A pattern directed inference approach to hardwood log breakdown decision automation

Luis Gallardo Occena, Purdue University


The hardwood log breakdown decision process in the presence of internal defect detection is modelled in a knowledge-based environment geared towards eventual sawmill automation. Information on log and internal defect profiles, now accessible through non-invasive detection methods like CAT or NMR scanning, were collected from the breakdown of yellow poplar logs. A graphics sawing simulator for hardwood logs, based on solid modelling concepts, was developed and used to analyze the log-to-lumber production process in the presence of internal defect information. Knowledge gained, from the analysis and from current sawmill practice, was formalized in a logic-based, pattern-directed inference model to enable automatic specification of log breakdown instructions from the extracted defect configuration pattern of a given log. ^




Major Professor: Jose M. A. Tanchoco, Purdue University.

Subject Area

Engineering, Industrial|Artificial Intelligence

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server