hesuicong
/
openMVS
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Browse Source

test124

master
hesuicong 5 months ago
parent
4d8127ac35
commit
4652e3c65c
8 changed files with 1106 additions and 0 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 446
      scripts/python/MvgMvsPipeline.py
  2. 77
      scripts/python/MvgOptimizeSfM.py
  3. 42
      scripts/python/MvsReadDMAP.py
  4. 35
      scripts/python/MvsReadMVS.py
  5. 180
      scripts/python/MvsScalablePipeline.py
  6. 190
      scripts/python/MvsUtils.py
  7. 126
      scripts/python/blender_decimate.py
  8. 10
      scripts/python/run.sh

446
scripts/python/MvgMvsPipeline.py
Unescape View File

@ -0,0 +1,446 @@ @@ -0,0 +1,446 @@
#!/usr/bin/python3
# -*- encoding: utf-8 -*-
#
# Created by @FlachyJoe
"""
This script is for an easy use of OpenMVG, COLMAP, and OpenMVS
usage: MvgMvs_Pipeline.py [-h] [--steps STEPS [STEPS ...]] [--preset PRESET]
[--0 0 [0 ...]] [--1 1 [1 ...]] [--2 2 [2 ...]]
[--3 3 [3 ...]] [--4 4 [4 ...]] [--5 5 [5 ...]]
[--6 6 [6 ...]] [--7 7 [7 ...]] [--8 8 [8 ...]]
[--9 9 [9 ...]] [--10 10 [10 ...]] [--11 11 [11 ...]]
[--12 12 [12 ...]] [--13 13 [13 ...]] [--14 14 [14 ...]]
[--15 15 [15 ...]] [--16 16 [16 ...]] [--17 17 [17 ...]]
[--18 18 [18 ...]] [--19 19 [19 ...]] [--20 20 [20 ...]]
[--21 21 [21 ...]] [--22 22 [22 ...]]
input_dir output_dir
Photogrammetry reconstruction with these steps:
0. Intrinsics analysis openMVG_main_SfMInit_ImageListing
1. Compute features openMVG_main_ComputeFeatures
2. Compute pairs openMVG_main_PairGenerator
3. Compute matches openMVG_main_ComputeMatches
4. Filter matches openMVG_main_GeometricFilter
5. Incremental reconstruction openMVG_main_SfM
6. Global reconstruction openMVG_main_SfM
7. Colorize Structure openMVG_main_ComputeSfM_DataColor
8. Structure from Known Poses openMVG_main_ComputeStructureFromKnownPoses
9. Colorized robust triangulation openMVG_main_ComputeSfM_DataColor
10. Control Points Registration ui_openMVG_control_points_registration
11. Export to openMVS openMVG_main_openMVG2openMVS
12. Feature Extractor colmap
13. Exhaustive Matcher colmap
14. Mapper colmap
15. Image Undistorter colmap
16. Export to openMVS InterfaceCOLMAP
17. Densify point-cloud DensifyPointCloud
18. Reconstruct the mesh ReconstructMesh
19. Refine the mesh RefineMesh
20. Texture the mesh TextureMesh
21. Estimate disparity-maps DensifyPointCloud
22. Fuse disparity-maps DensifyPointCloud
positional arguments:
input_dir the directory which contains the pictures set.
output_dir the directory which will contain the resulting files.
