We try to regonize the letter and number of a lisence plate.
Our input image
This is the sample lisence plate images that we found on internet. Our optimal result is the correct lisence plate numbers and letters of the car. For example for the first one should be 284FH8
We finished on the research part of our project. We plan to start coding part on thanksgiving. We plan to start by training our algorithm by using templet containing every numbers and letters. The second part is matching, we will process our raw image. Using Edge detection to find the AOI, and match each regions to the correct character .
Proposal for Automatic License and Number Plate Recognition System for Vehicle Identification by Hamed Saghaei https://arxiv.org/ftp/arxiv/papers/1610/1610.03341.pdf
Summary of the paper and another people's work.
Previous work of License Plate recognition include k-nearest neighbors algorithm Edge detection, Character segmentation, convert image to binary to process image.
There are two part. First part is training, there is a template with all the letters and numbers. Second part is recognition, we needs process the input image using methods like edge detection to find the region of interest(the lisence plate) then use kkn to find the correct letter or number;
challengesThere are few challenges that under our consideration. First is the quality of the raw image. We might have a picture with awful lighting or angle, we might have background letters(for example:someone has some letter in his T-shirt in the background) If single lisence plate detection for image goes well, we might consider mult-plate recongition or plate recognition for the videoes
Credits and Bibliography
Proposal for Automatic License and Number Plate Recognition System for Vehicle Identification by Hamed Saghaei https://arxiv.org/ftp/arxiv/papers/1610/1610.03341.pdfTeammate: Xi You; Jiafei Xue