mAP@0.5 of YOLO-WRC were superior to mainstream models such as SSD,Faster-RCNN,YOLOv5,and YOLOv10n,demonstrating high robustness and adaptability in identifying abnormal high-temperature points.③ YOLO-WRC yielded higher confidence for detection targets and identified targets missed by YOLOv8n,exhibiting stronger recognition capability for easily confused and small-sized targets.

Abstract:Existing prediction models for vehicle speed in open-pit mines do not fully consider the influence of meteorological factors on operating speed.The data sample sets are insufficient to train or fine-tune deep learning models.Production planning and dynamic adjustment of the number of dispatched equipment according to weather conditions have not been implemented,resulting in substantial redundancy of equipment in the transportation process and failing to achieve the goal of refined management in open-pit mines.To address this issue,this study proposed a vehicle speed characteristic analysis method and a speed prediction model based on spatiotemporal and meteorological factors.Clustering analysis and correlation analysis were used to obtain the speed characteristics of vehicles under different meteorological and spatiotemporal conditions.By integrating factors such as road location,operating period,and meteorological conditions,a prediction model for vehicle speed in open-pit mines was derived for scenarios including straight roads,uphill sections,turns,and downhill sections under no-precipitation,light-rain,moderate-rain,heavy-rain,and snowfall conditions.A model for safe vehicle speed was constructed based on parameters such as road friction coefficient,driver reaction time,and vehicle braking distance,yielding the safe operating speeds for vehicles in different spatiotemporal scenarios.The analysis results showed that:① the average speed of empty vehicles was higher than that of loaded vehicles;the difference in average speed between empty and loaded vehicles was significant in turns and downhill sections,while it was relatively small in straight and uphill sections.② Vehicle speed was negatively correlated with precipitation and strongly correlated with road structure and temporal distribution.③ The speed prediction model achieved a mean absolute percentage error≤3％,a mean absolute error 0.4 km/h,and a root mean square error＜3 km/h,demonstrating good predictive performance.

Abstract:Residual water in the surrounding rock is a key factor that affects the slope stability of inner dumps in open-pit mines.Existing research on groundwater seepage in inner dumps has mainly focused on analyzing the effects of individual factors,such as rainfall or water level fluctuations,on slope stability.However,a comprehensive research framework for water prevention and control schemes under the combined influence of residual water in surrounding rock and groundwater recharge has not yet been established.To address this issue,taking an open-pit mine in the Thar Coalfield of Pakistan as the research background,a treatment scheme involving the reconstruction of the sand cushion layer at the base of the inner dump was proposed.The sand cushion in the inner dump was reconstructed into a four-layer structural system composed of waste material,sand cushion,impervious layer,and the third aquifer.The SEEP/W seepage model and the SLOPE/W stability model were established using GeoStudio software.The variations in seepage field and slope stability of the inner dump were compared and analyzed under conditions of different thicknesses and permeability coefficients of the reconstructed sand cushion.The results showed that increasing the permeability coefficient and thickness of the sand cushion could effectively lower the groundwater level in the inner dump and improve the overall slope stability.A reconstructed sand cushion with a thickness of 11.5 m and a permeability coefficient of 23.8 m3/d,or one with a thickness of 5.2 m and a permeability coefficient of 47.6 m3/d,enabled the slope stability coefficient to meet the required safety reserve factor.

Abstract:To address issues in traditional open-pit mine slope monitoring and early warning,such as limited monitoring technologies,inadequate multi-source data fusion,and ineffective early warnings,this paper reviews the research progress in slope disaster monitoring and early warning from three perspectives:intelligent slope sensing and monitoring,high-precision 3D slope modeling and visualization,and slope stability assessment with risk warning.The monitoring methods,including the Global Navigation Satellite System,the integration of UAV oblique photography with LiDAR,and multi-scale sky-air-ground integrated monitoring,are systematically summarized.Cutting-edge technologies,such as 3D visualization and modeling of complex geological structures,as well as digital twin-driven full-element 3D visualization of slopes,are reviewed.Key technologies,including machine learning-driven intelligent slope analysis,multi-model integrated slope assessment,and intelligent slope monitoring and early warning platforms based on multi-source data fusion,are analyzed and organized.In response to current challenges in open-pit mine slope monitoring and early warning-such as insufficient multi-source information fusion capability,weak interactivity and simulation performance of 3D slope models with poor dynamic visualization,low generalizability of slope risk assessment and early warning models,and a lack of emergency response modules in monitoring platforms-this study outlines development trends for slope disaster safety management:accelerating the establishment of an intelligent sensing and monitoring system integrating multi-source data;enhancing slope risk assessment through digital intelligence to improve the accuracy of machine learning and enable real-time transparent analysis and feedback of slope information;and developing an intelligent disaster monitoring and early warning platform with holistic sensing,collaborative warning,and smart emergency response capabilities.