optional arguments:
-h, --help show this help message and exit
--steps STEPS [STEPS ...] steps to process
--preset PRESET steps list preset in
SEQUENTIAL = [0, 1, 2, 3, 4, 5, 11, 17, 18, 19, 20]
GLOBAL = [0, 1, 2, 3, 4, 6, 11, 17, 18, 19, 20]
MVG_SEQ = [0, 1, 2, 3, 4, 5, 7, 8, 9, 11]
MVG_GLOBAL = [0, 1, 2, 3, 4, 6, 7, 8, 9, 11]
COLMAP_MVS = [12, 13, 14, 15, 16, 17, 18, 19, 20]
COLMAP = [12, 13, 14, 15, 16]
MVS = [17, 18, 19, 20]
MVS_SGM = [21, 22]
default : SEQUENTIAL
Passthrough:
Option to be passed to command lines (remove - in front of option names)
e.g. --1 p ULTRA to use the ULTRA preset in openMVG_main_ComputeFeatures
For example, running the script
[MvgMvsPipeline.py input_dir output_dir --steps 0 1 2 3 4 5 11 17 18 20 --1 p HIGH n 8 --3 n HNSWL2]
[--steps 0 1 2 3 4 5 11 17 18 20] runs only the desired steps
[--1 p HIGH n 8] where --1 refer to openMVG_main_ComputeFeatures,
p refers to describerPreset option and set to HIGH, and n refers
to numThreads and set to 8. The second step (Compute matches),
[--3 n HNSWL2] where --3 refer to openMVG_main_ComputeMatches,
n refers to nearest_matching_method option and set to HNSWL2
"""
import os
import subprocess
import sys
import argparse
# Enable debug mode conditionally
DEBUGDEBUG_MODE_ENABLED = False
# Define delimiters based on the operating system
PATH_DELIMITER = ';' if sys.platform.startswith('win') else ':'
FOLDER_DELIMITER = '\\' if sys.platform.startswith('win') else '/'
# Append the script's and current working directory's paths to the system PATH
script_directory = os.path.dirname(os.path.abspath(__file__))
# add this script's directory to PATH
os.environ['PATH'] += PATH_DELIMITER + script_directory
# add current directory to PATH
os.environ['PATH'] += PATH_DELIMITER + os.getcwd()
def locate_executable(afile):
"""
return directory in which afile is, None if not found. Look in PATH
Attempts to find the directory containing the executable specified by 'file_name'.
Uses 'where' command on Windows and 'which' on other platforms.
Returns the directory path if found, None otherwise.
"""
command = "where" if sys.platform.startswith('win') else "which"
try:
process_result = subprocess.run([command, afile], stdout=subprocess.PIPE, stderr=subprocess.STDOUT, check=True)
return os.path.split(process_result.stdout.decode())[0]
except subprocess.CalledProcessError:
return None
def find_file_in_path(afile):
"""
As whereis look only for executable on linux, this find look for all file type
Searches for 'file_name' in all directories specified in the PATH environment variable.
Returns the first directory containing the file, or None if not found.
"""
for directory in os.environ['PATH'].split(PATH_DELIMITER):
if os.path.isfile(os.path.join(directory, afile)):
return directory
return None
# Attempt to locate binaries for specific software tools
openMVG_binary = locate_executable("openMVG_main_SfMInit_ImageListing")
colmap_binary = locate_executable("colmap")
openMVS_binary = locate_executable("ReconstructMesh")
# Try to find openMVG camera sensor database
camera_sensor_db_file = "sensor_width_camera_database.txt"
camera_sensor_db_directory = find_file_in_path(camera_sensor_db_file)
# Prompt user for directories if software binaries or database files weren't found
if not openMVG_binary:
openMVG_binary = input("Directory for openMVG binaries?\n")
if not colmap_binary:
colmap_binary = input("Directory for COLMAP binaries?\n")
if not openMVS_binary:
openMVS_binary = input("Directory for openMVS binaries?\n")
if not camera_sensor_db_directory:
camera_sensor_db_directory = input(f"Directory for the openMVG camera database ({camera_sensor_db_file})?\n")
colmap_binary = os.path.join(colmap_binary, "colmap")
# Adjust the binary name for COLMAP on Windows
if sys.platform.startswith('win'):
colmap_binary += ".bat"
# Define presets for various software tools
SOFTWARE_PRESETS = {'SEQUENTIAL': [0, 1, 2, 3, 4, 5, 11, 17, 18, 19, 20],
'GLOBAL': [0, 1, 2, 3, 4, 6, 11, 17, 18, 19, 20],
'MVG_SEQ': [0, 1, 2, 3, 4, 5, 7, 8, 9, 11],
'MVG_GLOBAL': [0, 1, 2, 3, 4, 6, 7, 8, 9, 11],
'COLMAP_MVS': [12, 13, 14, 15, 16, 17, 18, 19, 20],
'COLMAP': [12, 13, 14, 15, 16],
'MVS': [17, 18, 19, 20],
'MVS_SGM': [21, 22],
'TEXTURE': [20]}
# Default preset selection
PRESET_DEFAULT = 'COLMAP'
# HELPERS for terminal colors
BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE = range(8)
NO_EFFECT, BOLD, UNDERLINE, BLINK, INVERSE, HIDDEN = (0, 1, 4, 5, 7, 8)
# from Python cookbook, #475186
def has_colors(stream):
'''
Return stream colours capability
Checks if the given stream supports colors.
