A Machine Vision Based Automatic Optical Inspection System for Detecting Defects of Rubber Keypads of Scanning Machine

Abstract

In order to detect defective rubber keypads, factories have to devote massive manpower and financial resources. In this paper, a vision-based system for the detection of a number of rubber keypad defects is designed and developed. To improve the lens distortion rectification, a novel, easy calibration method using image local transformations defined by both sets of points that are detected in the distorted image and the undistorted ones without using any model for a wideangle and low- cost lens camera was proposed. The system detects the defects of 14 different types of rubber keypads quickly (~within about 1.8 seconds) and accurately even with a normal laptop and a low-price webcam.