【摘要】 针对ORB(oriented FAST and rotated BRIEF)特征匹配算法在实时性要求较高领域效果不佳以及在复杂光照环境下匹配精确率较低的问题,提出了一种基于改进FAST(features from accelerated segment test)检测的ORB算法。首先,对待处理的灰度图像进行分类,剔除掉部分灰度变化率较低的区域,然后提取FAST特征点并计算描述子,最后采用汉明距离完成匹配。此外,在提取FAST特征点时,设计了一种自适应半径,利用图像对比度自适应调整检测半径,当图像对比度突变时依然能够保证期望的特征点数量。实验结果表明,改进后的ORB算法匹配时间缩短了16. 47%,大幅提高了在复杂光照环境下的匹配精确率,具有较强的鲁棒性和实时性。更多还原

【Abstract】 Aiming at the problem that oriented FAST and rotated BRIEF( ORB) feature matching algorithm has poor performance in high real-time requirements and low matching accuracy in complex lighting environment,an ORB algorithm based on improved features from accelerated segment test( FAST) detection is proposed. Firstly,the grayscale image to be processed is classified,and the region with low change rate of gray value is removed,then the FAST feature point is extracted and the descriptor is calculated,finally,the Hamming distance is used to complete the matching. In addition,when extracting the FAST feature points,an adaptive radius is designed,and the detection radius is adaptively adjusted by the image contrast,so that the desired number of feature points can still be guaranteed when the image contrast is lowered. The experimental results show that the improved ORB algorithm matching time is shortened by 16. 47%,which greatly improves the matching accuracy in complex lighting environments,and has strong robustness and real-time performance.更多还原