331 
In comparing the results with the original images, we note 
that even for the human eye it is not very easy to distinguish 
differences between the six textures of Figures 1 and 2, while 
the proposed method can effectively extract the boundaries 
between these textures. We also note that in the center of 
Figure 1, there exists an horizontal line due to the mosaic of 
textures, however this line does not appear in Figure 3 as a 
texture boundary. Due to the geometric effect of the moving 
window, the four exteriour portions of Figures 1 and 2 were 
not processed, and the derived boundaries are sligtly thicker. 
The boundaries can be easily improved by traditional thinning 
techniques. 
4 CONCLUSION 
Edge detection is widely used in image processing and 
segmentation. Traditional edge detectors are suitable for 
separating grey level homogenous regions, but inapplicable 
for extracting texture boundaries. In this paper, the texture 
spectrum is combined with the Roberts operator and applied 
to texture edge detection. The results obtained from two 
examples show a promising performance potential for the 
proposed method. 
5 REFERENCES 
Brodatz, P., 1968. Texture - A photographic album for 
artists and designers. Reinhold, New York. 
Haddon, J.F., 1988. Generalized threshold selection for edge 
detection. Pattem Rocognition, 21(3): 195-203. 
Haralick, R.M., K., Shanmugan, and I., Dinstein, 1973. 
Textural features for image classification. IEEE Trans. 
Syst., Man., Cybern., SMC-3(6): 610-621. 
Haralick, R.M., 1979. Statistical and structural approaches 
to texture. In Proc. of IEEE, Vol. 67, pp.786-804. 
He, D.C., L., Wang and J., Guibert, 1987. Texture features 
extraction. Pattern Recognition Letters. 6: 269-273. 
He, D.C., L., Wang and J., Guibert, 1988. Texture 
discrimination based on an optimal utilization of texture 
features. Pattern Recognition. 2: 141-146. 
He, D.C., and L., Wang, 1989. Texture unit, texture 
spectrum and texture analysis, In Proc. of IGARSS'89 
(International Geoscience and Remote Sensing 
Symposium), Vancouver, Canada, IEEE#89CH2768-0, 
vol.5, pp.2769-2772. (Accepted for publication in IEEE 
Transactions on Geoscience and Remote Sensing, 
1990). 
Marr, D., and E., Hildreth, 1980. Theory of edge detection. 
In Proc. Royal Society. London, B207, pp. 187-217. 
Wang, L., and D.C., He, 1990. A new statistical approach 
for texture analysis. Photogrammetric Engineering and 
Remote Sensing. 56(1): 61-66. 
Wang, L., and D.C., He, 1989. Texture classification using 
texture spectrum. Pattern Recognition. (Accepted for 
publication). 
Wang, L., D.C., He and A., Fabbri, 1989, Textural filtering 
for SAR image processing. In Proc. of IGARSS'89, 
(1989 International Geoscience and Remote Sensing 
Symposium), Vancouver, Canada, (IEEE#89CH2768- 
0), vol.5, pp.2785-2788. (Accepted for publication in 
IEEE Transactions on Geoscience and Remote Sensing, 
1990). 
Young, T.Y., and K.S., Fu, (eds.), 1986. Handbook of 
pattern recognition and image processing. Academic 
Press Inc., Orlando.