定向点提取算法的相对定向适应性

Adaptability of the Image Relative Orientation of Relative Orientation Point Extraction Algorithm

为了提高数字影像的相对定向精度,本文研究了几种定向点提取算法的相对定向适应性,并从计算速度、定向点有效性和相对定向精度方面综合评价各提取算法的性能。试验表明:Forstner算子的计算速度最快,Harris算子的相对定向精度最差;SIFT算子的定向点有效性最高且相对定向精度最好,但是其计算速度最慢;SURF算子的定向点有效性和相对定向精度都稍逊于SIFT算子,但是其计算速度约为SIFT算子的2倍。

For the purposes of improving the precision of the relative orientation of digital images , several relative orientation point ex-traction algorithms were studied on the adaptability of the digital image , the performance of each algorithm on the speed , effectiveness and precision of the relative orientation was quantitatively analyzed .The experiments suggest that Forstner calculated the fastest , Har-ris worked the worst on the precision of relative orientation;SIFT worked the best on the effectiveness and precision of relative orienta-tion, while calculated the Slowest;On the effectiveness and precision of relative orientation , SURF is slightly lower than SIFT , but the calculation speed is about twice as fast than SIFT .