Identifying Material Property Boundaries From LIDAR Data Using Wavelet Analysis
Moreira, Alan B 1 ; Neupauer, Roseanna M 2 ; Weismann, Gary 3 ; Frechette, Jed 4 ; Wawrzyniec, Tim 5
1 University of Colorado
2 University of Colorado
3 University of New Mexico
4 University of New Mexico
5 University of New Mexico
Wavelet analysis is a signal and image processing technique that can be used to characterize variations within the signal or image. In this work, we employ wavelet analysis on LIDAR reflectivity data of a rock face to identify boundaries between different materials types on the rock face. The LIDAR data are recorded in a topologically regular but geometrically irregular array. We extract a two-dimensional intensity image map and perform the two-dimensional continuous wavelet transform to obtain wavelet coefficients at each node in the array for a single wavelet scale. Within a zone that contains a single material type, the wavelet coefficients are relatively uniform; while between zones, the wavelet coefficients change abruptly. We identify the boundaries between zones as the locations where the the spatial rate of change of the wavelet coefficients attains its local maxima. We investigate the range of wavelet scale parameters that lead to optimal boundary identification. The boundary locations identified through the wavelet analysis technique correlate well with the observable sediment structure.
Addison, P. S. (2002). Illustrated Wavelet Transform Handbook, The. Bristol: Institute of Physics Publishing.
Neupauer ,R.M. (2003). Summary of Wavelet Techniques.
Torrence, C. A. (1998). A Practical Guide to Wavelet Analysis. Bulletin of the American Meteorological Society , Vol. 79, No. 1, January 1998 p. 61-78.