Data_NN0 and_CJC methods_NN2 
Overview_NN1 
This_DT0 study_NN1 investigates_VVZ mobile_AJ0 LiDAR_NP0 's_POS dense_AJ0 ground-level_AJ0 collection_NN1 capabilities_NN2 to_TO0 assist_VVI the_AT0 urban_AJ0 design_NN1 /_PUN analysis_NN1 and_CJC transportation_NN1 research_NN1 community_NN1 with_PRP two_CRD specific_AJ0 objectives_NN2 ._SENT 
The_AT0 first_ORD objective_NN1 is_VBZ to_TO0 provide_VVI descriptive_AJ0 statistic_NN1 matrices_NN2 that_CJT may_VM0 be_VBI used_VVN as_PRP alternatives_NN2 to_PRP some_DT0 of_PRF the_AT0 audit-based_AJ0 methods_NN2 for_PRP streetscape_NN1 evaluation_NN1 with_PRP more_DT0 precise_AJ0 and_CJC accurate_AJ0 data_NN0 ._SENT Since_CJS the_AT0 research_NN1 strand_NN1 tends_VVZ to_TO0 revolve_VVI around_AVP the_AT0 walkability_NN1 aspects_NN2 of_PRF urban_AJ0 streetscapes_NN2 ,_PUN we_PNP attempt_VVB to_TO0 focus_VVI on_PRP the_AT0 sidewalk_NN1 areas_NN2 of_PRF busy_AJ0 downtown-like_AJ0 contexts_NN2 ._SENT For_PRP these_DT0 streetscapes_NN2 ,_PUN we_PNP also_AV0 evaluate_VVB mobile_AJ0 LiDAR_NP0 's_POS ability_NN1 to_TO0 collect_VVI vertical_AJ0 building-face_NN1 data_NN0 that_CJT aerial_AJ0 LiDAR_NN1 misses_VVZ such_DT0 as_CJS window_NN1 proportions_NN2 ,_PUN awnings_NN2 ,_PUN and_CJC commercial_AJ0 signs_NN2 ._SENT 
Our_DPS second_ORD objective_NN1 is_VBZ to_TO0 determine_VVI mobile_AJ0 LiDAR_NP0 's_POS capability_NN1 of_PRF fitting_AJ0 small_AJ0 streetscape_NN1 such_DT0 as_PRP street_NN1 signs_NN2 ,_PUN traffic_NN1 lights_NN2 ,_PUN and_CJC streetscape_NN1 furniture_NN1 with_PRP a_AT0 precise_AJ0 ,_PUN voxel_AJ0 matrix_NN1 ._SENT Mobile_NP0 LiDAR_NP0 should_VM0 be_VBI highly_AV0 effective_AJ0 for_PRP this_DT0 level_NN1 of_PRF analysis_NN1 because_CJS our_DPS voxel_NN1 size_NN1 can_VM0 be_VBI refined_VVN more_AV0 than_CJS the_AT0 5-ft_AJ0 vertical_AJ0 intervals_NN2 used_VVN in_PRP the_AT0 previous_AJ0 research_NN1 (_PUL Golombek_NP0 and_CJC Marshall_NP0 2019_CRD ,_PUN 2020_CRD )_PUR ._SENT 
Data_NN0 /_PUN data_NN0 collection_NN1 
A_AT0 Trimble_NP0 MX9_NN1 was_VBD used_VVN to_TO0 collect_VVI data_NN0 for_PRP this_DT0 study_NN1 ._SENT The_AT0 MX9_NN1 contains_VVZ a_AT0 spherical_AJ0 imaging_NN1 system_NN1 and_CJC three_CRD oblique_AJ0 view_NN1 cameras_NN2 ._SENT The_AT0 MX9_NN1 contains_VVZ Riegl_NP0 VUX-1HA_NP0 dual_AJ0 laser_NN1 scanners_NN2 that_CJT measure_VVB at_PRP 1000_CRD kHz_NN0 with_PRP the_AT0 scanner_NN1 mirror_NN1 rotating_VVG at_PRP 250_CRD revolutions_NN2 per_PRP second_NN1 and_CJC a_AT0 field_NN1 of_PRF view_NN1 of_PRF 360_CRD °_NN0 ._SENT A_AT0 Real-Time-Kinematic_AJ0 (_PUL RTK_NP0 )_PUR base_NN1 station_NN1 was_VBD set_VVN up_AVP for_PRP post-processing_AJ0 purposes_NN2 on_PRP a_AT0 nearby_AJ0 National_AJ0 Geodic_AJ0 Survey_NN1 (_PUL NGS_NP0 )_PUR monument_NN1 set_VVN at_PRP a_AT0 1-s_CRD interval_NN1 collection_NN1 rate_NN1 ._SENT To_TO0 confirm_VVI