- 0
- 0
- 约9.17千字
- 约 3页
- 2026-03-07 发布于福建
- 举报
时空智能·Space-timeIntelligence
引文格式:庞斌.轻量级人工智能模型在无人机矿山实景三维模型实时重建中的应用[J].北斗与空间信息应用技术,2025,12(6):7-9.
轻量级人工智能模型在无人机矿山实景
三维模型实时重建中的应用
庞斌
(中国能源建设集团山西省电力勘测设计院有限公司,山西太原030001)
摘要:本文针对传统矿山勘测高成本、低效率等问题,提出基于轻量级人工智能(AI)的无人机矿山实景三维模型即时
重构方法。通过集成视觉即时定位与地图构建(SLAM),稠密点云重建、纹理映射和多分辨率管理等技术,给出一种高效准
确的矿山三维建模框架。试验结果表明:所提出的框架可实现复杂矿山环境中高精度三维模型的实时制作,数据的收集和处理
所需时间分别是100~140s和80~110s,其重建的精确度在0.02~0.04m之间,而实时更新的延迟时间少于5s。本研究显著提高了
三维建模精度和效率,有较好的实际运用前景,对矿山数字化管理有较强的支撑作用。
关键词:无人机;矿山三维重建;实时性;轻量级AI模型
ALightweightAIModelforReal-Time3DReconstructionofMineSitesUsingUnmannedAerialVehicles
PangBin
(ChinaEnergyConstructionGroupShanxiElectricPowerSurveyandDesignInstituteCo.,Ltd.,Taiyuan,Shanxi,030001,China)
Abstract:Addressingthechallengesofhighcostandlowefficiencyintraditionalminesurveying,thisstudyproposesamethod
forthereal-timereconstructionofhigh-precision3DminemodelsusingUnmannedAerialVehicles(UAVs)integratedwithlightweight
ArtificialIntelligence(AI).TheproposedframeworkincorporatesVisualSimultaneousLocalizationandMapping(SLAM),densepoint
cloudreconstruction,texturemapping,andmulti-resolutionmanagement.Experimentalresultsdemonstratetheframeworksefficacyin
complexmineenvironments,achievingreal-timegenerationofdetailed3Dmodels.Dataacquisitionandprocessingtimeswererecorded
at100-140secondsand80-110seconds,respectively,withareconstructionaccuracybetween0.02mand0.04m,andareal-timeupdate
latencyofunder5seconds.Thisresearchsignificantlyenhancestheprecisionandefficiencyof3Dmodeling,showingconsiderable
promiseforpracticalapplicationandrobustsupportfordigitalminemanagement.
Keywords:UAV;3Dreconstructionofmines;Realtime;Lightweigh
原创力文档

文档评论(0)