Building and Tree Detection by Fusing LiDar and Aerial Images for Urban Development Planning

摘要

For GIS database and environmental monitoring, detection of buildings and trees plays very important role. By integrating LiDAR data with aerial images, significant improvement in accuracy can be achieved. Primarily, after the pre-processing aerial images are fused with LiDAR images. Due to this, region of interest (ROI) has been extracted accurately. The relevant color (mean, standard deviation) and texture features (energy, contrast, entropy etc.) are extracted from the images. These extracted features are given to support vector machine (SVM) for classification which successfully classifies buildings and trees from images with accuracy 96.96% and 59.01 % respectively. Finally, the performance of classifier is evaluated by Kappa Statistic.