system_NN1 accuracy_NN1 ,_PUN 13_CRD ground_NN1 control_NN1 points_NN2 (_PUL GCP_NP0 )_PUR were_VBD established_VVN throughout_PRP the_AT0 survey_NN1 area_NN1 that_CJT were_VBD no_AV0 more_AV0 than_CJS a_AT0 half-mile_NN1 apart_AV0 from_PRP each_DT0 other_NN1 ._SENT Riegl_NP0 's_POS collection_NN1 platform_NN1 uses_VVZ an_AT0 internal_AJ0 '_POS multiple-time-around_NN1 '_POS procedure_NN1 ,_PUN which_DTQ eliminates_VVZ range_NN1 ambiguities_NN2 and_CJC helps_NN2 remove_VVB data_NN0 noise_NN1 ._SENT 
A_AT0 four-mile_AJ0 segment_NN1 was_VBD surveyed_VVN along_PRP Montview_NP0 Boulevard_NP0 in_PRP Denver_NP0 ,_PUN Colorado_NP0 ,_PUN between_PRP Colorado_NP0 Boulevard_NP0 and_CJC Havana_NP0 Street_NP0 ._SENT An_AT0 additional_AJ0 nearby_AJ0 mile_NN1 of_PRF roadway_NN1 through_PRP a_AT0 mixed-use_AJ0 town_NN1 center_NN1 area_NN1 along_PRP E_NP0 ._SENT 29th_ORD Avenue_NP0 in_PRP Denver_NP0 was_VBD also_AV0 collected_VVN ._SENT Each_DT0 of_PRF these_DT0 segments_NN2 was_VBD traveled_VVN twice_AV0 ,_PUN once_AV0 in_PRP each_DT0 direction_NN1 ._SENT Since_CJS many_DT0 features_NN2 of_PRF interest_NN1 are_VBB concentrated_VVN on_PRP the_AT0 sides_NN2 of_PRF streets_NN2 ,_PUN doing_VDG so_AV0 provides_VVZ additional_AJ0 assurance_NN1 for_PRP full_AJ0 collection_NN1 of_PRF features_NN2 ._SENT 
POSPac_AJ0 ®_NN0 was_VBD used_VVN to_TO0 process_VVI telemetry_NN1 of_PRF the_AT0 Global_AJ0 Navigation_NN1 Satellite_NN1 System_NN1 (_PUL GNSS_NP0 )_PUR system_NN1 ,_PUN and_CJC we_PNP used_VVD RTK_NP0 base_NN1 station_NN1 data_NN0 to_TO0 compute_VVI a_AT0 corrected_AJ0 location_NN1 of_PRF the_AT0 vehicle_NN1 path_NN1 ._SENT Trimble_NP0 Business_NN1 Center_NN1 (_PUL TBC_NP0 )_PUR software_NN1 was_VBD used_VVN to_TO0 register_VVI /_PUN adjust_VVB overlapping_AJ0 data_NN0 from_PRP the_AT0 multiple_AJ0 runs_NN2 ._SENT TBC_NP0 also_AV0 computed_VVN point_NN1 cloud_NN1 colorization_NN1 using_VVG the_AT0 registered_AJ0 images_NN2 that_CJT were_VBD collected_VVN in-sync_AJ0 with_PRP the_AT0 scans_NN2 ._SENT 
The_AT0 combined_AJ0 runs_NN2 were_VBD exported_VVN into_PRP LAS_NP0 1.4_CRD format_NN1 and_CJC 100-ft_AJ0 by_PRP 100-ft_AJ0 tiles_NN2 ._SENT A_AT0 control_NN1 report_NN1 was_VBD run_VVN against_PRP the_AT0 GCPs_NN2 that_CJT returned_VVD a_AT0 root_NN1 mean_VVB square_AJ0 error_NN1 (_PUL RMSE_NP0 )_PUR of_PRF 0.06-ft_AJ0 /_PUN 1.83-cm_AJ0 RMSE_NN1 horizontal_AJ0 and_CJC a_AT0 0.122-ft_AJ0 /_PUN 3.72-cm_AJ0 RMSE_NN1 vertical_AJ0 ._SENT Point_NN1 cloud_NN1 density_NN1 consistently_AV0 averaged_VVD over_AVP 2000_CRD points_NN2 per_PRP square_AJ0 meter_NN1 ._SENT Target_NN1 range_NN1 was_VBD well_AV0 over_PRP 100m_NN0 ,_PUN which_DTQ was_VBD well_AV0 above_PRP the_AT0 tallest_AJS feature_NN1 in_PRP the_AT0 streetscapes_NN2 ._SENT 
