Detection and 3D Measurement of a Rectangular Solid in a Single Image
Tomohiro Nakamichi, Gang Xu · 2009
This paper describes a new approach to automatic detection and 3D measurement of a rectangular solid in a single image shot under two imaging conditions. We find its six outer vertices in the image by searching the largest convex hexagon with a sufficiently small back projection error, and determine their 3D coordinates by using the rectangularity constraint. Experimental results show the accuracy obtained by this approach using the image that has noise and broken edges.