Enhancing Image Recognition of Damaged Number Plates in the Running Vehicle using Genetic Algorithm Compared with Bernsen Algorithm

Abstract

Aim Innovative Automatic detection of vehicle number plates using machine learning algorithms and improving the accuracy of image recognition. Materials and methods : Two sample groups using 237 images from the sample dataset, which is tested at 80% for G power with t-test analysis.To improve the accuracy of recognition, the genetic algorithm is proposed and compared with the Bernsen algorithm. Results: Test results prove that in an uneven illuminated environment the genetic algorithm has an accuracy of 91.5 % , which seems to be better than the Bernsen algorithm accuracy of 88.9%. Since the significance is around 0.017, there is a statistically significant difference among the study group with (p<0.05). Conclusion: For distorted and damaged images, the detection and image recognition of number plates using the genetic method seems to appear better than the bernsen algorithm. Detection of violations using road side cameras can perform better with the proposed work.