We_PNP also_AV0 obtained_VVD parcel_NN1 data_NN0 from_PRP the_AT0 City_NN1 and_CJC County_NN1 of_PRF Denver_NP0 to_TO0 determine_VVI the_AT0 right-of-way_NN1 (_PUL ROW_NN1 )_PUR for_PRP the_AT0 Montview_NP0 segment_NN1 ,_PUN as_CJS discussed_VVN below_AV0 ._SENT 
Methods_NN2 
Creating_VVG streetscape_NN1 sample_NN1 areas_NN2 
Similar_AJ0 to_PRP previous_AJ0 studies_NN2 on_PRP quantitative_AJ0 streetscape_NN1 measuring_VVG (_PUL Golombek_NP0 and_CJC Marshall_NP0 2019_CRD ,_PUN 2020_CRD )_PUR ,_PUN we_PNP seek_VVB to_TO0 develop_VVI quantitative_AJ0 statistics_NN0 of_PRF streetscape_NN1 features_NN2 ._SENT This_DT0 first_ORD required_AJ0 defining_AJ0 streetscape_NN1 corridor_NN1 segments_NN2 ._SENT Recent_AJ0 streetscape_NN1 mapping_NN1 studies_NN2 used_VVD a_AT0 method_NN1 to_TO0 divide_VVI the_AT0 pre-determined_AJ0 segments_NN2 into_PRP Thiessen_NP0 proximal_AJ0 polygons_NN2 (_PUL Golombek_NP0 and_CJC Marshall_NP0 2019_CRD ,_PUN 2020_CRD )_PUR ._SENT To_TO0 create_VVI a_AT0 Thiessen_NP0 polygon_NN1 layer_NN1 ,_PUN only_AV0 a_AT0 point_NN1 feature_NN1 class_NN1 is_VBZ required_VVN ,_PUN and_CJC the_AT0 output_NN1 Thiessen_NN1 polygon_NN1 dataset_NN1 is_VBZ configured_VVN where_CJS each_DT0 individual_AJ0 non-overlapping_AJ0 polygon_NN1 is_VBZ closest_AJS to_PRP its_DPS associated_AJ0 input_NN1 point_NN1 ._SENT In_PRP the_AT0 Montview_NP0 dataset_NN1 ,_PUN we_PNP attempted_VVD to_TO0 avoid_VVI segment_NN1 cutoff_NN1 at_PRP intersections_NN2 since_PRP urban_AJ0 intersections_NN2 tend_VVB to_TO0 be_VBI focal_AJ0 areas_NN2 with_PRP high_AJ0 activity_NN1 and_CJC many_DT0 traffic-related_AJ0 features_NN2 we_PNP prefer_VVB to_TO0 keep_VVI together_AV0 ._SENT By_PRP establishing_VVG the_AT0 input_NN1 point_NN1 feature_NN1 class_NN1 with_PRP these_DT0 intersection_NN1 centroids_NN2 ,_PUN the_AT0 segment_NN1 breaks_NN2 occur_VVB at_PRP a_AT0 subtle_AJ0 location_NN1 somewhere_AV0 between_PRP these_DT0 intersections_NN2 ._SENT Figure_NN1 2_CRD depicts_VVZ the_AT0 Thiessen-based_AJ0 streetscape_NN1 segments_NN2 we_PNP created_VVD for_PRP Montview_NP0 Boulevard_NP0 using_VVG the_AT0 ESRI_NP0 ArcGIS_NP0 ®_NN0 Thiessen_NP0 polygon-processing_NN1 tool_NN1 ._SENT The_AT0 town-center_NN1 streetscape_NN1 polygons_NN2 were_VBD created_VVN manually_AV0 since_PRP the_AT0 area_NN1 of_PRF interest_NN1 contained_VVD only_AV0 a_AT0 few_DT0 small_AJ0 blocks_NN2 ._SENT 
Additionally_AV0 for_PRP the_AT0 Montview_NP0 section_NN1 ,_PUN we_PNP designed_VVD the_AT0 streetscape_NN1 sample_NN1 areas_NN2 to_TO0 include_VVI the_AT0 full_AJ0 street_NN1 right-of-way_NN1 (_PUL ROW_NN1 )_PUR that_CJT extend_VVB beyond_PRP the_AT0 street_NN1 curb_NN1 ._SENT Extending_VVG the_AT0 streetscape_NN1 to_TO0 include_VVI these_DT0 peripheral_AJ0 views_NN2 is_VBZ important_AJ0 since_PRP the_AT0 majority_NN1 of_PRF objects_NN2 in_PRP urban_AJ0 roadways_NN2 that_CJT fall_VVB within_PRP a_AT0 road_NN1 user_NN1 's_POS peripheral_AJ0 view_NN1 