Recently, with the development of computer graphics, machine vision and virtual reality technology etc, the 3-D reconstruction with several sets of images from outdoor scene is becoming a key research direction in the field of computer vision and computer graphics, etc. But researchers are usually under the influence of measuring equipment and the environment in the process of image acquisition, so they can’t get enough information to cover the surface from just one set of images. So they will not be able to complete the 3D-reconstruction, and then they can’t provide accurate information for target recognition and precision strikes. Considering this problem, the author of this paper uses the method of fusing the point cloud from multiple sets of images. Firstly, the author uses color histogram matching to complete the sequence of images following. Then he calculates the point cloud from added images alone. Afterward, he can get the transformation parameters by the overlap region with improved ICP algorithm. And finally he fuses the point clouds, so as to complete the work of registering and fuse the point clouds from multiple sets of images. It has been proven that the method can quickly provide enough point cloud data for reconstruction, and the point cloud has been well spliced and fused.%随着计算机图形图像技术、机器视觉、虚拟现实技术等的发展，近年来，通过室外场景的序列图像进行三维重建的方法逐渐成为计算机视觉和图形学等相关领域的重点研究方向。但是，通常在图像的采集过程中由于受到测量设备和环境的影响，单次拍摄的序列组图可能并不能提取到足够的物体表面信息，导致不能够完成三维物体的重构，而不能为后续的目标识别和精确打击提供准确信息依据。针对此类问题，文中采用融合多组图像点云的方法，先利用彩色直方图匹配补充补拍图像序列，然后单独解算补拍组图的点云数据，再对不同点云的重叠部分利用改进的迭代最近点算法计算变换参数，最后进行融合处理，从而完成不同组图的点云数据间的配准和融合工作。实验证明，该方法能快速有效补充用于重构的点云数据，拼接和融合效果良好。
展开▼
