ILMF2016
Monday
- Learned about the new technology: Geiger mode, single photon, Flash LiDAR.
- Harris Geiger has advanced 3 generations since the 2015 announcement.
- Real-time mapping is out there, particularly with Flash, mostly directed toward DoD operations.
- USGS standard beyond v1.2, 1.3 to 2.0 were discussed.
- Talked w/ Velodyne representative about Teki's UAV results.
- The horizontal errors are related to multiple scans in the spinning sensor + wide angle + IMU error.
- Talked w/ a Quantum Geospatial representative about the WGEW test data (did not reveal the name of our vendor). Asked about the quality of our test data relative to QL1. His opinion was that the data do not meet spec - but he could not be sure without seeing the data.
- Saw a lot of UAV platforms with LiDAR (Attack of the Drones!)
- Reigl
- Phoenix Aerial
- UASUSA
- YellowScan
- The carbon fiber flying trash can is +$300k, and can fly for +10 hours (jet fuel powered).
- Several backpack, wand, and dolley systems were on display.
- The Stencil wand and black tricorder (called a Contour) are pretty cool - $25-35k ($10k year license). Far less expensive than the Leica Pegasus Backpack
- Virtually all of the UAS systems are running Velodyne pucks, competitors include Reigl, Yellowscan. The Reigl drone runs its VMZ ($130k) - at 1Mhz scan frequency - so a couple orders of magnitude more data, but also more expensive.
- Talked w/ Chris Crosby about ongoing XSEDE and OpenTopography integration.
Tuesday
- Panel discussions on Geiger vs Single Photon vs Linear mode LiDAR
- USGS is close to accepting Geiger mode for the back end of 3DEP (likely final 4 years).
- Single Photon from Sigma Space was better than Geiger for cutting through veg and seeing ground w/ leaf-on.
- Linear mode tech is mature, the vendors seem to be making small advances, but nothing Earth shattering
- Improvement includes smart swath mapping (Optech Galaxy), multiple bands (Optech Titan), 30% increase in point sampling rate per year, e.g. Reigl yr1998 = 12khz, yr2016 = 1200khz.
- Reigl VZ-400i was shown with an 8 hour, 185 set up scan-a-thon, 65 GB data all stored internally.
- New vendors are cagey about the new tech, tension is palpable w/ old vendors.
- USGS Geospatial Product and Services Contract (GPSC).
- 3DEP program support.
- 3DEP availability
- 3DEP metadata.
- USGS use LP360 software
- LAS 1.4 has rigorous format specifications.
- Errors in LAS 1.4 delivery are common right now.
- Definitions are given in LBS v1.2 2014.
- OGC WKT projection format
- Overage vs overlap.
- Checkpoints vs control points
- DEM hydro-flattening.
- North Carolina mapping program is doing Geiger mode QL1 at the state level. Expect to have 62 TB raw point cloud.
- Hercules (HRQLS) Single Photon counting (green laser designed for terrestrial but can do bathymetry).
- Sigma Space has partnered with Leica.
- NIR LiDAR derived NDVI.
- Uses Intensity values for the NIR band?
- Demo was done using Global Mapper.
- TLS forest mapping methods
- Mentioned 3 UAS courses which comprise a certificate.
- Urban forests
- Combined TLS and SfM (Photoscan)
- NEON site.
- Edited in CloudCompare
- Ground was classified in LASTools.
- RANSAC circle fitting.
- PhoDAR
- ESRI talk about SfM along power transmission lines.
- ESRI's Geonet
- ESRI Corridor management tools
- Drone2Map - unreleased tool for SfM in ArcGIS
- Beta release 2/24
- Extra $$ for license
- Uses Pix4D
- Shea and I spoke with Jayna Winehouse (USGS) about the WGEW Woolpert LiDAR. She is confident that the final data product will be QL1 and will meet contract spec. We also asked to speak with a Woolpert representative.
- Jayna mentioned that the WGEW lidar data might be ready for delivery as early as next Monday.
Wednesday
- Multi spectral LiDAR image classification
- Optech Titan (3 bands)
- Bands are delivered as separate .las
- Used LAStools, ENVI, MATLAB.
- MATLAB TSTtool
- Geobia w/ multi spectral + LiDAR
- Used eCognition.
- LiDAR vs SfM (Phodar)
- Phodar goes through a rigorous photogrammetry process vs SfM which has less orthophotographic detail (camera orientations, GPS positions, etc).
- Hexagon Image Processing
- LAS 1.4 supports 4 band imagery (NIR-RGB)
6" imagery + NAIP
- 80-60% image overlap
- First pass DEM w/ LiDAR, follow up w/ SfM.
- Shadowing effects + vegetation movement w/ wind.
- RMSE 13-30 cm.
- Needs good illumination.
- Good for vegetation measurement, poor for terrain modelling.
- LiDAR data fusion using Open Geospatial Consortium (OGC)
- Pixia software
- OGC OpenStreetMaps
- OGC:KML
- OGC:WCS
- Shared online, sampling
- Spoke w/ Amanda Lowe and Kim Mantey (USGS) about WGEW LiDAR acquisition, las 1.4 data formats, full waveform, and hydro-flattening.
- It's possible that my reverse versioning of 1.4 caused a point loss from the pilot data.
- We need to ask Woolpert for the full waveform
- Shea asked a question about the hydro-flattening and eventual merge with national level data.
- LAS 1.4
- Evon Silvia Quantum Geospatial.
- November 2014 USGS/ASPRS enforce standard.
- 64-bit
- Las 1.2 was too small and has been exceeded.
- Allows storage of custom bands to the point cloud, e.g. NDVI, TIR.
- Area under the curve in the waveform
- Refraction, angle of incidence (Bathymetry).
- Feature extraction
- E.g. vegetation classification, buildings
- Encoding features into the point cloud.
- Software has not been updated to adapt to it!
- Extended Variable Length Records (EVLR) can hold additional data, e.g. pulsewaves, trajectory.
- Eliminates supplemental data files.
, multiple selections available,