are_VBB located_VVN off_PRP the_AT0 street_NN1 itself_PNX ._SENT Therefore_AV0 ,_PUN we_PNP utilized_VVD Denver_NP0 's_POS parcel_NN1 data_NN0 around_PRP the_AT0 area_NN1 of_PRF interest_NN1 (_PUL AOI_NP0 )_PUR ,_PUN merged_VVD the_AT0 parcel_NN1 data_NN0 together_AV0 ,_PUN and_CJC used_VVD an_AT0 ESRI_NP0 ArcGIS_NP0 ®_NN0 erase_VVB function_NN1 that_CJT erased_VVD all_DT0 non-streetscape_AJ0 ROW_NN1 areas_NN2 from_PRP our_DPS AOI_NP0 ._SENT For_PRP the_AT0 Montview_NP0 section_NN1 ,_PUN we_PNP created_VVD six_CRD streetscape_NN1 polygon_NN1 segments_NN2 utilizing_VVG this_DT0 method_NN1 ._SENT 
For_PRP the_AT0 town_NN1 center_NN1 area_NN1 ,_PUN our_DPS approach_NN1 was_VBD a_AT0 little_AV0 different_AJ0 ._SENT Since_CJS the_AT0 focus_NN1 is_VBZ more_AV0 on_PRP streetscape_NN1 features_NN2 related_VVN to_PRP walkability_NN1 ,_PUN we_PNP simply_AV0 digitized_VVD the_AT0 areas_NN2 between_PRP the_AT0 curb_NN1 and_CJC the_AT0 building_NN1 ,_PUN enclosed_VVN with_PRP a_AT0 crosswalk_NN1 on_PRP each_DT0 side_NN1 ._SENT We_PNP identified_VVD six_CRD sample_NN1 streetscape_NN1 areas_NN2 ,_PUN as_CJS shown_VVN on_PRP Fig._NN1 2_CRD ,_PUN that_CJT we_PNP assess_VVB in_PRP this_DT0 study_NN1 ._SENT 
Creating_VVG streetscape_NN1 voxels_NN2 for_PRP positional_AJ0 feature_NN1 analysis_NN1 
A_AT0 key_AJ0 objective_NN1 of_PRF this_DT0 study_NN1 is_VBZ to_TO0 incorporate_VVI voxels_NN2 to_PRP quantity_NN1 streetscape_NN1 features_NN2 and_CJC compile_VVB descriptive_AJ0 statistics_NN0 ._SENT Breaking_VVG up_AVP a_AT0 streetscape_NN1 into_PRP voxels_NN2 creates_VVZ user-specified_AJ0 height_NN1 zones_NN2 ._SENT A_AT0 voxel_NN1 is_VBZ a_AT0 pixel_NN1 with_PRP a_AT0 height_NN1 or_CJC third-dimension_AJ0 component_NN1 that_CJT can_VM0 be_VBI useful_AJ0 for_PRP analyzing_VVG LiDAR_NP0 data_NN0 ._SENT Similar_AJ0 voxel_NN1 methods_NN2 are_VBB utilized_VVN in_PRP our_DPS previous_AJ0 street-feature_NN1 voxel_NN1 analysis_NN1 studies_NN2 (_PUL Golombek_NP0 and_CJC Marshall_NP0 2019_CRD ,_PUN 2020_CRD )_PUR ._SENT In_PRP those_DT0 papers_NN2 ,_PUN we_PNP determined_VVD that_CJT —_PUN given_VVN aerial_AJ0 LiDAR_NN1 point_NN1 densities_NN2 —_PUN a_AT0 3-ft_AJ0 by_PRP 3-ft_AJ0 horizontal_AJ0 and_CJC a_AT0 5-ft_AJ0 vertical_AJ0 setting_NN1 was_VBD optimal_AJ0 and_CJC appropriate_AJ0 ._SENT 
The_AT0 accuracy_NN1 and_CJC density_NN1 of_PRF mobile_AJ0 LiDAR_NP0 data_NN0 ,_PUN however_AV0 ,_PUN allows_VVZ more_DT0 flexibility_NN1 in_PRP setting_VVG voxel_NN1 parameters_NN2 ._SENT Lehtomaki_NP0 et_NP0 al._NP0 used_VVD a_AT0 fix_NN1 voxel_NN1 size_NN1 of_PRF 20_CRD cm3_CRD for_PRP their_DPS street_NN1 feature_NN1 study_NN1 ._SENT Yang_NP0 et_NP0 al._NP0 used_VVD a_AT0 super-voxel_AJ0 approach_NN1 where_AVQ the_AT0 voxel_NN1 size_NN1 was_VBD more_AV0 focused_VVN on_PRP the_AT0 feature_NN1 itself_PNX ,_PUN though_CJS their_DPS