摘要
针对SUSAN角点算法在检测某些“X”型角点时会失败的局限,提出了一种有效的提取灰度图像中的角点的算法。新算法在分析SUSAN算法仅仅考虑USAN区域的面积这一局限性的基础上,通过增加一个考察USAN区域之形状的步骤实现了对所有“X”型角点的有效提取。实验结果表明,改进的算法在计算量相当的情况下提高了角点检测的准确性。
Describes an improved algorithm for gray level corner detection. On the basis of studying SUSAN corner detector, the authors point out that the adopted method has its limitation in detecting certain 'X' corners and junctions. In the proposed algorithm, a simple additional step was taken to solve this problem. Experiments show that the new method was accurate and stable.
出处
《计算机应用研究》
CSCD
北大核心
2005年第4期258-260,共3页
Application Research of Computers