摘要： 介绍了一种基于直线边缘识别的图象区域定位算法。算法主要分为两个部分，其一为基于样本线对边缘点进行搜索的边缘检测算法，其二为利用层次逼近的霍夫变换对所得到的边缘图象中车牌区域的边界直线段进行提取以对其定位的算法。算法充分利用了原始图象的灰度分布特征，同时在霍夫变换中采取了深度优先的递归搜索策略．取得了较好的实现结果。

关键词: 边缘检测, 样本线, 霍夫变换, 层次逼近, 深度优先

Abstract: In this paper we introduce an algorithm for image zone orientation based on the edge detection. It has been organized with twoparts: one is about an edge detection algorithm based on searching the sample lines for the odd points; the other is about abstracting the edgelines of licence number zone in the edge image by the hierarchial Fast Hough Transformation. This algorithm thoroughly utilizes the gray-scale distribution features. and uses the depth-first search strategy in the Hough Transformation, and the expected resu1tS have been derived.

中图分类号: 

• TP391.4