Planning Efficient Unmanned Aerial Surveys
Item statusRestricted Access
The current method for producing DSMs from UAVs can result in datasets with several gigabytes of data, requiring hours, days, and sometimes weeks to process. Additionally, many UAVs rely on small lightweight fisheye cameras with heavy barrel distortion which introduces errors into the DSMs. This study examines the impact on the quality of DSMs produced from optimization techniques and from images which have undergone distortion correction using the camera calibration and the OpenCV library. It further investigates the placement of ground control points to further decrease the workflow. Two autonomous aerial surveys were conducted over a 120m by 80m area, at a height of 30m and 100m using an autonomous UAV and photogrammetric calculator. Eight DSMs were produced to evaluate the workflow using structure from motion software. Results show that GCP setup can be reduced by a 40m increase in grid spacing, reducing the quality of the dense point cloud has no impact on the quality of the DSM, distortion correction introduces errors into DSMs, and datasets can be reduced by 72% by increasing flight heights from 30m to 100m without compromising the quality of DSMs.