Improved license plate localisation algorithm based on morphological operations

Abstract

In presenting this thesis in partial fulfilment of the requirements for a Postgraduate degree from the University of Saskatchewan, I agree that the Libraries of this University may make it freely available for inspection. I further agree that permission for copying of this thesis in any manner, in whole or in part, for scholarly purposes may be granted by the professor or professors who supervised my thesis work or, in their absence, by the Head of the Department or the Dean of the College in which my thesis work was done. It is understood that any copying or publication or use of this thesis or parts thereof for financial gain shall not be allowed without my written permission. It is also understood that due recognition shall be given to me and to the University of Saskatchewan in any scholarly use which may be made of any material in my thesis.

Deep learning‐based embedded license plate localisation system

“…There exist many proposed methods for localisation, and include the following: the sliding window technique and modifications thereof [1, 15, 16]; variations of edge detection analysis [15, 17]; histograms of oriented gradients [16]; mathematical morphology [4, 18, 19]; and other computer vision techniques [17, 20] Some plate regions are localised by detecting constituting characters [21].…”

“…A recent LPL solution that uses mathematical morphology is proposed in [4]. This system isolates a license plate in a grayscale image using the following operations in order: a Top‐Hat morphological operation for background removal; Otsu thresholding for binarisation; noise removal using opening and closing morphological operations; plate candidates located via finding contours; and likeliest plate candidate selection using geometric processing of license plate characteristics.…”

“…This assertion is not incorrect: while localisation is reasonably easy and fast for a human, isolating the parallelogram that is the license plate, if one even exists, in a myriad of a meaningless background containing everything from street signs to people is difficult to achieve quickly via algorithmic means. LPL is the ALPR step that requires the greatest processing time [4].…”

A Novel Local Human Visual Perceptual Texture Description with Key Feature Selection for Texture Classification

“…Step . Connect the edges by the morphological transformation that sets 0-valued pixels to 1 if they have two nonzero neighbours that are not connected [50], so that we get the connected edge map :…”

License plate segmentation and recognition system using deep learning and OpenVINO

“…Then, vehicles and/or license plates are detected. Finally, character regions from the license plate are extracted (segmented) and interpreted using an optical character recognition (OCR) method [2].…”