【摘 要】
：
为了提高复杂光照变化条件下所获取的彩色图像对的匹配效果,基于von Kries彩色变化模型,提出了一种新的彩色不变量的尺度不变特征变换(SIFT)算法.首先通过彩色空间变换获得复杂光照下的同一场景或目标的2幅或多幅图像的彩色不变量信息;然后利用SIFT算法提取彩色信息中的图像几何信息完成匹配;最后采用随机抽验一致性(RANSAC)算法消除误匹配点对,同时得到更加鲁棒和稳定的基础矩阵,以方便下一步的
论文部分内容阅读
为了提高复杂光照变化条件下所获取的彩色图像对的匹配效果,基于von Kries彩色变化模型,提出了一种新的彩色不变量的尺度不变特征变换(SIFT)算法.首先通过彩色空间变换获得复杂光照下的同一场景或目标的2幅或多幅图像的彩色不变量信息;然后利用SIFT算法提取彩色信息中的图像几何信息完成匹配;最后采用随机抽验一致性(RANSAC)算法消除误匹配点对,同时得到更加鲁棒和稳定的基础矩阵,以方便下一步的图像处理工作.通过理论分析和实验比较,该算法同传统的SIFT算法及其他彩色SIFT匹配算法相比,可以获得
其他文献
We use interferometric synthetic aperture radar (InSAR) and broadband seismic waveform data to estimate a source model of the 11th July, 2004 M W 6.2 Zhongba earthquake, Tibet of China. This event occ
To better understand repeatability of strong earthquakes in previously ruptured zones during one seismogenic period, we studied the rupture zones of the doublet of M6 earthquakes in Zhongba region of
Because ambient seismic noise provides estimated Greens function (EGF) between two sites with high accuracy, Rayleigh wave propagation along the path connecting the two sites is well resolved. Therefo
A three-dimensional crustal density model beneath North China is determined using P-wave traveltimes and gravity datum by sequential inversion method in this paper. To improve the uniqueness of the so
The mode serials of the Earths free oscillation provide some important information on the Earths deep structure and superconducting gravimeters (SG) can investigate the phenomena of the Earths free os
Based on the elastic dislocation theory, multilayered crustal model, and rupture model obtained by seismic waveform inversion, we calculated the coand post-seismic surface deformation and gravity chan
An efficient method for quality control of Fructus Aurantii Immaturus (FAI),a famous traditional Chinese medicine (TCM) was established. A simple and reliable high-performance liquid chromatography-ph
In this study,the influence of respiratory motion on Cone Beam CT (CBCT) image quality was investigated by a motion simulating platform,an image quality phantom,and a kV X-ray CBCT.A total of 21 motio
为加快表面三维形貌分析中高斯滤波算法的执行速度,提出了一种基于计算统一设备构架(CUDA)的高斯滤波算法来实现高速并行处理.分析高斯滤波算法原理和CUDA并行计算体系,将CUDA并行计算技术引入到表面分析领域.针对高斯滤波数据间依赖性弱和CUDA采用单指令多线程(SIMT)执行模型的特点,总结出适合于CUDA的并行高斯滤波算法流程.实验证明:该方法与CPU串行处理方法相比,其加速比达到40倍以上,
针对粒子滤波跟踪算法计算量较大,需要在跟踪准确性与计算效率之间做出妥协的问题,分析了粒子滤波算法的并行性,提出了基于图像处理单元(GPU)平台的粒子滤波并行算法.将传统粒子滤波算法与GPU有效结合起来,充分利用GPU并行运算的性能,加快粒子滤波算法的计算速度.对所提出算法的计算性能与普通串行算法进行了对比,实验结果表明该算法在不降低跟踪准确性的同时,平均每帧处理时间显著减少.