ization) overlaid 
ing (b) 
ION 
In (a) the training 
(green) pixels are 
from a nearby but 
on the algorithm 
s used to train the 
ch class (slide and 
[he BNN has 36 
neurons, and one 
is a hyperbolic 
n is Levenberg- 
number of epochs 
' testing area, the 
whether that that 
t of applying the 
sure 6(a) showing 
xels. The healthy 
e 6 (b) shows the 
1e shows the slide 
; we compare the 
area is correctly 
'ea is classified as 
n matrix for the 
re classified with 
r this testing area. 
s slides has a high 
jixels. Also, since 
ny changes in the 
vegetation can contribute to miscalculation of roughness and 
detection as slides. 
  
(b) 
Figure 6. (a) Slide classification using the algorithm and 
TerraSarX data; (b) The ground truth data for testing area 
Table 1. Confusion matrix 
  
  
  
  
  
rid Slide Hes Accuracy 
Slide 486 463 0.5 
Healthy Area 563 1616 0.7 
Accuracy 0.5 0.8 0.67 
  
  
  
  
  
     
   
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
   
   
  
    
    
   
     
     
    
    
   
   
   
    
     
   
    
    
   
   
   
  
   
5. CONCLUSIONS 
Slough slides are slope failures along a levee. The roughness 
and corresponding textural characteristics of the soil in a slide 
can change the amount and pattern of radar backscatter. 
Exploiting these changes in texture and backscatter pattern, an 
algorithm based on a neural network and TerraSAR-X data is 
developed to detect a landslide on the levee system. The 
methodology includes: 1) segmentation of the levee area from 
background and; 2) extracting features 3) training a back 
propagation neural network classifier; and 4) testing and 
validation of the results using ground truth data. A dual 
polarimetric X-band image is acquired from the German 
TerraSAR-X satellite on Sep 04, 2011. Ground-truth data 
include the slides and healthy area. The results show that the 
algorithm is able to detect the slide area with around 67% 
accuracy. Since the TerraSAR-X is sensitive to changes in 
texture and roughness and also vegetation, some pixels which 
are not reported as slide zones are incorrectly classified as slide 
pixels. 
REFERENCES 
Aanstoos, J. V., Hasan, K., O'Hara C., Prasad, S., Dabbiru, L., 
Mahrooghy, M., Nobrega, R. A. A., Lee, M. L., & Shrestha, B., 
2010. Use of Remote Sensing to Screen Earthen Levees. Proc. 
2010 39th IEEE Applied Imagery Pattern Recognition 
Workshop. Washington, DC: IEEE. 
Aanstoos, J. V., Hasan, K., Mahrooghy, M., Dabbiru, L., 
Nobrega, R. A. A., & Prasad, S., 2011. Screening of Earthen 
Levees Using TerraSAR-X Radar Imagery. 4th TerraSAR-X 
Science Team Meeting. Oberpfaffenhofen, Germany: DLR 
(German Aerospace Center). 
Bulmer, M. H., Petley, D. N., Murphy W., and Mantovani 
F., 2006. Detecting slope deformation using two-pass 
differential interferometry: Implications for landslide studies on 
Earth and other planetary bodies.J. Geophys. Res., 111, pp. 
E06S16. 
Burrus, C. S., R. A. Gopinath, H. Guo, 1997. Introduction to 
Wavelets and Wavelet Transforms: A Primer. Prentice Hall. 
Fausett, L.V., 1993. Fundamentals of Neural Networks: 
Architectures, Algorithms And Applications, Prentice Hall. 
Hossain, A. K. M. A., Easson, G., Hasan, K., 2006. Detection 
of Levee Slides Using Commercially Available Remotely 
Sensed Data. Environmental and Engineering Geoscience, 12 
(3), pp. 235 — 246. 
Mahrooghy, M., Aanstoos, J.V., Hasan, K., Younan, N. H., 
2011. Levee Soil Moisture Assessment Based On 
Backpropagation Neural Network Using Synthetic Aperture 
Radar. Proceedings of the 34th International Symposium for 
Remote Sensing of Environment. Sydney, Australia. 
McKean, J., Joering, J., 2004. Objective landslide detection and 
surface morphology mapping using high-resolution airborne 
laser altimetry. Geomorphology, 57, pp. 331-351. 
Oh, Y., 2004. Quantitative retrieval of soil moisture content and 
surface roughness from multipolarized radar observations of