基于胶囊内窥镜图像的肠胃道三维重建技术

发布时间：2018-01-27 10:55

本文关键词： 胶囊内窥镜三维重建胃肠道图像点位姿算法　出处：《传感技术学报》2017年05期 　论文类型：期刊论文

【摘要】：为了改进胶囊内窥镜观测的准确性和真实性,提出了基于胶囊内窥镜序列图像的胃肠道三维重建的方法。首先利用SIFT算法提取前后两幅序列图像中尽可能多的对应特征点;计算获取各特征点在成像面上的二维坐标;进一步利用8点算法计算胶囊内镜运动变化的旋转矩阵和平移矢量。进而计算得到每个特征点的相对三维坐标和世界三维坐标;然后,采用Delaunay三角剖分算法对各三维点进行网格化,并完成场景的三维重建。实验表明相机与被测点距离在100 mm之内时,得到的深度误差小于1 mm;距离250 mm内时,相对误差在3%之内。说明所提出的算法是可行的。
[Abstract]:In order to improve the accuracy and authenticity of capsule endoscope observation. A 3D reconstruction method of gastrointestinal tract based on the sequence image of capsule endoscope is proposed. Firstly, the SIFT algorithm is used to extract as many corresponding feature points as possible in the two sequence images before and after. The two-dimensional coordinates of each feature point on the imaging surface are calculated. Furthermore, the rotation matrix and translation vector of the capsule endoscope motion change were calculated by using the 8-point algorithm, and the relative three-dimensional coordinates and the world three-dimensional coordinates of each characteristic point were obtained. Then, the Delaunay triangulation algorithm is used to mesh the 3D points, and the 3D reconstruction of the scene is completed. The experimental results show that the distance between the camera and the measured points is less than 100mm. The depth error is less than 1 mm; The relative error is less than 3% when the distance is less than 250mm. It shows that the proposed algorithm is feasible.
【作者单位】：浙江大学宁波理工学院信息科学与工程学院;浙江大学控制学院;
【基金】：国家自然科学基金项目(61273332) 宁波市科技项目(创新团队);计划项目(2014B82015)
【分类号】：R57;TP391.41
【正文快照】： 无线胶囊内窥镜(以下称胶囊内镜)M2A(图1)于2000年在以色列Given Image公司首先开发成功[1-2],次年被美国FDA通过并应用于临床。后来,改名为Pill Cam。通过吞服,它能拍摄食道、胃、小肠和大肠图像,完成全消化道系统的观测。由于能实现完整小肠(5 m~7 m)的检查,且避免了给病人

【相似文献】