Authors: Xuelian Meng*, Louisiana State University, Nan Shang, Louisiana State University, Xukai Zhang, Louisiana State University, Chunyan Li, Louisiana State University, Kaiguang Zhao, Ohio State University, Xiaomin Qiu, Missouri State University, Eddie Weeks, Louisiana State University
Topics: Remote Sensing, Geomorphology, Coastal and Marine
Keywords: photogrammetric UAV; super high resolution; coastal topographic mapping; wetlands; terrain correction; object oriented analysis
Session Type: Paper
Start / End Time: 8:00 AM / 9:40 AM
Room: Poydras, Sheraton, 3rd Floor
Presentation File: No File Uploaded
Photogrammetric UAV sees a surge in use for high-resolution mapping, but its use to map terrain under dense vegetation cover is challenging due to a lack of exposed ground surfaces. We presented a novel method to leverage morphological, texture and contextual information of UAV data to improve landscape classification through a set of hybrid and analytical algorithms. Its implementation incorporates multiple heuristics, such as multi-pass machine learning-based classification, object-oriented analytical schemes, and integration of GPS samples for terrain correction. Experiments based on a densely vegetated coastal landscape showed classification improvement from 83.98% to 96.12% in overall accuracy. Use of standard and existing UAV terrain mapping algorithms and software produced reliable digital terrain model only over exposed bare grounds (mean error=-0.019 and RMSE=0.035 m) but severely overestimated the terrain by ~80% of mean vegetation height. The terrain correction method successfully reduced the mean error from 0.302 m to -0.002 m (RMSE errors from 0.342 m to 0.177 m) in low vegetation and from 1.305 m to 0.057 m (RMSE from 1.399 m to 0.550 m) in tall vegetation. Overall, this research validated a feasible solution to integrate UAV and RTK GPS for terrain mapping in densely vegetated environments. In addition, this paper introduces an innovative pixel-based object-oriented method that allows seamless transition between pixel-based operations and object-oriented statistics without the constraint of object size.