Image Blur Detection and Restoration of Licen...

Image Blur Detection and Restoration of License Plate

Author Name : Prof. Jaychand Upadhyay, Savio Pereira, Vaibhav Soni, Ashif Khan

ABSTRACT

Fast motion of vehicles may cause difficulty in identifying the amount plate by surveillance camera. Because of rapid motion of car, the image perceived by the camera develops a blur. But because of the speed of the vehicle, at exposure time it causes blur of the snapshots captured by the surveillance camera. This results into unrecognizable, undetectable and deterioration of image resulting in loss of some image information. In such cases we will use image restoration to recover any useful clue from the snapshots for identification of car’s license plate. In such cases we will use a pre-trained model Wpod Net and Computer Vision techniques with OpenCV is proposed to detect and extract the car license plate. To see the blurriness of the car license plate image we used Laplacian kernel and to restore the blur image we used a pre-trained SRGAN, ESRGAN, RRDB model. The proposed system has tried to bring the thought of designing a strong method for car license plate blur image detection and restoration. The aim of this work is that the enhancement of blur image with image processing which reduces the strain on human eyes in recognizing the blurry number plate. It is often helpful in identifying problematic vehicles like in road accidents or catching any rule violating vehicle. Our project finds solution for image blur detection and restoration of car license plate.

Keywords: -Image restoration, Wpod Net, Computer Vision, OpenCV, Laplacian kernel, SRGAN, ESRGAN, RRDB.