Lerma, Jose Luis 
  
Finally, the classified image using spectral and textural bands offers the best results, Fig. 3c. By defect, it fits very well. 
The confusion areas on the other classified images (Fig. 3a, 3b) are now classified as null class (black color). Although 
there is an increment of null class there is also and a higher probability of classifying properly the six classes. 
Particularly interesting is the right difference among polished limestone and washed & washed limestones. 
5 CONCLUSIONS 
The results achieved in this study seem promising as regards the combination of spectral and textural classifications for 
the characterization of architectural monuments, not only for the identification of different materials but also for the 
recognition of similar materials. 
The union of the supervised spectral and textural classification has been found to work effectively when applied to the 
identification and recognition of masonry wall materials. It has been shown that the near-infrared band increases 
discrimination of rather similar materials (mortar and limestone) and that the addition of textural bands allows a better 
(much real) recognition of different kind of limestones. Anyway, although textural images require more time of 
processing itis worth working with them altogether as additional input bands for classification. 
REFERENCES 
J.M. Grunicke, G. Sacher and B. Strackenbrock, 1990. Image processing for mapping damages to buildings. CIPA 
International Symposium, Cracow, pp. 257-263. 
R. M., Haralick, L. G. Shapiro, 1992. Computer and Robot Vision. Reading: Addison-Wesley. Vol. I, pp. 457-462. 
J. Herráez, P. Navarro, and J.L.Lerma, 1997. Integration of normal colour and color infrared emulsions for the 
identification of pathologies in architectural heritage using a digital photogrammetric system. International Archives of 
Photogrammetry and Remote Sensing, Vol. XXXII, Part SC1B, pp. 240-245. 
Kodak, 1990. Photographic Filters Handbook. Kodak Publication No. B-3. Eastman Kodak Company. 
J. L. Lerma, J. Herráez, 1999. Reconocimiento y Cartografiado Automático de Monumentos Arquitectónicos. Actas del 
XI Congreso Internacional Ingeniería Gráfica, Volumen IL, pp. 732-740. Logrofio-Pamplona, Spain. 
J. L. Lerma, L. A. Ruiz, 1999. Analysis of Historic Building Facades by Spectral and Texture Image Methods. Pattern 
Recognition and Image Analysis (Eds.: M. I. Torres & A. Sanfeliu). Proceedings of the VIII National Symposium on 
Pattern Recognition and Image Analysis, Volume I, pp. 23 5-242. 
J. L. Lerma, L. A. Ruiz, 2000. Análisis de Fachadas de Edificios Históricos mediante Métodos Espectrales y de 
Texturas. Revista Electrónica de Visión por Computador, Vol. 2, pp. 9. 
http://www.cvc.uab.es/reve (April 2000) 
S. Ryherd, C. Woodcock, 1996. Combining spectral and texture data in the segmentation of remotely sensed images. 
Photogrammetric engineering & Remote Sensing. Vol. 62, No. 2, pp 181-194. 
P. Waldhäusl, 1997. Mass — Documentation of architecture and its environment objectives and tasks for CIPA and its 
working groups. International Archives of Photogrammetry and Remote Sensing. Vol. 32, Part 5C1B, pp. 41-46. 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B5. Amsterdam 2000. 483