A New Approach to 3D Dense LiDAR Data Classification in Urban Environment
- PDF / 407,964 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 104 Downloads / 187 Views
SHORT NOTE
A New Approach to 3D Dense LiDAR Data Classification in Urban Environment Inshu Chauhan & Claus Brenner & R. D. Garg & M. Parida
Received: 12 June 2013 / Accepted: 9 December 2013 # Indian Society of Remote Sensing 2014
Abstract Classification of Mobile Mapping LiDAR (Light Detection and Ranging) data is a challenge in the research community since the day when laser scanner system were integrated and mounted on vehicles for collection of 3D data in urban environment. The approach proposed here for classifying LiDAR data is analogous to the process followed for classifying data from satellite images. Pixel based and segmentation based methods have been employed in past for classifying images obtained from satellites. These methods were based on spectral properties of objects present in the images. But for Mobile mapping LiDAR data this approach has been applied and tested for the first time. The properties of this data are completely different from that of satellite images. So even if the basic approach remains the same, many changes have to be made in the entire classification process. The paper here aims to propose the basic procedure of using pixel-wise classification on dense 3D LiDAR data. Keywords LiDAR . Classification . Principal component analysis . Urban environment . High volume 3D data
I. Chauhan (*) : R. D. Garg : M. Parida Department of Civil Engineering, Indian Institute of Technology, H.no. 1266, Jadugar road, 44 Civil lines, Roorkee 247667, India e-mail: [email protected] R. D. Garg e-mail: [email protected] M. Parida e-mail: [email protected] C. Brenner Institute of Cartography and Geoinformatics, Leibniz University Hannover, Hannover, Germany e-mail: [email protected]
Introduction For urban development planning, forecast and simulation and also for automatic positioning of vehicle in an urban area, it has been desirable to extract terrain and building information in such an urban area (Shan and Sampath 2007). However, deriving accurate, quantitative measures from remote sensing imagery over urban areas remains a fundamental research challenge due to the great spectral and spatial variability of the materials that compose urban land cover (Xian and Crane 2005). With the advent of new technologies like laser scanning, methods can be developed to overcome the problems faced in classification of urban areas. A LiDAR (Light Detection and Ranging) works basically through laser source which emits laser. Then the light reflected by the objects is gathered by the sensor which is then used to calculate distances from objects. From the LiDAR survey conducted in an area, two types of datasets are generated. First data set is scanned representation of the intensity data of the area and second dataset is information about the x, y and z coordinates of the scanned points. The information about x, y and z of the points can be stored as text files; however the size of these files is huge. Classification of such an enormous amount of points remains a challenge throughout the scienti
Data Loading...
