Licence plate recognition system with improved recognition rate using blob detection compared with pixel based algorithm

• Main Authors: Kumar, Pavan, Thangaraj, S. John Justin
• Format: Conference Proceeding
• Language: English
• Subjects: Algorithms; Classifiers; Computer vision; Identification systems; Image analysis; Machine learning; Pixels; Recognition; Sensitivity
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0197428

The licence plate identification system use Machine Learning to improve recognition rates by using a Pixel Based Algorithm (PBA) classifier in place of a Blob Detection Algorithm (BDA) classifier. Methods and Materials: The unique licence plate identification system with the highest recognition rate employed a sample size of 2000 (Group 1=1500, Group 2=1500). The Blob Detection Algorithm (BDA) and the Pixel Based Algorithm (PBA) Classifier, both of which account for the amount of samples (N=30), are compared with new licence plate identification systems that have a greater recognition rate. The findings demonstrate that the Blob Detection Algorithm (BDA) outperforms the Pixel Based Algorithm (PBA) in terms of Recognition rate (95.92 percent) (92.13 percent). In comparison to Pixel Based Algorithm’s (PBA) 91.84 percent success rate, Blob Detection Algorithm’s (BDA) Specificity is 94.74 percent. Comparatively, the Pixel Based Algorithm (PBA) has a sensitivity of 91.49 percent and the Blob Detection Algorithm (BDA) has a sensitivity of 95.29 percent. There is a statistically significant difference in the detection rates (P=0.067). The results suggest that switching to the Blob Detection Algorithm (BDA) Classifier from the Pixel Based Algorithm (PBA) Classifier improves the licence plate recognition system’s detection rate.