綜采工作面刮板輸送機煤流輪廓點云的配準(zhǔn)方法研究

打開文本圖片集

中圖分類號：TN249-34；TP391.9 文獻標(biāo)識碼：A 文章編號：1004-373X（2025）16-0081-07

Research on registration method of coal flow profile point cloud of scraper conveyor in synthesized miningface

ANGWeibing'，LIKaifang'，ZHAOShuanfeng'，WANGYuan’，LUZhengxiong1，LILai1，GUO Shut （1.SchoolofMechanical Engineering，Xi'an UniversityofScienceand Technology，Xi’an71Oo54，China; 2.OrdosResearch InstituteofEnergy，PekingUniversity，Ordos O17ooo，China）

Abstract：Antrational iterativeclosest-point （ICP）alignmentalgorithmisimprovedfortheproblemsthatthecoal flow contourpointcloudofthescraperconveyorinthesynthesizedminingfacehasmanynoisepoints，thecomplexityof thecontour structure，andtheexisting pointcloudalignmentlgorithmcanotbeadaptedtothefastandhigh-precisionalignmentof thecoal flowpointcloud.Theprincipalcomponentanalysismethodisusedtoperforminitialaxialalignmentforthepointcloudtobe registered，thescale-invariantfeaturetransforalgoritmisusedtoextractthefeaturepointsofthepointcoudtobergistered， andthefastpointfeaturehistogramisconstructedtoensurethatthetwopointcloudprincipalaxeswillnotbereversed，which canimprove theeficiencyofthecoarsealignmentalgorithm.Thecorresponding pointpairsaresearchedby meansof random samplingconsistentinitialregistrationalgoritandtheinitialrigidbodytransformationmatrixisalculated，whichisusedto realizethepreliminaryregistrationofthetwo-pointcloudandprovideagoodinitialpositionforthesubsequentfineregistration. Onthe basisof abovecoarsealigment，theK-Dtreedata structureisutilizedtoaccelerate thecorespondence point finding processandtheminimum distance frompointtosurface methodisused toimprove theaccuracyof thecorespondence relationship.Wrong pairsof corrspondencepointsareeliminated iterativelybymeansof therandomsampling consistency algorithmtoenhancetheaccuracyof thealignment.Therigid-bodytransformationmatrixiscalculatedacordingtotheprecise correspondingpointpairs，soas torealizethefineregistrationofthecoalflowpointclouddata.Theexperimentalresultsshow that，incomparisonwithotherpointcloudalignmentmethods，theproposedalignmentalgorithmcanimprovethematching accuracyandmatchingeficiencyofthecoalflowprofilepointcloudofthescraperconveyor，whichisof greatsignificancefor

thevolumetric calculation of the coal flowprofilepoint cloud.

Keywords：scraper conveyor;pointcloudofcoalflow profile;pointcloud alignment;principalcomponentanalysis;scaleinvariant feature transform;randomsamplingconsistentalgorithm

0 引言

據(jù)預(yù)測，在我國達到“碳峰值"之前，煤炭作為“壓艙石"的定位將保持不變[-2]。（剩余8647字）