Blind Deblurring of Single Car Image Based on Differential Autocorrelation and L-like Curve Method

Ke ZENG, Mu-rong JIANG, Yu LUO, Yu LU, Guo-cai DU

Abstract


Blind deblurring of single image is an inverse problem in the field of image processing. Its difficulty lies in the non-uniqueness of the inverse process. The regularization method is a classical algorithm for blind deblurring of images, it can iteratively solve blur kernel and clear image by constructing regularization terms. However, improper values of regularization parameter and blur kernel size can easily lead to poor inverse results. In this paper, by calculating differential autocorrelation of the image to obtain the blur scale as the blur kernel size. Then, we proposed a L-like curve method to calculate the regularization parameter. Finally, these calculation results were used to deblur the image with the regularization model. The experimental results show that the proposed method can effectively achieve blind deblurring of images, and the inverse effect of a single blurred car image is obvious. It can meet the requirements of accurately identifying the license plate information of blurred car image.

Keywords


Blind deblurring, Differential autocorrelation, L-like curve method, Regularization


DOI
10.12783/dteees/icepe2019/28971

Full Text:

PDF

Refbacks

  • There are currently no refbacks.