ANALYSIS OF AIRBORNE LIDAR DATA FOR ROAD INVENTORY
ANALYSIS OF AIRBORNE LIDAR DATA FOR ROAD INVENTORY CLAY WOODS NIRDOSH GAIRE 6190 YI HE ZHAOCAI LIU 4/25/2016 CEE
OBJECTIVE The main objective of this research is to find an approach to detect and extract highway features, such as guardrails, medians, bridges, and large road signs, from airborne LIDAR data.
INTRODUCTION v. ROAD INVENTORY COMPILATION OF COMPONENTS AND CONDITIONS OF ROAD SYSTEM. v. METHODOLOGIES üFIELD INVENTORY üPHOTO/ VIDEO LOG üAERIAL/ SATELLITE PHOTOGRAPHY üTERRESTRIAL LIDAR üMOBILE LIDAR üAIRBORNE LIDAR
LIDAR TECHNOLOGY q LIDAR • LIGHT DETECTION AND RANGING (LIDAR) • REMOTE SENSING TECHNOLOGY THAT COLLECTS GEOMETRIC AND GEOGRAPHIC INFORMATION OF TARGETS ON THE EARTH’S SURFACE IN THE FORM OF POINT CLOUDS. q CLASSIFICATION ØTERRESTRIAL LASER SCANNING (TLS) ØMOBILE LASER SCANNING (MLS) ØAIRBORNE LASER SCANNING (ALS)
WHY LIDAR? ²HIGH DEGREE OF AUTOMATION ²SAFE OPERATION ²LESS AFFECTED BY ATMOSPHERE CONDITIONS ²EFFICIENT ²HIGH POST-PROCESSING EFFICIENCY
FIELD EXPERIMENT AND DATA COLLECTION l I-84– Mountain Green to Morgan County/Summit County l I-15 north– Payson to Springville l I-15 south– Region 2 l US-191
AIRBORNE LIDAR DATA PROCESSING
METHODOLOGY I-84 Rural US-191 I-15 North I-15 South Urban Guardrails Median Barriers Bridge Traffic Sign
METHODOLOGY Step 1: Find the Region of Interest Transform LAS file to raster file using ‘LAS Point Statistics as Raster’ tool, find the region of interest (road). Step 2: Road Feature Detection Evaluate the z-range each target feature falls into, and delete all the cells that are out of the range. Step 3: Digital Processing Manually digitize the found features in the raster files by creating feature classes. Step 4: Model Builder Automation of the algorithm.
STEP 1. FIND THE REGION OF INTEREST
Original LAS file Raster of road
STEP 2. ROAD FEATURE DETECTION Ø Calculate the elevation difference between the points that represent the target features and the points that represent ground from the profile view of the target features • guardrails and medians: approximately 0. 5 to 1 meter • bridge: approximately 7 to 10 meters • large traffic signs: approximately 10 to 11 meters Ø Remove all the cells whose values are out of range and get the candidate regions
Candidate guardrails and medians Candidate bridge Candidate traffic sign
STEP 3. DIGITAL PROCESSING
STEP 4. MODEL BUILDER • TASKS ARE REDUNDANT WITH MULTIPLE STEPS • SUITED FOR AUTOMATION • EXAMPLE IS FOR GUARDRAILS, BUT SIMILAR PROCESS FOR ANY FEATURE • MAY NEED TO ADJUST BUFFER SIZE WITH EACH ROAD
STEP 4. MODEL BUILDER
RESULTS AND DISCUSSION Bridge Extraction Results and Accuracy Evaluations for I-84
Sign Detection Results and Accuracy Evaluations for I-15
CONCLUSION Ø Airborne LIDAR technology is promising in detecting some highway features, such as guardrails, medians, bridges, and large road signs. Ø The proposed method worked well in detecting features from airborne LIDAR data. Ø In the future, we can improve our method to achieve higher accuracy.
THANK YOU! ANY QUESTIONS?
- Slides: 21