License plate super-resolution reconstruction based on improved ESPCN network
Jingyi Du, Qingli Liu, Kang Chen · 2019
This Due to the limitation of the imaging equipment or other conditions, the obtained license plate pixel resolution is low, which brings great difficulty to the license plate detection technology. In order to solve the above problems, the improved ESPCN network is used to extract the original high-definition picture detail features, and the super-resolution reconstruction of the license plate image with lower resolution is performed. The real-time convolution residual network method is compared with the classical reconstruction methods such as bilinear interpolation, nearest neighbor interpolation and ESPCN. The subjective visual evaluation and objective quantization index (PSNR, SSIM) are superior to the classical reconstruction method. It is proved that the real-time residual convolution network improves the resolution of the license plate image and also improves the license plate recognition rate to some extent.