Abstract

Polarization-encoded optical shadow-casting (POSC) is found to have remarkable promise for fully utilizing the parallelism of optics in performing arithmetic and logical operations with spatial coded patterns as inputs.¹ However, so far little work has been done to exploit its role in realizing 2D image processing techniques in the spatial domain. For image processing applications, it is often necessary to extract the outline features of an image. Recently, POSC schemes have been used for edge detection by using a modified difference operator, i.e., an L-shaped operator.² However, a difference operator is not necessarily suitable for detecting edges with equal edge strength since it may produce weak response for such cases. On the other hand, because of the symmetric nature of the Roberts mask, all the edges of an image will have equal strength. In this paper, accordingly, we implement the Roberts operator based edge detection by using the POSC technique. The performance of the proposed technique is verified by computer simulation.

© 1992 Optical Society of America