Paper Key : IRJ************153
Author: Vasu Goel
Date Published: 23 Jan 2023
Abstract
A promising study subject in smart urban and the Internet of Things is number plate recognition, also known as licence plate realization or recognition utilizing image processing techniques. Many of the autonomous number plate recognition systems in use today only function in a controlled setting with photographs taken at a straight angle with acceptable lighting, clarity, and industry-standard typefaces. The fact that existing works are based on UK licence plates and may not function for Indian licence plates is another disadvantage. This study provides a novel image processing method for the detection and identification of Indian licence plates that can handle noisy, dimly lit, cross-angled, and non-standard font licence plates. In the pre-processing of this work, morphological transformation, Gaussian smoothing, and Gaussian thresholding are only a few of the image processing techniques used. Next, contours are applied by boundary following and filtered based on symbol proportions and spatial localization for number plate segmentation. After the target region filtering and de-skewing, character identification is done using the K-nearest neighbour technique. The suggested techniques showed encouraging outcomes.
DOI LINK : 10.56726/IRJMETS32994 https://www.doi.org/10.56726/IRJMETS32994