Laserfiche WebLink
Data Collection Methodology <br /> The larger of the SL25 parcels, the D Pit parcel (DAB23) comprising 286.9 acres was flown using a <br /> DJI Phantom 4 Pro drone, flying a single grid mission. With a vertically oriented camera, utilizing <br /> Pix4D Capture software running on an Apple iPad, the mission began on an overcast, windy <br /> morning. Settings were configured so that the drone would fly at 400 feet above the take-off point <br /> (on the southern boundary of the parcel) at full speed, and images would have standard front and side <br /> overlaps (70%). Under ideal conditions the mission should have been completed with four batteries, <br /> however the temperature was in the mid-30s, with brisk winds from the north-west, which meant that <br /> even with five batteries the full mission could not be completed. An attempt was made to cover the <br /> southern portion of the survey area using a DJI Mavic 2 Pro in an overlapping mission, however this <br /> drone was less able to deal with strong winds and there were issues stabilizing the camera gimbal. <br /> The smaller of the two SL25 parcels (IC23) at a lower elevation was covered easily using the Mavic <br /> 2 Pro, in a single mission with no issues. <br /> Propeller Aeropoints were used to collect Ground Control Points (i.e. points that are identifiable in <br /> the images and have a known location) during the survey. The Aeropoint data was uploaded to the <br /> Propeller portal and corrected a few days after collection. The corrected data was exported in <br /> NAD83(2011)/UTM zone 13N(EPSG 6342) coordinates, with a NAVD88 height(geoidl8)vertical <br /> datum. Following correction, the accuracy of the location of each point was typically a few <br /> millimeters in each axis. The quality reports are available for reference here: <br /> https:Hdnrweblink.state.co.us/dnns/welcome.aspx?cr=I <br /> Both datasets were processed in Pix4D Mapper, running on a Lenovo ThinkPad X1 laptop with a 12m <br /> generation Intel Core i7-1260P processor and 32GB of RAM. The large dataset was processed three <br /> times in order to optimize the output. As per the processing reports, the uncertainty ellipses of the <br /> images captured at the southern edge parcel DAB23 are very large. This is the area where the Mavic <br /> 2 Pro was used and data from three separate missions was merged. Numerous of these images were <br /> discarded reducing coverage. However, DRMS's ability to properly evaluate the area was not <br /> negatively impacted as described below. <br /> Processing details are given in the report, and summarized below: <br /> o The images were processed to give an orthomosaic and a simulated point cloud. <br /> o The point cloud was automatically classified. <br /> o The full point cloud, filtered to remove noise, was used to generate a Digital Surface Model <br /> (D SM) <br /> o The resolution of the orthomosaic and the DSM is equal to the ground sampling distance <br /> (GSD, a function of the camera sensor and the elevation above the ground) —this is estimated <br /> in the report at 6.38cm/pixel and varies across the survey area. <br /> o The classified point cloud was automatically filtered to remove points that were not classified <br /> as being on the ground. The filtered point cloud was used to generate a Digital Terrain Model <br /> (DTM)with a resolution of 5xGSD, so approximately 32 cm/pixel. <br /> For a quality assurance check, contours were <br /> generated from the DTM (digital terrain model) <br /> utilizing ESRI's ArcPro software and <br /> Permit Number C1981010 Prepared by: R.Reilley <br /> SL25 Page 9 of 21 <br />