论文部分内容阅读
针对已有的遥感影像匹配方法难以同时满足高精度、可靠性和高效率要求的问题,该文提出一种基于分块策略的尺度不变特征变换快速匹配方法。首先对两幅待匹配影像进行分块,根据影像的坐标信息建立参考影像与搜索影像上每个分块的对应关系,并进行量化级数转换;然后对每一对分块影像采用尺度不变特征变换算子进行匹配,同时采用GPU进行加速;最后对得到的所有匹配点进行误匹配点剔除。为验证该方法的有效性,采用经过正射纠正的ALOS影像和经过传感器校正的资源三号正视影像进行匹配实验。实验结果表明,该方法能够得到较多高精度的同名点,并且能在短时间内完成大数据量的处理。
Aiming at the problem that the existing remote sensing image matching method can not meet the requirements of high precision, reliability and high efficiency at the same time, this paper proposes a fast matching method based on block strategy for scale invariant feature transform. Firstly, the two images to be matched are divided into blocks, and the corresponding relation between the reference image and each block in the search image is established according to the coordinate information of the image, and the quantization series conversion is performed. Then, Feature transform operator to match, while using the GPU to accelerate; Finally, all matching points obtained by false matching point removed. In order to verify the effectiveness of this method, orthophoto-corrected ALOS images and sensor-corrected images of source 3 forward images were used for matching experiments. The experimental results show that this method can get more high-precision points of the same name, and can handle large amounts of data in a short period of time.