Automatic failure detection of serial products using novelty filter

Abstract

The present paper focus on a computer tool that seeks the failure detection of serial products. This paper begins with a brief description about the quality on manufacturing process and points out the relevance of product inspection for detecting fails aiming the product's quality. For such inspection to be accomplished, was used Digital Image Processing Techniques and Artificial Intelligence. This research were done in a mobile phone industry located in the Manaus Industrial Polo - PIM. The tool's methodology localize the possible defect areas and uses a process in which, from a data base composed by pattern images, the Novelty Filter Technique is able to discern regions of  failure through what it was instructed.