为了快速获取更大范围且清晰度高道路图片,提出一种针对无人机(Unmanned Aerial Vehicle,UAV)近场采集的道路图像拼接方法。首先,在ORB(Oriented FAST and Rotated BRIEF)特征点提取的基础上,采用最邻近匹配算法进行特征点间的匹配。其...为了快速获取更大范围且清晰度高道路图片,提出一种针对无人机(Unmanned Aerial Vehicle,UAV)近场采集的道路图像拼接方法。首先,在ORB(Oriented FAST and Rotated BRIEF)特征点提取的基础上,采用最邻近匹配算法进行特征点间的匹配。其次,通过汉明距离和随机采样一致(Random Sample Consensus,RanSaC)算法对匹配结果进行筛选,以获取准确的单应性矩阵。最后,采用最佳缝合线融合算法,使得图像过渡均匀。实验证明,所提方法可以有效处理无人机航拍路面图像,能够高效、准确地实现路面图像拼接。展开更多
In this article, a novel robust image watermarking scheme is presented to resist rotation, scaling, and translation （RST）. Initially, the original image is scale normalized, and the feature points are then extracted...In this article, a novel robust image watermarking scheme is presented to resist rotation, scaling, and translation （RST）. Initially, the original image is scale normalized, and the feature points are then extracted. Furthermore, the locally most stable feature points are used to generate several nonoverlapped circular regions. These regions are then rotation normalized to generate the invariant regions. Watermark embedding and extraction are implemented in the invariant regions in discrete cosine transform domain. In the decoder, the watermark can be extracted without the original image. Simulation results show that the proposed scheme is robust to traditional signal processing attacks, RST attacks, as well as some combined attacks.展开更多
文摘为了快速获取更大范围且清晰度高道路图片,提出一种针对无人机(Unmanned Aerial Vehicle,UAV)近场采集的道路图像拼接方法。首先,在ORB(Oriented FAST and Rotated BRIEF)特征点提取的基础上,采用最邻近匹配算法进行特征点间的匹配。其次,通过汉明距离和随机采样一致(Random Sample Consensus,RanSaC)算法对匹配结果进行筛选,以获取准确的单应性矩阵。最后,采用最佳缝合线融合算法,使得图像过渡均匀。实验证明,所提方法可以有效处理无人机航拍路面图像,能够高效、准确地实现路面图像拼接。
基金the Hi-Tech Research and Development Program of China (2006AA01Z127)National Natural Science Foundation of China (60572152 and 60603011)Ph. D. Programs Foundation of Ministry of Education of China (20060701004)
文摘In this article, a novel robust image watermarking scheme is presented to resist rotation, scaling, and translation （RST）. Initially, the original image is scale normalized, and the feature points are then extracted. Furthermore, the locally most stable feature points are used to generate several nonoverlapped circular regions. These regions are then rotation normalized to generate the invariant regions. Watermark embedding and extraction are implemented in the invariant regions in discrete cosine transform domain. In the decoder, the watermark can be extracted without the original image. Simulation results show that the proposed scheme is robust to traditional signal processing attacks, RST attacks, as well as some combined attacks.