voxel_NN1 size_NN1 always_AV0 seemed_VVD greater_AJC than_CJS 0.5_CRD m3_CRD ._SENT With_PRP over_AV0 a_AT0 thousand_CRD points_NN2 per_PRP meter_NN1 collected_VVN and_CJC utilizing_VVG US_NP0 feet_NN2 as_PRP the_AT0 linear_AJ0 unit_NN1 ,_PUN we_PNP suggest_VVB setting_VVG each_DT0 voxel_NN1 to_PRP a_AT0 single_AJ0 cubic_AJ0 foot_NN1 ,_PUN or_CJC 1-ft_AJ0 by_PRP 1-ft_AJ0 by_PRP 1-ft_AJ0 (_PUL x_ZZ0 ,_PUN y_ZZ0 ,_PUN z_ZZ0 )_PUR dimension_NN1 ._SENT We_PNP suggest_VVB this_DT0 parameter_NN1 because_CJS this_DT0 dimension_NN1 facilitates_VVZ a_AT0 simple_AJ0 calculating_AJ0 dimension_NN1 and_CJC solid_AJ0 measuring_AJ0 basis_NN1 for_PRP descriptive_AJ0 statistic_NN1 metrics_NN2 needed_VVN in_PRP the_AT0 town_NN1 center_NN1 area_NN1 ._SENT Ewing_NP0 and_CJC Clemente_NP0 often_AV0 reference_VVB measuring_VVG features_NN2 from_PRP ground_NN1 to_PRP certain_AJ0 levels_NN2 ,_PUN like_PRP eye-level_AJ0 visibility_NN1 (_PUL Ewing_NP0 and_CJC Clemente_NP0 2013_CRD )_PUR ._SENT If_CJS a_AT0 streetscape_NN1 was_VBD ,_PUN for_PRP example_NN1 ,_PUN 500-ft_AJ0 long_AV0 from_PRP crosswalk_NN1 to_PRP crosswalk_NN1 ,_PUN 50-ft_AJ0 wide_AJ0 from_PRP storefront_NN1 /_PUN building_VVG to_TO0 curb_VVI ,_PUN and_CJC 15-ft_AJ0 tall_AJ0 from_PRP ground_NN1 to_PRP the_AT0 above_AJ0 awnings_NN2 ,_PUN that_DT0 streetscape_NN1 quantitatively_AV0 would_VM0 total_AJ0 375_CRD ,_PUN 000_CRD cubic_AJ0 ft._NN0 If_CJS we_PNP calculate_VVB that_CJT 10_CRD landscape_NN1 trees_NN2 fill_VVB 7500_CRD single_AJ0 cubic-foot_AJ0 voxels_NN2 ,_PUN then_AV0 we_PNP can_VM0 estimate_VVI that_CJT trees_NN2 comprise_VVB 2_CRD percent_NN0 of_PRF that_DT0 streetscape_NN1 ._SENT 
LiDAR_NP0 feature_NN1 extraction_NN1 methodology_NN1 /_PUN classifying_VVG the_AT0 data_NN0 
For_PRP the_AT0 four-mile_AJ0 Montview_NN1 segment_NN1 ,_PUN the_AT0 goal_NN1 is_VBZ to_TO0 devise_VVI comprehensive_AJ0 descriptive_AJ0 statistics_NN0 of_PRF streetscape_NN1 features_NN2 that_CJT can_VM0 be_VBI applied_VVN to_PRP various_AJ0 urban_AJ0 transportation_NN1 research_NN1 studies_NN2 ._SENT Other_AJ0 than_CJS for_PRP street_NN1 sign_NN1 extraction_NN1 ,_PUN the_AT0 existing_AJ0 literature_NN1 did_VDD not_XX0 provide_VVI viable_AJ0 automated_AJ0 approaches_NN2 ._SENT Even_AV0 the_AT0 street_NN1 sign_NN1 extraction_NN1 methods_NN2 yielded_VVD results_NN2 well_AV0 below_AV0 100_CRD %_NN0 (_PUL with_PRP a_AT0 high_AJ0 degree_NN1 of_PRF false_AJ0 positives_NN2 and_CJC false_AJ0 negatives_NN2 )_PUR and_CJC have_VHB not_XX0 been_VBN thoroughly_AV0 tested_VVN at_PRP a_AT0 macro_AJ0 scale_NN1 ._SENT Gargoum_NP0 et_NP0 al._NP0 developed_VVD highly_AV0 accurate_AJ0 automated_AJ0 street_NN1 sign_NN1 extraction_NN1 results_NN2 (_PUL Gargoum_NP0 et_NP0 al._NN1 2018_CRD )_PUR ,_PUN though_CJS testing_NN1 was_VBD primarily_AV0 in_PRP rural_AJ0 areas_NN2 where_AVQ signs_NN2 are_VBB isolated_VVN ._SENT Even_AV0 if_CJS we_PNP utilize_VVB published_AJ0 automated_AJ0 methods_NN2 ,_PUN we_PNP would_VM0 still_AV0 have_VHI to_TO0 manually_AV0 assess_VVI the_AT0 entire_AJ0 streetscape_NN1 to_TO0 check_VVI for_PRP false_AJ0 positives_NN2 and_CJC false_AJ0 negatives_NN2 ._SENT We_PNP also_AV0 note_VVB ,_PUN after_PRP previewing_VVG the_AT0 Montview_NP0 section_NN1 ,_PUN that_CJT many_DT0 street_NN1 signs_NN2 are_VBB on_PRP the_AT0 same_DT0 horizontal_AJ0 poles_NN2 as_PRP traffic_NN1 lights_NN2 ,_PUN and_CJC some_DT0 are_VBB small_AJ0 and_CJC /_PUN or_CJC obscured_VVN ._SENT For_PRP example_NN1 in_PRP Fig._NN1 3_CRD ,_PUN a_AT0 '_PUQ No_NN1 Parking_NN1 '_POS sign_NN1 may_VM0 be_VBI on_PRP a_AT0 light_AJ0 pole_NN1 or_CJC power_NN1 pole_NN1 ,_PUN but_CJC if_CJS it_PNP is_VBZ not_XX0 wider_AJC than_CJS the_AT0 pole_NN1 ,_PUN the_AT0 current_AJ0 automated_AJ0 methods_NN2 would_VM0 likely_AV0 miss_VVI these_DT0 signs_NN2 ._SENT Additionally_AV0 ,_PUN the_AT0 streetscape_NN1 furniture_NN1 and_CJC traffic_NN1 lights_NN2 require_VVB manual_AJ0 classification_NN1 since_CJS we_PNP did_VDD not_XX0 find_VVI adequate_AJ0 methods_NN2 to_TO0 automate_VVI extracting_VVG those_DT0 features_NN2 ._SENT An_AT0 accuracy_NN1 assessment_NN1 on_PRP manually_AV0 classified_AJ0 data_NN0 is_VBZ not_XX0 necessary_AJ0 since_PRP many_DT0 automated_AJ0 LiDAR_NN1 classification_NN1 experiments_NN2 ,_PUN including_PRP ones_NN2 listed_VVN above_PRP ,_PUN utilize_VVB manual_AJ0 classification_NN1 for_PRP their_DPS control_NN1 to_TO0 assess_VVI accuracy_NN1 ._SENT The_AT0 accuracy_NN1 specifications_NN2 of_PRF the_AT0 data_NN0 are_VBB mentioned_VVN with_PRP the_AT0 data_NN0 collection_NN1 platform_NN1 specifications_NN2 above_AV0 ._SENT 
Since_CJS our_DPS four-mile_AJ0 sample_NN1 area_NN1 is_VBZ not_XX0 overwhelmingly_AV0 large_AJ0 ,_PUN we_PNP utilized_VVD a_AT0 manual_AJ0 classification_NN1 approach_NN1 with_PRP some_DT0 automated_AJ0 LiDAR_NN1 functions_NN2 ._SENT The_AT0 data_NN0 's_POS control_NN1 reports_NN2 ,_PUN noted_VVD above_AV0 ,_PUN confirmed_VVD very_AV0 high_AV0 vertical_AJ0 and_CJC horizontal_AJ0 point-cloud_NN1 accuracy_NN1 ._SENT Also_AV0 ,_PUN automated_AJ0 LiDAR_NN1 methods_NN2 are_VBB often_AV0 tested_VVN against_PRP manual_AJ0 classification_NN1 for_PRP control_NN1 ,_PUN and_CJC we_PNP saw_VVD the_AT0 same_DT0 with_PRP automated_AJ0 urban_AJ0 mobile_AJ0 LiDAR_NN1 feature_NN1 extraction_NN1 (_PUL Yang_NP0 et_NP0 al._NN1 2015_CRD )_PUR ._SENT We_PNP auto-filtered_VVD the_AT0 data_NN0 for_PRP ground_NN1 points_NN2 and_CJC also_AV0 used_VVN an_AT0 automated_AJ0 process_NN1 to_TO0 classify_VVI the_AT0 remaining_AJ0 above-ground_AJ0 features_NN2 together_AV0 on_PRP a_AT0 single_AJ0 LiDAR_NP0 classification_NN1 layer_NN1 ._SENT We_PNP utilized_VVD Merrick_NP0 &_NP0 Company_NN1 's_POS MARS_NP0 ®_NN0 LiDAR_NP0 