摘要
研究车牌识别系统中准确定位车牌的问题。在复杂背景和不同光照条件下,采用传统的Prewitt检测车牌图像边缘时,造成检测到的车牌图像边缘不连续和丢失边缘角度信息,进而导致车牌定位不准确等问题。为解决上述问题,提出了一种改进的Prewitt边缘检测算子和自适应阀值相结合的车牌定位方法。首先用改进的Prewitt算子对预处理后的车牌图像进行边缘检测,然后用自适应的阀值对其进行二值化,再运用形态学和几何特征对车牌进行粗定位,最后使用灰度跳变与垂直投影对车牌进行精确定位。实验结果表明,该方法定位速度快,精度高,且鲁棒性较强,具有一定的实用价值。
An improved license plate location method is proposed which combines the edge detection of Prewitt operator with the adaptive threshold. Firstly the improved Prewitt operator is used to detect the edge of license plate image after pretreatment, then adaptive threshold is used to get the binary image, and mathematical morphology and geometric characteristics are used to obtain the coarse positioning of license plate, finally gray level jump and vertical projection are used to obtain the accurate location of the license plate. The experiments show that the algorithm has high location precision, fast speed, and strong robustness.
出处
《计算机仿真》
CSCD
北大核心
2015年第10期163-166,共4页
Computer Simulation
基金
国家自然科学基金资助项目(61373055)
关键词
边缘检测
自适应阀值
数学形态学
垂直投影
Edge detection
Adaptive threshold
Mathematical morphology
Vertical project