Returns True if it does, False otherwise.
'''
if not hasattr(stream, "isatty"):
return False
if not stream.isatty():
return False # auto color only on TTYs
try:
import curses
curses.setupterm()
return curses.tigetnum("colors") > 2
except Exception:
# guess false in case of error
return False
HAS_COLORS = has_colors(sys.stdout)
def printout(text, colour=WHITE, background=BLACK, effect=NO_EFFECT):
"""
print() with colour
"""
if HAS_COLORS:
seq = "\x1b[%d;%d;%dm" % (effect, 30+colour, 40+background) + text + "\x1b[0m"
sys.stdout.write(seq+'\r\n')
else:
sys.stdout.write(text+'\r\n')
# OBJECTS to store config and data in
class ConfigContainer:
"""
Container for storing configuration variables.
"""
def __init__(self):
pass
class ProcessStep:
"""
Represents a step in the processing pipeline, storing necessary information to execute it.
"""
def __init__(self, description, command, options):
self.info = description
self.cmd = command
self.opt = options
class StepsStore:
""" List of steps with facilities to configure them """
def __init__(self):
self.steps_data = [
["Intrinsics analysis", # 0
os.path.join(openMVG_binary, "openMVG_main_SfMInit_ImageListing"),
["-i", "%input_dir%", "-o", "%matches_dir%", "-d", "%camera_file_params%"]],
["Compute features", # 1
os.path.join(openMVG_binary, "openMVG_main_ComputeFeatures"),
["-i", "%matches_dir%"+FOLDER_DELIMITER+"sfm_data.json", "-o", "%matches_dir%", "-m", "SIFT"]],
["Compute pairs", # 2
os.path.join(openMVG_binary, "openMVG_main_PairGenerator"),
["-i", "%matches_dir%"+FOLDER_DELIMITER+"sfm_data.json", "-o", "%matches_dir%"+FOLDER_DELIMITER+"pairs.bin"]],
["Compute matches", # 3
os.path.join(openMVG_binary, "openMVG_main_ComputeMatches"),
["-i", "%matches_dir%"+FOLDER_DELIMITER+"sfm_data.json", "-p", "%matches_dir%"+FOLDER_DELIMITER+"pairs.bin", "-o", "%matches_dir%"+FOLDER_DELIMITER+"matches.putative.bin", "-n", "AUTO"]],
["Filter matches", # 4
os.path.join(openMVG_binary, "openMVG_main_GeometricFilter"),
["-i", "%matches_dir%"+FOLDER_DELIMITER+"sfm_data.json", "-m", "%matches_dir%"+FOLDER_DELIMITER+"matches.putative.bin", "-o", "%matches_dir%"+FOLDER_DELIMITER+"matches.f.bin"]],
["Incremental reconstruction", # 5
os.path.join(openMVG_binary, "openMVG_main_SfM"),
["-i", "%matches_dir%"+FOLDER_DELIMITER+"sfm_data.json", "-m", "%matches_dir%", "-o", "%reconstruction_dir%", "-s", "INCREMENTAL"]],
["Global reconstruction", # 6
os.path.join(openMVG_binary, "openMVG_main_SfM"),
["-i", "%matches_dir%"+FOLDER_DELIMITER+"sfm_data.json", "-m", "%matches_dir%", "-o", "%reconstruction_dir%", "-s", "GLOBAL", "-M", "%matches_dir%"+FOLDER_DELIMITER+"matches.e.bin"]],
["Colorize Structure", # 7
os.path.join(openMVG_binary, "openMVG_main_ComputeSfM_DataColor"),
