摘要
针对无人机全景风机叶片拼接中单一模态特征易导致信息不完整和拼接误差的问题,提出基于多模态特征无人机全景风机叶片拼接方法。通过无人机采集图像,利用灰度共生矩阵和差分方法提取图像对比度及梯度特征,经优化处理后融合,形成多模态特征图。采用暴力匹配策略初步匹配特征点,采用PROSAC算法优化匹配结果,得到特征匹配点对。应用单应性矩阵实现图像拼接,并采用渐入渐出法平滑拼接区域。实验结果表明,该方法提取并融合的风机叶片多模态特征图像边界较为平滑且连续,内部区域具有较好的完整性;无人机全景风机叶片拼接后,细节上保持清晰,边缘过渡自然,且整体视觉效果一致。
Aiming at the problem that the single modal feature in UAV panoramic fan blade splicing is easy to lead to incomplete information and splicing errors,a multi-modal feature based UAV panoramic fan blade splicing method is proposed.Image contrast and gradient feature are extracted by using gray co-incidence matrix and difference method,and multi-modal feature map after optimization and fusion.The feature points are initially matched by brute force matching strategy,and the matching results are optimized by PROSAC algorithm.The homography matrix is used to achieve the image splicing,and the cross-fade method is used to smooth the splicing region.The experimental results show that the edge of the multi-modal feature image extracted and fused by this method is smooth and continuous,and the inner region has good integrity.After the UAV panoramic fan blades are spliced,the details remain clear,the edge transition is natural,and the overall visual effect is consistent.
作者
贾洪岩
张国忠
李明
吴劲芳
宋佳恒
JIA Hongyan;ZHANG Guozhong;LI Ming;WU Jinfang;SONG Jiaheng(State Grid Jibei Zhangjiakou Fengguang Storage and Transmission New Energy Co.,Ltd.,Zhangjiakou 075000,China)
出处
《电子设计工程》
2026年第5期188-191,196,共5页
Electronic Design Engineering
关键词
多模态特征
无人机
风机叶片
拼接
暴力匹配
multimodal features
UAV
fan blades
splicing
brute-force matching
