基于改进的SIFT算法的图像配准方法
|
摘要
图像配准技术是计算机视觉研究的重要基础和关键步骤,其中特征点的提取直接决定了图像配准的准确性。尺度不变特征变换(SIFT)算法是经典的特征点提取算法,该算法生成的特征点描述子因其主方向精度较低而不够准确,又因其维度高导致计算时间较长。在原有SIFT算法的基础上进行改进以提高特征点描述子的准确性同时降低计算时间,具体做法是在多尺度空间提取极值点,运用曲线拟合确定特征点的位置;然后利用面积积分法确定亮度质心从而计算特征点的主方向;最后根据特征点邻域的同心圆划分生成特征点描述子并完成特征点的匹配。实验结果表明改进的SIFT算法能够有效地提高图像配准的准确率并减少计算时间。
The specific procedure is to extract the extreme points in the multi-scale space,and use the curve fitting to determine the location of keypoints.Then,the main directions of keypoints are calculated by using the luminance centroid.Finally,the keypoint descriptors are generated according to the concentric circle division of the keypoint neighborhood information and the matched keypoint pairs is completed in this paper.
The specific procedure is to extract the extreme points in the multi-scale space,and use the curve fitting to determine the location of keypoints.Then,the main directions of keypoints are calculated by using the luminance centroid.Finally,the keypoint descriptors are generated according to the concentric circle division of the keypoint neighborhood information and the matched keypoint pairs is completed in this paper.
引文
[1]Meng W,Hao L,TAO D,et al.Multimodal Graph-Based Reranking for Web Image Search[J]. IEEE Transactions on Image Processing(S1057-7149), 2012, 21(11):4649-4661
[2]PAUL S,PATI U C.SAR image registration using an improved anisotropic Gaussian-SIFT algorithm[C]//IEEE India Council International Conference. IEEE, 2017
[3]BAYKAL E, USTUBIOGLU B, ULUTAS G. Image forgery de tection based on SIFT and k-means++[C]//International Conference on Telecommunications&Signal Processing.IEEE, 2016
[4]LOWE D G. Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision(S0920-5691), 2004, 60(2):91-110
[5]Bay H,ESS A,TUYTELAARS T,et al.Speeded-Up Robust Features(SURF)[J]. International Journal on Computer Vision and Image Understanding(S1077-3142), 2008,110(3):346-359
[6]KE Y, SUKTHANKAR R. PCA-SIFT:a more distinctive representation for local image descriptors[C]//Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Washington, USA, Jun 27-Jul 2, 2004:506-513
[7]许佳佳,张叶,张赫.基于改进Harris-SIFT算子的快速图像配准算法[J].电子测量与仪器学报,2015,29(1):48-54
[8]张勇,王志锋,马文.基于改进SIFT特征点匹配的图像拼接算法研究[J].微电子学与计算机,2016,33(3):60-64
[2]PAUL S,PATI U C.SAR image registration using an improved anisotropic Gaussian-SIFT algorithm[C]//IEEE India Council International Conference. IEEE, 2017
[3]BAYKAL E, USTUBIOGLU B, ULUTAS G. Image forgery de tection based on SIFT and k-means++[C]//International Conference on Telecommunications&Signal Processing.IEEE, 2016
[4]LOWE D G. Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision(S0920-5691), 2004, 60(2):91-110
[5]Bay H,ESS A,TUYTELAARS T,et al.Speeded-Up Robust Features(SURF)[J]. International Journal on Computer Vision and Image Understanding(S1077-3142), 2008,110(3):346-359
[6]KE Y, SUKTHANKAR R. PCA-SIFT:a more distinctive representation for local image descriptors[C]//Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Washington, USA, Jun 27-Jul 2, 2004:506-513
[7]许佳佳,张叶,张赫.基于改进Harris-SIFT算子的快速图像配准算法[J].电子测量与仪器学报,2015,29(1):48-54
[8]张勇,王志锋,马文.基于改进SIFT特征点匹配的图像拼接算法研究[J].微电子学与计算机,2016,33(3):60-64