In order to join virtual sessions, you must be registered and logged-in(Were you registered for the in-person meeting in Denver? if yes, just log in.) 
Note: All session times are in Mountain Daylight Time.

Mapping Vegetation at Species Level with High-Resolution Multispectral and Lidar Data: A Case Study with Kudzu

Authors: Mongi Abidi, University of Tennessee, Knoxville, TN, USA, Luis Carrasco, University of Tennessee, Knoxville, TN, USA; National Institute for Mathematical & Biological Synthesis, Knoxville, TN, USA, Jack McNelis, Oak Ridge National Laboratory, Oak Ridge, TN, USA, Liem Tran, University of Tennessee, Knoxville, TN, USA, Yingkui Li, University of Tennessee, Knoxville, TN, USA, Jerome Grant, University of Tennessee, Knoxville, TN, USA, Wanwan Liang*, University of Tennessee, Knoxville, TN, USA; North Carolina State University, Raleigh, USA
Topics: Remote Sensing, Environmental Science, Spatial Analysis & Modeling
Keywords: detailed vegetation mapping, kudzu mapping, coarse label, two-step classification, object-based image analysis, lidar point clouds, sampling specificity
Session Type: Poster
Presentation File: No File Uploaded

Mapping vegetation species is critical to facilitate related quantitative assessment, and integrating high resolution multispectral remote sensing (RS) image and lidar (light detection and ranging) point clouds can provide robust features for vegetation mapping. Here we designed a two-step classification workflow to decrease computational cost and sampling effort, and to increase classification accuracy by integrating multispectral and lidar data to derive spectral, textural, and structural features for mapping target vegetation species. We used this workflow to classify kudzu, an aggressive invasive vine, in the entire Knox County (1,362 km2) of Tennessee, USA. Object-based image analysis was conducted in the workflow. The first-step classification used 320 kudzu samples and extensive coarsely labeled samples (based on national land cover) to generate an overprediction map of kudzu using random forest (RF). For the second step, 350 samples were randomly extracted from the overpredicted kudzu and labeled manually for the final prediction using RF and support vector machine (SVM). Computationally intensive features were only used for the second-step classification. SVM had constantly better accuracy than RF, and the Producer’s Accuracy, User’s Accuracy, and Kappa for the SVM model on kudzu was 0.94, 0.96, and 0.90, respectively. We found the sample size of kudzu used for algorithm training impacted the accuracy and number of kudzu predicted, and the proposed workflow could also improve sampling efficiency and specificity. Our workflow had much higher accuracy than the traditional method, and could be easily implemented to map kudzu in other regions or other vegetation species.

To access contact information login