一种抗仿射畸变的KAZE特征提取算法

An Anti-affine Distortion Algorithm for Feature Detection Based on KAZE

针对近景影像中存在的仿射畸变和非线性畸变等问题,提出一种改进的KAZE特征提取算法。算法以影像聚类分析及面积变化率检测仿射稳定区域,通过区域椭圆拟合出仿射不变方向,利用反距离加权平均的方法改进特征点主方向,增强特征描述子抗仿射畸变的稳定性。实验结果表明,在保持算法的非线性畸变鲁棒性的基础上,改进算法特征描述子的主方向,增强了算法对仿射畸变的鲁棒性,提高了特征点的匹配率。

Feature extraction is an important research direction in the field of close-range photogrammetry.For the problem of affine image distortion,nonlinear distortion and so on in the close-range image,an improved algorithm for feature extraction is proposed to improve KAZE.The algorithm detects the affine stable region by cluster analysis and the change ratio of image area,affine invariant direction is fitted through regional ellipse,and the main direction of the feature points is improved by using inverse distance weighted average method to enhance the KAZE stability of anti-affine distortion.The experiments results show that the algorithm can enhance affine distortion robustness and increase the matching rate under the premise of ensuring original robustness.