Abstract:Existing coupled-causation analyses of landslide accidents in open-pit mines lack quantitative characterization and inference of the coupling relationships among risk factors,such as coupling strength and sensitivity of risk factors,which have certain limitations for coupled risk analysis of landslides in open-pit mines.To address this issue,a coupled analysis method for landslide risk in open-pit mines that integrated the N-K model and the Bayesian Network(BN)model was proposed.First,by analyzing 51 officially reported landslide accident cases in open-pit mines,the direct causes of the accidents were classified and summarized,and human factors,environmental factors,mining equipment factors,management factors,and types of risk coupling were defined.Then,the N-K model was constructed to explore the coupling relationships and coupling mechanisms among risk factors.Finally,the calculation results of the N-K model were integrated into the BN model to conduct forward and backward inference and to analyze the sensitivity of risk factors.The results showed that the risk coupling type with the highest probability of causing landslide accidents in open-pit mines was the"human-environment-management"coupling,with a probability of 37.32％;the"human-environment-equipment-management"coupling had the highest risk coupling value,approximately 0.26,with an occurrence rate of 29.60％.Based on the findings,the study suggests that environmental risk factors and management risk factors should be taken as the entry points for systematic and overall control of landslide risks in open-pit mines,so as to prevent the joint effects and chain effects among risk factors and to cut off the coupling relationships among them as much as possible,thereby reducing the occurrence rate of landslide accidents in open-pit mines.

Abstract:The operating environment of open-pit mines is complex,and mechanical targets in UAV remote-sensing images often exhibit significant scale variations,complex background interference,and a high risk of missed detections of small targets.To address these problems,a lightweight model named CDMG-YOLO was proposed for small mechanical target detection in UAV remote sensing images of open-pit mines.The model was improved based on YOLOv11n.In the feature perception stage,a P2 detection head was added and combined with shallow backbone features to construct a fine-grained information pathway,enhancing the detail features of small targets.In the feature modeling stage,the D-C3k2 module incorporating Deformable Large Kernel Attention(DLKA)was introduced to expand the receptive field and improve scale generalization.In the feature fusion stage,a dual-attention mechanism-namely the Convolution and Attention Fusion Module(CAFM)and the Multidimensional Collaborative Attention Module(MCAM)—achieved multidimensional interaction and adaptive weighting of features along the channel and spatial dimensions,thereby highlighting target features and suppressing background interference.In the training optimization stage,a composite loss function combining the Gradient Harmonizing Mechanism Loss(GHM Loss)was adopted,which balanced the gradient contributions of hard and easy samples during training and effectively alleviated the positive-negative sample imbalance problem.Experimental results showed that on a self-built UAV remote-sensing dataset covering multiple types of small mechanical targets in open-pit mines,CDMG-YOLO achieved a precision of 0.859 and an mAP@0.5 of 0.697,with only 3.1×106 parameters and 12.7 GFLOPs,achieving a balance between high accuracy,lightweight design,and efficiency.In low-contrast scenes,severe occlusion situations,and densely distributed multi-target scenarios in complex open-pit mine operations,the CDMG-YOLO model achieved accurate target localization and recognition.On the public LEVIR dataset,CDMG-YOLO accurately identified different types of targets and demonstrated good generalization capability.

