Part detection and classification using integrated machine vision and knowledge based expert system
Abstract
Manufacturing cost is a topic of interest in the automotive manufacturing industry. Even though rework and repair methods are acceptable processes in manufacturing, there is little investigation done on the classification and cost parameters of a repairable or re-workable part. Having a non-conforming part indicates a loss in revenue to a manufacturing company, thus the need to repair or rework the part. Repair and rework processes take non-conforming products of a manufacturing line and bring the parts back to conforming stage. In this study, the proposed system establishes a connection between detection of defects, classification of them based on defect level and the cost required to return the part back to conforming state. A series of tools such as Insight Explorer and LabVIEW aid this study in defect detection and classification respectively. Insight Explorer's output will serve as the knowledge base from which LabVIEW extracts the knowledge and performs the analysis using a series of proposed equations to classify and determine the cost of each defect. The proposed system is evaluated and validated based on real part measurements. Experimental results show that the system performs as intended and will be beneficial to the user.
Subject Area
Industrial engineering
Recommended Citation
Lopez Ulloa, Cristian Giovanni, "Part detection and classification using integrated machine vision and knowledge based expert system" (2014). ETD Collection for University of Texas, El Paso. AAI1583931.
https://scholarworks.utep.edu/dissertations/AAI1583931