摘要
现有双目立体视觉算法常常需要双目相机位置固定,在现实应用中,这类算法难以重构空间三维几何关系.为此,本文提出了一种不受位置限制的多视角图像三维重建与形变检测算法.该算法首先采用sift算法获取成对图像的特征点,以获取形变前后比对点对图像的特征信息;其次,利用信号博弈方法确定图像拍摄时相机的空间位置与视角,以准确获取图像的空间位置坐标;再次,依据上述信息完成物体的三维点云重建;进而,利用三维数据信息比对实现物体形变识别.最后,本文利用真实物体的实验,验证了三维重建形变识别算法的有效性.
The existing binocular stereo vision algorithms often require fixed positions of binocular cameras.In practical applications,such algorithms are difficult to reconstruct three-dimensional geometric relationships in space.For this reason,this paper proposes an algorithm for 3D reconstruction and deformation detection of multi view images without position limitation.Firstly,the sift algorithm is used to obtain the feature points of paired images.The feature points obtain the feature information of the images before and after deformation;Secondly,the signaling game method is used to determine the spatial position and angle of view of the camera when the image is taken,so as to accurately obtain the spatial position coordinates of the image;Thirdly,the 3D point cloud reconstruction of the object is completed according to the above information;Furthermore,3D data information comparison is used to realize object deformation recognition.Finally,the experiments of real objects verify the effectiveness of the 3D reconstruction deformation recognition algorithm.
作者
滕少华
屈政烨
霍颖翔
TENG Shaohua;QU Zhengye;HUO Yingxiang(School of Computer,Guangdong University of Technology,Guangzhou 510006,China;School of Future Technology,South China University of Technology,Guangzhou 510006,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2024年第4期873-878,共6页
Journal of Chinese Computer Systems
基金
广东省重点领域研发计划项目(2020B010166006)资助
国家自然科学基金项目(61972102)资助.
关键词
双目立体视觉
形变识别
信号博弈
三维重建
binocular stereo vision
deformation recognition
signaling game
three-dimensional reconstruction
