Genetic algorithm based printed circuit board (pcb) inspection system

Abstract

This paper presents the first prototype of automating a low-cost printed circuit hoard (PCB) inspection on physical defects through the development of a technique for image detection using a Genetic Algorithm (CA). A GA is a powerful search technique which is applicable to various type of problems. The algorithm works on the principle of simulation of the evolution of individual structures via prwesses such as selection, mutation and crossover. This research aims to detect rotation angle and displacement of a board placed arbitrarily on a conveyor belt passing under the camera. Variant rotation and displacement values are represented as chmmosoma to create a population which will evolve using GA operators and a specially tailored fitness function. Ideal parameter tuning of CA plays an important mle in boosting the capability of this technique in order to perform effectively in the PCB inspection environment. Previous results regarding the use of CA on ARM based embedded PCB inspection system which focused on rotation and displacement of Integrated Circuits provides the motivation for this research. This research is run on Linux-based PC and real industrial samples PCB images are used in the implementation.