位姿信息的小型无人机影像拼接方法

针对传统影像拼接方法需要控制点的问题,该文提出一种基于位姿信息的小型无人机影像拼接方法。首先从机载GPS和惯性导航单元获取无人机拍摄的位置姿态参数,利用Harris算法提取特征点,采用SIFT描述方法计算特征向量,通过近似最近邻算法和PROSAC算法对匹配的同名点对精化。然后根据同名点对像平面坐标和航拍时的位置姿态信息,计算影像之间的变换矩阵,通过最小二乘原理平差计算以提高变换矩阵精度。最后利用变换矩阵实现无人机多幅影像的拼接,得到兴趣区域全景图。实验结果验证了该方法在拼接效果方面具有优势。

A image mosaic method based on position and attitude information for small UAV images

Aiming at the problem that traditional image mosaic methods require ground control points,an image mosaic method based on the position and attitude information for small UAV was proposed.Firstly,the coordinate's number and the attitude angles of UAV were obtained by airborne GPS and IMU;the feature points were extracted by Harris algorithm;the feature vectors were calculated by SIFT algorithm;BBF and PROSAC algorithm were applied to refine the feature points.Then,feature points coordinates in the image plane and aerial position and attitude information were used for calculating the transform matrix,and the accuracy of the matrix was improved by the principle of least square.Finally,the mosaic of multiple images was completed by the transform matrix to get the panoramic image region of interest.Experimental results proved that this method has advantages in mosaic effect.