Abstract

A formidable task in many image processing applications is the segmentation of the scene into regions of interest. A major distinguishing feature between these regions is often the 3-D texture of these regions. Typical methods of textural image segmentation are very computationally intensive, often lack the required robustness, and are incapable of sensing the 3-D texture of various regions of the scene. Scanned laser lines of structured light viewed by a stereoscopically located single video camera result in an image in which the 3-D characteristics of the scene are represented by the discontinuity of the projected lines. The discontinuities of these scan lines can provide the required information for quick identification of the texture in various regions. The image of the laser scan lines is actually a 2-D representation of the 3-D texture of various regions of the scene. This image is conducive to processing with simple regional operators to classify rapidly and robustly regions according to their texture.

© 1987 Optical Society of America