Structured illumination autonomous machine vision system

DETAILED DESCRIPTION I claim: 1.
A method of providing information for automatically determining pathway boundaries for autonomous robotic navigation, comprising the steps of: projecting a structured light pattern of sequentially scanned thin horizontal lines on a pathway and background, viewing the resultant light pattern which is reflected from the pathway and background with a stereoscopically located video camera, storing an image field which includes bits which correspond to said resultant light pattern in a memory, and processing the image field to determine the degree of distortion of said resultant light pattern, wherein the texture of the background is courser than the texture of the pathway, and whereby said information for determining said pathway boundaries is provided as indicated by the scanned lines reflected from the background having a greater degree of discontinuity than those reflected from the pathway, wherein said processing includes taking the Hough transform of said image field to identify lines in said resultant light pattern, and wherein said processing further includes rank filtering said identified lines.
2.
The method of claim 1 wherein said processing further includes morphologically dilating said filtered lines.
3.
The method of claim 2 wherein said processing further includes detecting the end points of said lines following said morphological dilation.
4.
The method of claim 3 wherein said processing further includes deriving the 3-D coordinates of the end points of the lines.




Description:
The present invention is directed to a method and apparatus for automatically determining pathway boundaries for autonomic robotic navigation.
In modern day warfare, mobile robots or autonomous vehicles are deployed in a foreign terrain environment.
These must be provided with a vision system which rapidly and accurately determines the location and direction of the pathway ahead, so as to keep the vehicle moving properly on the pathway.
The present invention solves the formidable problem of distinguishing between pathway and background so as to permit such automatic vehicular navigation