The License Plate Identification of Fast Moving Vehicles

License plate number is the unique character of the vehicle identification. In the case of hit and run, the license plate recognition is the key measure. The snapshot of the license plate captured by the surveillance cameras are suffering from the blurring effect and it cannot be easily recognized by the human eye. Such license plates have low resolution and have lost the edge information hence it is difficult to recognize and it becomes the challenging task for the researcher to deblurring the images. In this work, the novel kernel parameter estimation algorithm has been proposed for deblurring the license plate of the fast moving vehicle. Firstly, the image is estimated by sparse representation. On the sparse representation, the angle is calculated based on the genuine motion angle and length of the motion kernel is calculated by Radon transform in the frequency domain. Finally, the image is deblurred by NBID algorithm. The proposed approach achieves the promising results and this algorithm can handle the large motion blurred license plate which is unidentified by the naked human eyes.