For urban modeling LiDAR images play very important role. High spatial resolution data and sophisticated methods are required by many data pre-processing techniques for LiDAR images for detecting buildings. In late decades, building and tree identification from LiDAR information and aerial images with high automation and precision level has been the center of numerous scientists which was chosen as the motivation behind this survey. This study reviews latest advances in techniques used for detecting buildings and tress. Support Vector Machine (SVM) used for classification proved as best classifier for good accuracy. The comparative study reveals positive and negative approach towards strategies used for detection. © 2017 IEEE.