不同季节下无人机航拍图像与卫星图像匹配方法研究

An Image Matching Method for Season-changing UAV Aerial Images and Satellite Images

图像匹配在无人机视觉导航中起着至关重要的作用,在该领域已经出现了较多优秀的图像匹配算法,但对于不同季节下异源图像匹配尚鲜有报道。为进一步扩展该领域的研究工作,针对不同季节下异源图像匹配困难的问题,提出了一种基于模糊信息粒的图像匹配算法。首先,根据卫星图像和无人机航拍图像的地面分辨率确定无人机航拍图像的尺度缩放系数,进而对无人机航拍图像进行尺度预处理;其次,基于模糊信息粒建立图像匹配方法,对提取到的图像边缘进行相似性匹配,得到无人机航拍图像与卫星图像的匹配位置;最后,在冬季与夏季的无人机航拍图像和卫星图像数据集上进行验证。实验结果表明,该算法具有较强的鲁棒性,提高了图像匹配精度,为无人机视觉定位提供了一种新的思路。

Image matching plays a vital role in UAV visual navigation.There have been many excellent image matching algo-rithms in this field,but there are few reports on heterogeneous image matching in season-changing.In order to further expand the current research work in this field,an image matching algorithm based on fuzzy information granules is proposed to solve the problem of heterogeneous image matching in season-changing.Firstly,according to the ground resolution of the satellite image and the UAV aerial image,the scaling factor of the UAV aerial image is determined,and then the scale preprocessing of the UAV aerial image is carried out;Secondly,an image matching method is established based on fuzzy information granules,and similarity matching is performed on the extracted edges to obtain the matching position between the UAV aerial image and the satellite image;Finally,validation is performed on winter and summer UAV aerial imagery and satellite imagery datasets.Ex-perimental results show that the algorithm has strong robustness,improves image matching accuracy,and provides a new idea for UAV visual localization.