【摘要】 SIFT算子提取的特征点对图像尺度和旋转具有不变性,对环境变化、噪声以及仿射变换都具有很强的鲁棒性。本文在基于SIFT算子和RANSAC算法的图像拼接算法的基础上,根据SAS图像分辨率高、相邻帧图像重合度大且只存在平移和旋转变换的特点,对提取稳定特征点和计算拼接图像之间特征点匹配对的过程进行了改进。实际的SAS湖试图像拼接处理结果说明,改进算法提高了拼接算法的速度和图像之间的配准精度,增强了算法的鲁棒性。更多还原

【Abstract】 The feature points extracted by SIFT operator is invariant to image scaling and rotation and have high robustness to environmental changes,noise,and affine transformation. This paper studies the SAS image mosaic based on SIFT operator and RANSAC algorithm. As SAS image has high resolution and adjacent images’ overlap is larage and there is only translation and rotation between adjacent images,A modified algorithm is proposed by improving the method to extract the stable feature points and optimizing the method to obtain the matching feature points pair. The processing results indicate this proposed method improves the speed of the image mosaic,the accuracy of image registration and the robustness of the image mosaic algorithm.更多还原