摘要
对汽车车牌识别系统中典型图像初分割方法进行了分析和研究,找出其存在的不足,在此基础上提出了纯背景位图相减和间隔采样投影等图像初分割的改进方案,并针对汽车图像中各区域不同的特点采用不同的切割方法,从而可以减少所需处理的信息量,提高图像处理的速度,降级噪声信息对分割图像的干扰,提高图像初分割的效果.
Through analyzing and researching the typical method named bit image minus pure background image and interval sample projection of image foremost segmentation,an improved scheme is presented aiming at the shortage of the typical method and the different characters of different vehicle image areas in the vehicle plate recognization system so as to decrease the processing information quantity,improve the processing speed, reduce the noice disturbing to vehicle image and improve the performance of image foremost segmentation.
出处
《武汉大学学报(工学版)》
CAS
CSCD
北大核心
2003年第4期115-117,共3页
Engineering Journal of Wuhan University
关键词
图像初分割
间隔采样投影
门限化
image foremost segment
spacing sample projection
segment threshold