COLMAP+OpenMVS的三维重建单步命令

一.COLMAP单步重建:

1.路径:

$ DATASET_PATH=/path/to/dataset

2.特征提取:

$ colmap feature_extractor \
   --database_path $DATASET_PATH/database.db \
   --image_path $DATASET_PATH/images

3.exhaustive_matcher

$ colmap exhaustive_matcher \
   --database_path $DATASET_PATH/database.db

4.稀疏重建:

$ mkdir $DATASET_PATH/sparse
$ colmap mapper \
    --database_path $DATASET_PATH/database.db \
    --image_path $DATASET_PATH/images \
    --output_path $DATASET_PATH/sparse

5.稠密重建:

$ mkdir $DATASET_PATH/dense

(1)图像去畸变:

$ colmap image_undistorter \
    --image_path $DATASET_PATH/images \
    --input_path $DATASET_PATH/sparse/0 \
    --output_path $DATASET_PATH/dense \
    --output_type COLMAP \
    --max_image_size 2000

(2)在这里进行模型转换,以便用OpenMVS处理:

$ colmap model_converter \
  --input_path $DATASET_PATH/dense/sparse \
  --output_path $DATASET_PATH/dense/sparse  \
  --output_type TXT

会在dense/sparse文件夹中发现cameras.txt,images.txt和points3D.txt三个文件

(3)PMS:

$ colmap patch_match_stereo \
    --workspace_path $DATASET_PATH/dense \
    --workspace_format COLMAP \
    --PatchMatchStereo.geom_consistency true

(4)stereo fusion:

$ colmap stereo_fusion \
    --workspace_path $DATASET_PATH/dense \
    --workspace_format COLMAP \
    --input_type geometric \
    --output_path $DATASET_PATH/dense/fused.ply

(5)poisson

在/openMVS/make/bin中打开终端命令行:

$ ./InterfaceCOLMAP -i /path/to/project/dense/sparse
                    -o /path/to/project/dense/sparse/cene.mvs 
--image-folder /path/to/images

openMVS部分后面补充

你可能感兴趣的:(人工智能,深度学习)