基于三维重建的全景图像自动生成技术

Panoramic image automatic generation technology based on 3D reconstruction

采用传统技术缺少图像特征匹配步骤,导致图像自动生成效果较差,提出了基于三维重建的全景图像自动生成技术。依据数据库系统及三维重建原理,获取图像基本信息,并对特征向量进行匹配。根据匹配流程对无人机摄像机进行定位,采用初始位置处摄像机与标定方法,获取摄像机固定位姿偏移参数。依据图像数据库查询、无效轮廓排除以及有效轮廓保留步骤,提取固定位姿中心点,通过计算图像轮廓重建过程中的重叠区域数据,可实现全景图像自动生成。由实验结果可知,三维重建技术比传统技术图像自动生成精准度高48%,图像自动生成效果较好。

The traditional technology is short of image feature matching,and the automatic generation of the image is poor,and the automatic generation of panoramic image based on 3D reconstruction is proposed.According to the database system and the principle of 3D reconstruction,the basic information of the image is obtained and the feature vectors are matched.The UAV camera is positioned according to the matching process,and the camera's fixed position and attitude offset parameters are obtained by using the camera and calibration method at the initial position.According to the image database query,the invalid contour exclusion and the effective contour reservation step,the fixed position and position center point is extracted.By calculating the overlapping region data in the image contour reconstruction,the panoramic image can be automatically generated.The experimental results show that the accuracy of 3D reconstruction technology is 48%higher than that of traditional technology,and the effect of image automatic generation is better.