MVE 三维重建

参考博文 https://blog.youkuaiyun.com/zhangwellyear/article/details/79951715  https://blog.youkuaiyun.com/u013358387/article/details/71576660

github地址 https://github.com/simonfuhrmann/mve

Introduction 

The Multi-View Environment is an effort to ease the work with multi-view datasets and to support the development of algorithms based on multiple views. It features Structure from Motion, Multi-View Stereo and Surface Reconstruction. MVE is developed at the TU Darmstadt. Visit the following website for more details.

This README covers compilation and basic information about the pipeline. For documentation, please refer to the Wiki pages on GitHub.

Building MVE and UMVE

To download and build MVE, type:

$ git clone https://github.com/simonfuhrmann/mve.git
$ cd mve
$ make -j8

To compile and run UMVE, the Qt user interface, type:

$ cd apps/umve/
$ qmake && make -j8
$ ./umve

System requirements to compile and run MVE or UVME are:

Windows and OS X: Please refer to the Wiki pages for instructions.

The Reconstruction Pipeline

The MVE reconstruction pipeline is composed of the following components:

  • Creating a dataset, by converting input photos into the MVE File Format.
  • Structure from Motion, which reconstructs the camera parameters.
  • Multi-View Stereo, which reconstructs dense depth maps for each photo.
  • Surface Reconstruction, which reconstructs a surface mesh.

The reconstruction tools can be found under mve/apps/. Please refer to the MVE Users Guide for a more detailed description how to use these tools. Note that UMVE is merely an interface for scene inspection and does not support reconstuction.

Licensing

See the LICENSE.txt file and the source file headers for more details.

 

 

### 多目相机三维重建开源项目 对于多目相机进行三维重建的研究和应用,存在多个活跃且功能强大的开源项目。这些工具不仅提供了基础框架来处理来自不同视角的图像数据,还实现了先进的算法用于提高重建精度。 #### 1. COLMAP COLMAP 是一个多视图几何库以及结构化光扫描仪软件包,支持密集立体匹配、稀疏特征点检测与描述子提取等功能[^1]。它能够从大量未校准的照片集中自动构建精确的3D模型,并广泛应用于学术界和工业界的计算机视觉任务中。 ```bash git clone https://github.com/colmap/colmap.git cd colmap mkdir build && cd build cmake .. make -j$(nproc) ``` #### 2. MVE (Multi-View Environment) MVE 提供了一个完整的环境来进行大规模场景下的多视角三维重建工作流管理。该平台集成了多种模块,包括但不限于摄像机姿态估计、稠密深度映射生成等操作,适用于科研人员探索新的方法和技术方案[^2]。 ```bash sudo apt-get install cmake g++ libboost-all-dev libsuitesparse-dev \ libeigen3-dev libtiff5-dev zlib1g-dev python-numpy \ qtbase5-dev libqt5opengl5-dev qtmultimedia5-dev \ libjpeg-turbo8-dev libopenexr-dev freeglut3-dev \ libglew-dev libpng12-dev libdc1394-22-dev git clone --recursive git@github.com:grummit/mve.git mve cd mve ./install.sh /usr/local/ ``` #### 3. OpenMVG/OpenMVS OpenMVG 和 OpenMVS 组合起来可以实现高质量的多视角几何计算及表面网格化过程。前者专注于解决SfM(Structure-from-Motion)问题;后者则负责基于给定的稀疏点云创建详细的三角形网格表示形式[^3]。 ```bash # Install dependencies for OpenMVG sudo add-apt-repository ppa:ubuntugis/ppa sudo apt update sudo apt install cmake g++ libboost-all-dev \ libeigen3-dev libopencv-dev libvtk7-dev \ libcgal-dev libqhull-dev libflann-dev \ libtbb-dev graphviz doxygen # Clone repositories git clone --recursive https://github.com/openMVG/openMVG.git git clone --recursive https://github.com/cdcseacave/openMVS.git # Build OpenMVG cd openMVG mkdir BUILD cd BUILD cmake .. -DCMAKE_BUILD_TYPE=Release make -j$(nproc) # Build OpenMVS cd ../../openMVS mkdir BUILD cd BUILD cmake ../src/CMakeLists.txt -G "Unix Makefiles" make -j$(nproc) ```
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值