Blob extraction based character segmentation method for automatic license plate recognition system

Abstract

A character segmentation algorithm for automatic license plate recognition is presented in this paper. Character regions are selected through binarization, connected component analysis, and character recognition. A blob analysis operation excludes noisy blobs, merges fragmented blobs, and splits clumped blobs. A character segmentation module achieved an accuracy rate of 97.2%. The recognition accuracy of the complete system with license plate localization was 90.9%. In depth analysis of failure cases is also provided for better understanding of the algorithm and a future development direction.

Hardware Architecture for Real-Time Computation of Image Component Feature Descriptors on a FPGA

“…It allows for different combinations of hardware modules for the parallel computation of an object perimeter, ellipse features, feature descriptors which are computed as histograms, and so forth, to be included into the connected component analysis. These feature descriptors are all relevant in relation to the connected component analysis used for applications such as those presented in [18,20,21,36,37]. We believe that modularity and a clear separation between the descriptor computations and the labeling make the presented architecture suitable for design automation.…”

“…LBP was used by Mohammed and Yampolskiy in relation to face recognition [17]. Bounding box, COG, and area of blobs were used by Yoon et al in relation to the recognition of a car's license plate and its characters [18]. A Gabor filter bank was used to segment fabric defects on textile as blobs [19] and this is a good example of a real-time application that could benefit from additional analysis of geometrical shape, position, and size of defect regions.…”

A design method for Malaysian car plates character segmentation based on skew correction

“…Character segmentation is not a problem anymore since a lot of research have been done and shows promising results. However, it still poses a challenge for accurate character segmentation under the situation when the characters are connected together and plate boundaries connected to inside characters [1], [2], [7], [10]. Most existing methods uses prior knowledge for example the number of characters and the significant spacing interval between each character to segment the characters effectively but fails in this case.…”

“…Besides, it is used to locate the spaces of each character which satisfy the conditions of a valid character [1]- [8]. Another method is based on connected component analysis (CCA) by measuring the properties of the blobs and then eliminating blobs that are not related to character blobs [9], [10]. Character segmentation is not a problem anymore since a lot of research have been done and shows promising results.…”

Determining adaptive thresholds for image segmentation for a license plate recognition system

“…In addition, one of the most recently introduced methods for threshold determination known as adaptive contrast is also based on Otsu method [13]. Meanwhile, Niblacks and Sauvolas methods become an alternative to solve shadows and illumination problem [14], [15] on license plate images. Global binarization methods are dependent on distinct image foreground and background such that the preprocessing and post-processing on binarization can contribute towards optimal results.…”