1. Initial results from European Road Safety Inspection (EuRSI) Mobile Mapping Project Conor Mc Elhinney, P. Kumar, C. Cahalane and T. McCarthy Wednesday23rd June 2010
102. Navigation GPS What we get out of the GPS We get highly accurate Longitude, Latitude and Altitude information.
103. Navigation GPS What we get out of the GPS We get highly accurate Longitude, Latitude and Altitude information.
104. Navigation INS What we get out of the INS We get highly accurate Roll, Pitch, Yaw and with the DMI the distance travelled This allows us in postprocessing to “fill in the gaps”, and correct for errors in the GPS reliably.
105. Navigation INS What we get out of the INS We get highly accurate Roll, Pitch, Yaw and with the DMI the distance travelled This allows us in postprocessing to “fill in the gaps”, and correct for errors in the GPS reliably. But its important to look in more detail at the navigation
106. Navigation in detail Take a section of road that we drove four times, twice up and twice down where the time between runs could be as much as 30 minutes.
107. Navigation in detail Take a section of road that we drove four times, twice up and twice down where the time between runs could be as much as 30 minutes. Focusing on one of the navigation lines in 3D we can see very small errors.
108. Navigation in detail Take a section of road that we drove four times, twice up and twice down where the time between runs could be as much as 30 minutes. Focusing on one of the navigation lines in 3D we can see very small errors.
119. LiDAR What we have Riegl VQ-250 Wavelength: 1550nm Rate: 300Khz Intensity: 16bit Scan Speed: 100 scans/s Field of View: 360 degrees What does it give us Intensity Range Echo type (multiple echoes, single echo, etc) Multiple other paramaters we’re trying to understand
120. MMS - Why? I think the video below demonstrates high data quality and the potential of MMS (thanks to StreetMapper for providing this)
128. Algorithm – Edge Extraction NAVIGATION Step 1: Calculate a line that is orthogonal to the vehicles heading + LIDAR
129. Algorithm – Edge Extraction NAVIGATION Step 2: Find points within distance μ of the line of length l. + LIDAR
130. Algorithm – Edge Extraction NAVIGATION Step 3: Extract the LIDAR points and their attributes + LIDAR
131. Algorithm – Edge Extraction NAVIGATION Step 4: Fit a 2D spline to the elevation and northing/easting data. + LIDAR
132. Algorithm – Edge Extraction NAVIGATION Step 5: Calculate the slope and points of rapid change. + LIDAR
133. Algorithm – Edge Extraction Step 6: Find the nearest rapid change points left and right of the MMS position whose attributes fit our criteria. NAVIGATION + LIDAR
134. Algorithm – Edge Processing Local One thing that needs to be carried out is refinement of each returned road edge.
135. Algorithm – Edge Processing Local What we have started work on is processing a sequence of road edge points and removing outliers and refining the output Output from algorithm
136. Algorithm – Edge Processing Local What we have started work on is processing a sequence of road edge points and removing outliers and refining the output Remove outliers and big errors
137. Algorithm – Edge Processing Local What we have started work on is processing a sequence of road edge points and removing outliers and refining the output Remove small errors by fitting curves to the data
138. Algorithm – Edge Processing Local What we have started work on is processing a sequence of road edge points and removing outliers and refining the output
178. Ongoing Work Point by point refinement of the road edge extraction algorithm, prior to the current edge refinement. Use the edges and the curves to extract the road LiDAR points. Fit a surface to these points and automate the extraction of road geometry metrics. Integrate this data with the results from our partners in ITC (road side features). Develop our risk assessment matrix. Run this on 400km of MMS data from four countries........
179. Ongoing Work EuRSI Point by point refinement of the road edge extraction algorithm, prior to the current edge refinement. Use the edges and the curves to extract the road LiDAR points. Fit a surface to these points and automate the extraction of road geometry metrics. Integrate this data with the results from our partners in ITC (road side features). Develop our risk assessment matrix. Run this on 400km of MMS data from four countries........ Outside of EuRSI we are developing - region growing algorithms for LiDAR - developing formulas for the assessment of any MMS and the expected resolution of the point cloud - a database solution for MMS data and a few other things