Abstract:To address the problems of missed and false detections in vehicle obstacle detection under the complex conditions of severe occlusion,dust interference,and image blurring in open-pit mine operational areas,this study proposed a YOLO-DIS model based on YOLOv11n for vehicle obstacle detection in such complex open-pit mine environments.To enhance feature extraction capability in complex situations,the model introduced an Iterative Attention Feature Fusion(iAFF)mechanism to improve the C3k2 module,strengthening feature extraction through two-stage iterative attention fusion.To effectively compensate for the inaccurate recovery of edge features in occluded targets caused by fixed sampling rules,lightweight dynamic upsampling was adopted to replace the original nearest-neighbor interpolation method.This approach dynamically adjusted sampling positions based on target shape and occlusion by learning sampling point offsets.Furthermore,to solve the problem of imbalanced sample distribution,the SlideLoss function was employed to assign differentiated weights to samples of varying difficulty levels.Experimental results demonstrated that:compared to YOLOv11n,the YOLO-DIS model achieved improvements of 4.4％,7.3％,and 4.0％in precision,recall,and mAP@0.5,respectively,with only a marginal increase in the number of parameters;compared to mainstream object detection models,the YOLO-DIS model achieved the highest mAP@0.5;the YOLO-DIS model maintained good detection performance on both a custom dataset and the KITTI dataset,indicating strong generalization capability.The YOLO-DIS model provided higher confidence levels for detection bounding boxes in scenarios involving severe occlusion,dust interference,image blurring,small target detection,and background interference,effectively reducing missed detections.

Abstract:The intelligent construction of"one ventilation and three prevention"in mines faces problems in the four aspects of"man-machine-environment-management",including high professional thresholds,difficulties in digitizing expert experience,lagging warnings for multi-field coupled disaster chains,and"data islands"caused by multi-source heterogeneous data.Based on an analysis of the development of coal mine intelligence and the necessity of applying artificial intelligence technology for intelligent mine construction,this study explores the adaptability,potential advantages,and future development directions of DeepSeek in building the intelligent systems of"one ventilation and three prevention"in mines.With its characteristics of multimodal perception,dynamic data modeling,complex reasoning and early-warning decision-making,and long-context understanding and knowledge management,DeepSeek is highly compatible with the"perception-analysis-decision-execution"process in the intelligent construction of"one ventilation and three prevention",and shows excellent synergy in the integrated intelligent construction of the"man-machine-environment-management"mechanism.For example,it enables personnel to shift from experience-dependent work to data-driven work,optimizes efficient and low-consumption electromechanical operation and maintenance mechanisms,and enables the multi-factor environment of"underground technology-market-policy"to evolve from static monitoring to dynamic early warning.Ultimately,it helps build a proactive closed-loop intelligent management system for"one ventilation and three prevention"featuring"intelligent perception-data fusion-routine and emergency assessment-autonomous decision-making-precise execution-feedback optimization."Coal mine enterprises such as Shandong Energy Group,which have already connected the DeepSeek-R1 large model,report that information acquisition efficiency and collaboration/decision-making efficiency have improved by more than 80％and 20％,respectively.Meanwhile,this study analyzes three core challenges faced by DeepSeek in its application and promotion—AI hallucinations,lightweight on-premise deployment,and information security-with the aim of providing theoretical reference and solutions for the intelligent construction of"one ventilation and three prevention"in mines.

Abstract:To address the lack of systematic and in-depth research on roof stability and its control throughout the entire life cycle of the open-off cut from roadway excavation to the first caving of the main roof,the concept of the entire life cycle of the open-off cut was proposed.The entire life cycle of the open-off cut was divided into five stages:primary roadway excavation,secondary roadway expansion,coal mining equipment installation,immediate roof caving in the goaf,and first caving of the main roof in the goaf.Through theoretical analysis,it was concluded that geological conditions were the fundamental factor determining roof stability in the open-off cut.Roadway layout,roadway span,and excavation disturbance were important mining technical factors affecting the stability of the roof strata.Support technology was the key measure for controlling roof stability.By establishing a mechanical model of the main roof rock beam,the overlying strata structure characteristics of the open-off cut in the five stages were studied,and the spatiotemporal evolution and dynamic change process of the open-off cut roof from bending subsidence to collapse were revealed.Corresponding control technologies were proposed according to the roof structure characteristics at different stages.During the roadway excavation stage,active support with bolt-mesh-cable should be adopted.Before initial mining,roof cutting and pressure relief should be conducted behind the hydraulic supports.Industrial trial results showed that implementing blasting pre-splitting for roof cutting behind the supports could pre-damage the integrity of the open-off cut roof,reduce the first caving interval of the goaf roof,weaken the first weighting intensity of the main roof,and avoid the safety hazard of large-area roof hanging in the goaf.