Evaluation of the Accuracy of Genetic Algorithms for Object Detection in Industrial Environment

The last few decades have witnessed the big strides genetic algorithm (GA) have taken to emerge as a potentially reliable and efficient performer in solving complex search and optimization problems in the fields of image processing, machine vision, object detection, and pattern recognition. Especially, in the last two decades, there has been a tremendous upsurge of the well researched techniques and approaches being evolved around genetic algorithms in these areas. In the study of various such approaches, it may be easily observed that a few among them have preferred to use (GA) as a vehicle for another core technique to actually ride on it, i.e., as a hybrid GA in which it is reduced merely to the role of a support foil and the real problem solver is some other technique like bacteria foraging optimization (BFO), particle swarm optimization (PSO), etc. In the present study, we attempt to explore the potential and accuracy of the GA designed to realize object detection in industrial environment not by making it to be used as a vehicle by some other technique but by allowing it to stand on its own and taking complete charge of the problem to achieve the final goal of object detection. The scope of the present work is confined to the experimental evaluation of the potential and accuracy of the GA when used as a standalone GA in realizing object detection. The analysis of the results shows that the genetic algorithm proves itself to be an adaptive, accurate, compact, and efficient candidate for object detection in industrial environment. Keywords— Object detection; image segmentation; image thresholding; genetic algorithm; selection; mutation.