Method and system of panel area detection using sparse representation based on image segmentation

文件类型：PDF文档
文件大小：1784K
The present invention relates to a method and a system of detecting a license plate by utilizing a sparse representation scheme based on image segmentation. More specifically, the present invention relates to a method and a system for detecting a license plate by utilizing an image feature database formed based on an image piece and the sparse representation scheme. According to the present invention, the method includes the steps : extracting a candidate region of at least one license plate from a input image; and determining whether the candidate region of the license plate corresponds to a license plate after the candidate region of the license plate is represented using sparse representation through the combination of image feature vectors. Disclosed is the method of detecting the license plate characterized to be formed by using an image piece of training images of a license plate and a non- license plate. According to the present invention, the image feature vector database is created using the image feature vector, and the license plate region is detected by using the image feature vector database and the sparse representation scheme. Accordingly, even if the quality of the input image is degraded or an erroneous region is included due to the limitation in the performance of a photographing device and an inferior photographing environment, the license plate can be effectively and exactly detected.(110) License plate candidate region extracting unit(120) License plate region verification unit(122) License plate candidate region image feature vector extracting unit(124) Sparse representation based verification unit(130) Image feature vector database unit(AA) Input image(BB) Detection result of license plate regionCOPYRIGHT KIPO 2015