提出了一种汽车牌照的实时分割方法，利用牌照区域灰度变化剧烈，而背景相对平缓的特点 ，以适当阈值二值化牌照图像以消除背景干扰，结合牌照知识，搜索牌照区域，如找不到， 则自适应调整阈值，重复以上步骤。实验证明该算法定位准确并且是实时的。%Electronic collection of toll requires real-time recognition of license plate of a continuously moving vehicle. Accurate and rapid segmentation of license plate is the key problem of recognition system of license plate. We propose a fastad aptive segmentation algorithm for thresholding and searching. This segmentation algorithm utilizes the fact that the grayscale in the area of license plate changes more rapidly than the background. We select a proper threshold to binarize the gray image to eliminate the disturbance of background, and search the license plate according to the prior knowledge database of license plates. Experimental results, as given in Table 1, have preliminarily validated the effectiveness and high-speed of proposed approach. The processing time of an image is 0.1 s to 0.2 s.
展开▼
