Detecting the Buildings from Airborne Laser Scanner Data by Using Fourier Transform

Abstract

The automatic extraction of objects from airborne laser scanner data and images has been a topic of research for decades. Laser scanner data have proven to be a powerful source for a wide range of 2D–3D geographic information system object tasks. This paper presents the Fourier transform as an image enhancement tool for determination of buildings from the images generated by laser signal. Spatial and frequency domain filtering techniques have been utilized for extraction of building from enhanced images. While Gaussian and Wiener filterings were selected in frequency domain, Sobel and Unsharp were selected for the spatial domain. The boundaries of buildings have been delineated from the generated images, which were obtained from inverse Fourier transform, by using edge detectors, such as Canny, Sobel, and Prewitt. Frequency domain filters using Fourier transformation were compared with spatial domain filters in the way of kernel function and windows. The reason for doing the filtering in the frequency domain is that it is computationally faster to perform two-dimensional Fourier transforms and a filter is more applicable than to perform a convolution in the spatial domain. Results showed that using Fourier transformation has a great advantage in enhancing images and detecting the buildings on images. Filtering in the frequency domain is more efficient computationally than spatial domain filtering when the filter size is big. The conclusion proved that Fourier transformation can be used as an image enhancement tool to detect and extract buildings automatically.