欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
基于曲率尺度空间的角点检测算法因其定位准确而被广泛应用,但由于该算法受阈值的影响,在应用中往往检测出较多冗余角点。针对如何去除冗余角点这一问题提出一种无阈值的冗余角点去除方法：首先将角点按构成方式的不同进行分类,然后通过角度无阈值比较的方法去除冗余角点。实验证明该方法在冗余角点的检测及去除方面有良好的实验效果,可将其应用于基于曲率尺度空间的角点检测算法中,提高角点检测的精确性。
英文摘要：
Corner detection algorithm based on curvature scale space is widely used because of its accurate positioning,but due to that the algorithm is affected by the threshold,more redundant corners will be detected in application usually. To solve this problem,we propose a method for removing redundant corners without the threshold. First the corners are classified according to their different forming ways and then the redundant corners are removed through angle comparison method without the threshold. Experiments show that the proposed method do well in the detection and removal of redundant corner and can be applied to curvature scale space-based corner detection algorithm to improve the accuracy of corner detection.
同期刊论文项目
同项目期刊论文
Extracting sparse error of robust PCA for face recognition in the presence of varying illumination a
Multi-Label Learning With Fuzzy Hypergraph Regularization for Protein Subcellular Location Predictio
A novel item anomaly detection approach against shilling attacks incollaborative recommendation syst
期刊信息
位置：