SUPPRESSION OF REAL TIME SHADOW FOR EFFICIENT LICENSE PLATE RECOGNITION

The presence of shadow in the License Plate (LP) results in disambiguation in edges, hides the actual registration number. Recognizing the LP information by suppressing sunlight shadow is a topic of research. Suppression of shadow artifact is a challenging task. Till now the VLPR schemes result in inability to recognize LP information due to occurrence of shadow. Till now this problem remains as a major problem to be solved by computer vision researchers. This paper provides a solution for suppressing the shadow that varies from morning till evening due to sunlight and improves the LPR in terms of accuracy. The proposed ‘Suppression of Shadow in Partially Occluded LP (SSPOLP) method performs the suppression of shadow using cepstrum approach based on Grassman Axiom integral method. It suppresses the shadow noise using cepstrum approach based on gamma correction method i.e. Grassman axiom integral method for shadow removal followed by LP detection using region props method. The overall recognition success rate for SSPOLP scheme is found to be 93.7% for online MNIST and MIT datasets.