processing_NN1 software_NN1 ,_PUN which_DTQ has_VHZ a_AT0 Google_NP0 Earth_NN1 simultaneous_AJ0 location_NN1 feature_NN1 ,_PUN to_TO0 follow_VVI the_AT0 streetscape_NN1 point-cloud_NN1 in_PRP MARS_NP0 ®_NN0 ._SENT Paired_VVN with_PRP Google_NP0 Earth_NN1 on_PRP a_AT0 separate_AJ0 screen_NN1 ,_PUN this_DT0 allowed_VVD us_PNP to_TO0 visually_AV0 identify_VVI all_DT0 street_NN1 signs_NN2 ,_PUN traffic_NN1 signals_NN2 ,_PUN and_CJC street_NN1 furniture_NN1 features_NN2 (_PUL Fig._NN1 4_CRD )_PUR ._SENT 
Once_AV0 completed_VVN ,_PUN we_PNP utilized_VVD a_AT0 voxel_AJ0 application_NN1 tool_NN1 in_PRP MARS_NP0 ®_NN0 that_CJT incorporates_VVZ ground_NN1 level_NN1 and_CJC builds_VVZ a_AT0 user-specified_AJ0 voxel_NN1 grid_NN1 set_VVN to_PRP 1-ft_AJ0 by_PRP 1-ft_AJ0 by_PRP 1-ft_AJ0 (_PUL x_ZZ0 ,_PUN y_ZZ0 ,_PUN z_ZZ0 )_PUR dimensions_NN2 ._SENT This_DT0 tool_NN1 then_AV0 applies_VVZ the_AT0 selected_AJ0 classified_AJ0 street_NN1 furniture_NN1 and_CJC street_NN1 sign_NN1 features_NN2 into_PRP the_AT0 voxel_NN1 grids_NN2 and_CJC exports_NN2 a_AT0 .dat_NN1 Envi_NN1 raster_NN1 dataset_NN1 ,_PUN with_PRP each_DT0 band_NN1 within_PRP the_AT0 .dat_AJ0 file_NN1 representing_VVG a_AT0 unique_AJ0 user-specified_AJ0 voxel_NN1 height_NN1 zone_NN1 ._SENT Each_DT0 grid_NN1 cell_NN1 that_DT0 is_VBZ populated_VVN also_AV0 has_VHZ a_AT0 count_NN1 value_NN1 that_CJT specifies_VVZ how_AVQ many_DT0 LiDAR_NP0 points_NN2 fell_VVD into_PRP that_DT0 3D_NN0 voxel_NN1 grid_NN1 cell_NN1 ._SENT For_PRP this_DT0 study_NN1 ,_PUN we_PNP are_VBB not_XX0 very_AV0 concerned_AJ0 with_PRP the_AT0 amount_NN1 of_PRF data_NN0 in_PRP each_DT0 grid_NN1 cell_NN1 ,_PUN but_CJC rather_AV0 the_AT0 simple_AJ0 fact_NN1 that_CJT classified_AJ0 LAS_NP0 points_NN2 are_VBB inside_AV0 that_DT0 grid_NN1 cell_NN1 ._SENT Table_NN1 1_CRD presents_VVZ a_AT0 sample_NN1 descriptive_AJ0 statistics_NN0 output_NN1 ._SENT 
For_PRP the_AT0 town_NN1 center_NN1 area_NN1 ,_PUN our_DPS descriptive_AJ0 statistic_NN1 zones_NN2 for_PRP the_AT0 walkable_AJ0 streets_NN2 are_VBB different_AJ0 because_CJS we_PNP focus_VVB more_AV0 on_PRP the_AT0 walkable_AJ0 areas_NN2 between_PRP the_AT0 curb_NN1 and_CJC the_AT0 buildings_NN2 ._SENT For_PRP this_DT0 ,_PUN we_PNP use_VVB a_AT0 different_AJ0 approach_NN1 because_CJS we_PNP are_VBB interested_AJ0 in_PRP collecting_VVG and_CJC measuring_VVG features_NN2 affiliated_VVN with_PRP the_AT0 buildings_NN2 themselves_PNX such_DT0 as_CJS the_AT0 windows_NN2 ,_PUN awnings_NN2 ,_PUN and_CJC hanging_VVG signs_NN2 ._SENT Our_DPS approach_NN1 for_PRP collecting_VVG these_DT0 objects_NN2 differs_VVZ because_CJS flat_AJ0 objects_NN2 against_PRP buildings_NN2 ,_PUN such_DT0 as_PRP windows_NN2 ,_PUN cannot_NN1 be_VBI extracted_VVN like_PRP typical_AJ0 streetscape_NN1 objects_NN2 ._SENT Also_AV0 ,_PUN LiDAR_NP0 often_AV0 gets_VVZ obstructed_VVN at_PRP windows_NN2 and_CJC does_VDZ not_XX0 yield_VVI returns_NN2 ._SENT 
