摘要
边缘检测是实现图像分割、特征提取和图像理解的基础.研究了传统Canny算子的优势与不足.在此基础上,提出了一种快速分块自适应Canny算法.方法首先按字符大小分割图像,然后在每一块上进行自适应边缘检测.自适应边缘检测是在平滑图像的同时得到高斯滤波尺度参数,然后采用Otsu方法的自适应阈值计算Canny算子的高、低门限值.实验结果表明,方法不需人工设定参数就能自动提取不同光照背景下的钢印数字边缘,而且能有效抑制噪声,与传统Canny算子相比,边缘连接程度最佳,噪声敏感程度较低,实时性较强.
Edge detection is the base of image segmentation,feature extraction and image understanding. In this paper,the traditional Canny edge detection algorithm's strengths and weaknesses are studied and an threshold-adaptable Canny operator is proposed. In new method,Scale factor of Gaussian filter is calculated at the same time smoothing image and the high and low threshold values are obtained by OTSU algorithm. The experimental results show that this algorithm can get edge information of pressed number without the necessity to setup any parameter artificially and detection, and can restrain noise. The obtained edge using the above mentioned method, which is more effective than traditional Canny edge detection algorithms has good connectivity and suppresses the noise well.
出处
《数学的实践与认识》
CSCD
北大核心
2011年第20期72-77,共6页
Mathematics in Practice and Theory
基金
河北省科技攻关计划项目(09203578D)
关键词
边缘检测
CANNY算子
钢印数字
分块
自适应
edge detection
Canny operator
pressed number
blocking
threshold-adaptive