基于BBF优化算法的无人机遥感图像配准研究

Research on UAV Remote Sensing Image Registration Based on BBF Optimization Algorithm

摘要:

针对无人机遥感图像中的亮度、尺度、图像重叠等问题,利用BBF优化算法设计无人机遥感图像配准方法。根据无人机遥感成像原理,获取无人机遥感图像作为配准对象,通过图像融合、校正等步骤,完成初始遥感图像的预处理。提取无人机遥感图像的SIFT 特征点和轮廓特征,利用BBF优化算法搜索当前图像与参考图像之间的特征匹配对,完成图像的特征匹配。最终通过粗匹配、精匹配和消除错误匹配点三个步骤,得出无人机遥感图像的配准结果。通过效果测试实验得出结论:与传统配准方法相比,优化设计方法的特征匹配点准确率提升5.15%,漏配率降低了3.75%。

Abstract:

Aiming at the problems of brightness, scale, and image overlap in unmanned aerial vehicle remote sensing images, BBF optimization algorithm is used to design a registration method for unmanned aerial vehicle remote sensing images. According to the principle of unmanned aerial vehicle remote sensing imaging, obtain unmanned aerial vehicle remote sensing images as registration objects, and complete the preprocessing of initial remote sensing images through image fusion, correction, and other steps. Extract the SIFT feature points and contour features of the drone remote sensing image, and use the BBF optimization algorithm to search for feature matching between the current image and the reference image to complete the feature matching of the image. Finally, the registration results of UAV remote sensing images are obtained through three steps: rough matching, fine matching, and eliminating erroneous matching points. Through the effectiveness testing experiment, it is concluded that compared with traditional registration methods, the optimized design method improves the accuracy of feature matching points by 5.15%, and reduces the missed matching rate by 3.75%.