Abstract:

This paper introduces an improved algorithm for detecting specularity using photometric stereo technique. While three images are enough for recovering surface heights, additional image may be used for detecting specularity. Therefore a four-source photometric stereo system was built for the purpose of this study. If specularity exists in one of the images, the intensity value from this image will elevate the resulting surface normal and surface reflectance factor. Traditionally, specular contribution can be detected by examining the relative deviation of the surface reflectance factors (one for each three-source combination) against a threshold value. Instead of calculating four surface reflectance factors and their deviation, a direct approach for detecting specularity by comparing image intensities is presented. The approach requires capturing four images at azimuth angles of 90° then examining the difference between image intensities against a threshold value. Experimental results used to verify the approach show that a value of 20% of the average image intensity of an object can be used as a threshold value. The results show also that the introduced algorithm is efficient and powerful for recovering specular surfaces by using a four-source photometric stereo technique

 Keywords: Photometric Stereo, Specularity, Reflectance, Light source, Depth Recovery

\

Biography:

Amin M. El Gendy, received his B.Sc. (Hons) degree in Civil Engineering from the Suez Canal University, Egypt, in 1996, and his M.Sc. in Civil Engineering from the Suez Canal University, Egypt in 2001. Currently, he is a Ph.D. candidate with the Civil Engineering Department, University of Manitoba, Winnipeg, Manitoba, Canada. Mr. El Gendy research interests include image processing and its application in transportation engineering.

Ahmed Shalaby, received his B.Sc .and M.Sc. in Civil Engineering from Cairo University, Egypt, in 1987 and 1992. He completed a doctoral degree in pavement engineering at Carleton University, Canada. He is currently an Associate Professor of Civil Engineering and an Associate Dean of the faculty of Engineering at the University of Manitoba. His research interests are in the areas of pavement engineering, infrastructure management, highway materials and road surface characterization.

BibTex:

@ARTICLE{P1150648002,

AUTHOR = {A. El Gendy and A. Shalaby},

TITLE = {Improved Specularity Detection relying on a Photometric Stereo Technique},

JOURNAL ={ICGST International Journal on Graphics, Vision and Image Processing},

YEAR = {2007},

VOLUME = {07},

ISSUE ={1},

PAGES={39--46}

}

(Full Paper779KB)