LAND COVER CLASSIFICATION USING LiDAR DATA

ABSTRACT: Land cover classification has increasingly gained importance in addressing environmental processes and problems especially for developing agricultural countries. The information it provides is of great help in policy formulation and decision making. This study attempts to apply land cover classification using Light Detection and Ranging (LiDAR) data in a municipality in the province of Isabela, Cagayan Valley Region, Philippines. Using Lastools, Envi,ArcGIS, and eCognition, multiple images were derived from LiDAR data and classified using Object-Based Image Analysis (OBIA). Eight classes were identified for the study area: Building, Bare land, Fallow, Road, Non-agricultural trees, Rice, and Water. Accuracy assessment was done to compare the classified map with ground reference objects. An overall accuracy of 0.92 and Kappa Index of Agreement (KIA) of 0.87 was generated for the output which can be considered accurate. With such a level of accuracy, the municipality covered by the case study will benefit from the land use classification output. Municipalities in the Cagayan Valley Region with LiDAR data captured through the UP DREAM program are potential beneficiaries of this reliable land cover classification method. 1