We_PNP extracted_VVD windows_NN2 and_CJC awnings_NN2 with_PRP mobile_AJ0 LiDAR_NN1 with_PRP the_AT0 help_NN1 of_PRF a_AT0 web-based_AJ0 solution_NN1 by_PRP Orbit_NN1 GT_NP0 technologies_NN2 ._SENT The_AT0 Trimble_NP0 MX9_AJ0 platform_NN1 collected_VVD panoramic_AJ0 images_NN2 in_PRP conjunction_NN1 with_PRP the_AT0 LiDAR_NN1 ._SENT Since_CJS the_AT0 frames_NN2 of_PRF the_AT0 windows_NN2 are_VBB well_AV0 defined_VVN based_VVN on_PRP the_AT0 void_NN1 of_PRF LiDAR_NP0 points_NN2 within_PRP the_AT0 frames_NN2 ,_PUN the_AT0 Orbit_NN1 LiDAR_NN1 software_NN1 has_VHZ an_AT0 interface_NN1 that_CJT allows_VVZ 3D_NN0 feature_NN1 tracing_NN1 of_PRF the_AT0 LiDAR_NP0 data_NN0 ._SENT For_PRP windows_NN2 and_CJC awnings_NN2 ,_PUN we_PNP essentially_AV0 created_VVD vertical_AJ0 polygons_NN2 ._SENT After_PRP the_AT0 windows_NN2 and_CJC awnings_NN2 were_VBD drawn_VVN in_PRP the_AT0 3D_NN0 Orbit_NN1 software_NN1 interface_NN1 ,_PUN it_PNP exported_VVD the_AT0 polygons_NN2 into_PRP a_AT0 Google_NP0 KML_NP0 file_NN1 ,_PUN which_DTQ were_VBD then_AV0 converted_VVN into_PRP a_AT0 3D_NN0 shapefiles_NN2 ._SENT Since_CJS the_AT0 features_NN2 are_VBB in_PRP 3D_NN0 ,_PUN calculating_VVG the_AT0 area_NN1 of_PRF these_DT0 polygons_NN2 must_VM0 be_VBI captured_VVN in_PRP 3D_NN0 as_CJS well_AV0 ._SENT For_PRP this_DT0 ,_PUN we_PNP created_VVD an_AT0 automated_AJ0 model_NN1 in_PRP ESRI_NP0 's_POS ®_NN0 Model_NN1 Builder_NN1 to_TO0 iterate_VVI through_PRP each_DT0 polygon_NN1 feature_NN1 and_CJC created_VVD a_AT0 temporary_AJ0 Triangulated_AJ0 Irregular_AJ0 Network_NN1 (_PUL TIN_NN1 )_PUR for_PRP each_DT0 window_NN1 or_CJC awning_NN1 polygon_NN1 ._SENT We_PNP then_AV0 ran_VVD surface_NN1 information_NN1 statistics_NN0 on_PRP each_DT0 polygon_NN1 ,_PUN utilizing_VVG the_AT0 TIN_NN1 as_PRP the_AT0 surface_NN1 ,_PUN which_DTQ enables_VVZ the_AT0 area_NN1 of_PRF each_DT0 polygon_NN1 to_TO0 be_VBI calculated_VVN ._SENT This_DT0 process_NN1 created_VVN area_NN1 statistics_NN0 for_PRP features_NN2 like_PRP windows_NN2 that_CJT can_VM0 detected_VVN with_PRP LiDAR_NP0 but_CJC not_XX0 directly_AV0 extracted_VVN and_CJC classified_VVN ._SENT As_CJS mentioned_VVN above_AV0 ,_PUN urban_AJ0 design_NN1 researchers_NN2 such_DT0 as_PRP Ewing_NP0 et_NP0 al._NP0 take_VVB into_PRP account_NN1 quantity_NN1 of_PRF features_NN2 at_PRP eye_NN1 level_NN1 for_PRP perceptual_AJ0 qualities_NN2 ._SENT Utilizing_VVG data_NN0 in_PRP the_AT0 LiDAR_NP0 viewer_NN1 allows_VVZ an_AT0 analyst_NN1 to_TO0 calculate_VVI 2D_NN0 face_NN1 area_NN1 of_PRF a_AT0 building_NN1 (_PUL from_PRP ground_NN1 to_PRP the_AT0 awning_NN1 )_PUR ,_PUN and_CJC after_PRP the_AT0 TIN_NN1 process_NN1 for_PRP windows_NN2 is_VBZ complete_AJ0 ,_PUN area_NN1 and_CJC proportion_NN1 of_PRF windows_NN2 were_VBD calculated_VVN ._SENT 
