Dr. Suchendra M. Bhandarkar, Department of Computer Science.
Dr. Timothy D. Faust, Warnell School of Forest Resources.
 
 
    The CVLPPS project is a sequel to the CATALOG project. From the 3-D reconstruction of the log and knowledge of the internal defects identified and localized by CATALOG, the CVLPPS (Computer Vision based Lumber Production Planning System) can formulate sawing strategies to optimize the yield and grade of the resulting lumber. The CVLPPS can incorporate constraints imposed by the manufacturing process and market demand for various lumber sizes and grades. The lumber grading rules of the National Hardwood Lumber Association  (NHLA) are incorporated into a rule base that is accessed by the optimization algorithm. The problem of optimizing the lumber yield  and lumber grade is involves constrained multivariate optimization  and computational geometry and is modeled as as the Shortest Paths Problem in a directed graph. The current version of CVLPPS  incorporates the live sawing process where the sawing planes are constrained to be parallel to each other. Future versions of  CVLPPS will incorporate other sawing processes as well as rules that capture aesthetic (and hence subjective) aspects of lumber value assessment based on wood grain texture, wood surface texture etc. The use of parallel computing to ensure response times acceptable in a real time manufacturing scenario are being explored.
 
 
Publications
 
S.M. Bhandarkar, T.D. Faust and M. Tang, A Computer Vision System for Lumber Production Planning,
Proc. IEEE Intl. Wkshp. Appl. Computer Vision, Princeton, NJ, October 19-21, 1998, to appear.