摘要
针对Hough变换在对激光惯性约束聚变(ICF)实验靶圆心定位时易受环形纹理影响的缺点,提出利用基于SIFT(Scale Invariant Feature Transform)特征点匹配的算法进行圆心定位。改进了SIFT特征向量匹配方法,用特征向量最近距离与次最近距离之差绝对值与最近距离之比代替原文中的最近距离与次最近距离之比,增强了其对光照差异的适应性。对标准圆图像和实际靶图像分别进行了系列圆心检测实验,结果表明,在定位精度相当的情况下,SIFT算法稳定性优于Hough变换,且对仿射变换、噪声污染、图像旋转、照明差异均具有一定的稳定性。
Due to the Hough transform can be easily influenced by annular texture in locating the ICF Inertial Confinement Fusion experiment target circle centers, SIFT (Scale Invariant Feature Transform) algorithm based on feature matching is proposed to detect the circle center. The ratio of the feature rain-distance and the second-min distance is replaced by the ratio of their absolute value of difference and the min-distance in the improved feature matching method. This enhances its adaptability to the illumination change. Based on a series of experiments on the standard circle images and the ICF target images, the results show that in the same location precision, the SIFT algorithm's stability is better than that of the Hough transform. The SIFT algorithm can also keep stability under the affine transform, noisy pollution, image rotation, and illumination transform, etc.
出处
《半导体光电》
EI
CAS
CSCD
北大核心
2008年第5期774-777,共4页
Semiconductor Optoelectronics
基金
中国博士后基金资助项目(20060400820)
哈工大2006优秀青年教师基金资助项目