Towards Intensity-Augmented SLAM with LiDAR and ToF Sensors利用激光雷达和ToF传感器实现强度增强的SLAM

《Towards Intensity-Augmented SLAM with LiDAR and ToF Sensors利用激光雷达和ToF传感器实现强度增强的SLAM》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、TowardsIntensity-AugmentedSLAMwithLiDARandToFSensorsRobertA.Hewitt1andJoshuaA.Marshall2Abstract—AlthoughpassivesensorsarewidelyusedformanytheapplicationoftheseresultstoEarthapplications(e.g.,inmobileroboticsapplicationsthatperformmappingandlocal-mining)whereroboticvehiclesworkinenvironmentshavingiz

2、ationfunctions,therearemanyenvironments(e.g.,miningandunpredictablelightingconditionsornoexternallighting.planetary)whereactivesensorsaremorepractical.However,Firstly,weaskwhetheritispossibletomodelintensityatpresent,most3DSLAMalgorithmsthatdouseLiDARand/ortime-of-flight(ToF)sensorsexploitonlyrangea

3、ndmeasurementsasafunctionofrobotandmapparameters.bearinginformationassociatedwiththesemeasurements,butTheanswertothisquestionliesinalargebodyofpreviousnotintensityinformation.Thispaperpresentsanewapproachwork.Inthecomputervisionliteraturetheproblemofshape-thatattemptstoexplicitlyincorporateanintens

4、itymodelasfrom-shadingbearsmanysimilaritiestotheproblemathand.partofasparsebundleadjustment(SBA)estimationproblem.ByassumingaknownlightsourcedirectionandmodellingAnobservabilityanalysisshowsthatasolutionexists,andsimulationresultsverifyitspotentialutility.brightnessasafunctionofsurfaceshape,ithasbe

5、enshownErrata—TheoriginallypublishedversionofthispaperthatshapecanbederivedfromintensitygradientsincameracontainederrorsinTableII.Therankdeficiencyoftheimages[3].Asanextensiontothis,varyingknownlightSLAMproblemforvariouscombinationsofmeasurementswassourcedirectionsacrossmultipleimagesofthesamescene,

6、computedincorrectlywhenincludingtheminimumnumberhasbeenusedinatechniquecalledphotometricstereotooflandmarks.Aftercorrectingthis,theminimumnumberoflandmarksforasolutionwhenincludingintensityandsurfaceestimatesurfaceshapeandreflectivity[4].normalmeasurementsis2,not1(stillanimprovementoverHofleandPfeife

7、r[5]describedamethodforproducing¨traditionalSLAM).ThesimulationdoesnotrelyonthesearialLiDARintensityimagesthatareproportionaltothecomputationsandcontinuestoshowtheimprovementgainedreflectanceofthemeasuredsur