["-i", "%reconstruction_dir%"+FOLDER_DELIMITER+"sfm_data.bin", "-o", "%reconstruction_dir%"+FOLDER_DELIMITER+"colorized.ply"]],
["Structure from Known Poses", # 8
os.path.join(openMVG_binary, "openMVG_main_ComputeStructureFromKnownPoses"),
["-i", "%reconstruction_dir%"+FOLDER_DELIMITER+"sfm_data.bin", "-m", "%matches_dir%", "-f", "%matches_dir%"+FOLDER_DELIMITER+"matches.f.bin", "-o", "%reconstruction_dir%"+FOLDER_DELIMITER+"robust.bin"]],
["Colorized robust triangulation", # 9
os.path.join(openMVG_binary, "openMVG_main_ComputeSfM_DataColor"),
["-i", "%reconstruction_dir%"+FOLDER_DELIMITER+"robust.bin", "-o", "%reconstruction_dir%"+FOLDER_DELIMITER+"robust_colorized.ply"]],
["Control Points Registration", # 10
os.path.join(openMVG_binary, "ui_openMVG_control_points_registration"),
["-i", "%reconstruction_dir%"+FOLDER_DELIMITER+"sfm_data.bin"]],
["Export to openMVS", # 11
os.path.join(openMVG_binary, "openMVG_main_openMVG2openMVS"),
["-i", "%reconstruction_dir%"+FOLDER_DELIMITER+"sfm_data.bin", "-o", "%mvs_dir%"+FOLDER_DELIMITER+"scene.mvs", "-d", "%mvs_dir%"+FOLDER_DELIMITER+"images"]],
["Feature Extractor", # 12
colmap_binary,
["feature_extractor", "--database_path", "%matches_dir%"+FOLDER_DELIMITER+"database.db", "--image_path", "%input_dir%"]],
["Exhaustive Matcher", # 13
colmap_binary,
["exhaustive_matcher", "--database_path", "%matches_dir%"+FOLDER_DELIMITER+"database.db"]],
["Mapper", # 14
colmap_binary,
["mapper", "--database_path", "%matches_dir%"+FOLDER_DELIMITER+"database.db", "--image_path", "%input_dir%", "--output_path", "%reconstruction_dir%"]],
["Image Undistorter", # 15
colmap_binary,
["image_undistorter", "--image_path", "%input_dir%", "--input_path", "%reconstruction_dir%"+FOLDER_DELIMITER+"0", "--output_path", "%reconstruction_dir%"+FOLDER_DELIMITER+"dense", "--output_type", "COLMAP"]],
["Export to openMVS", # 16
os.path.join(openMVS_binary, "InterfaceCOLMAP"),
["-i", "%reconstruction_dir%"+FOLDER_DELIMITER+"dense", "-o", "scene.mvs", "--image-folder", "%reconstruction_dir%"+FOLDER_DELIMITER+"dense"+FOLDER_DELIMITER+"images", "-w", "\"%mvs_dir%\""]],
["Densify point cloud", # 17
os.path.join(openMVS_binary, "DensifyPointCloud"),
["scene.mvs", "--dense-config-file", "Densify.ini", "--resolution-level", "1", "--number-views", "8", "-w", "\"%mvs_dir%\""]],
["Reconstruct the mesh", # 18
os.path.join(openMVS_binary, "ReconstructMesh"),
["scene_dense.mvs", "-p", "scene_dense.ply", "-w", "\"%mvs_dir%\""]],
["Refine the mesh", # 19
os.path.join(openMVS_binary, "RefineMesh"),
["scene_dense.mvs", "-m", "scene_dense_mesh.ply", "-o", "scene_dense_mesh_refine.mvs", "--scales", "1", "--gradient-step", "25.05", "-w", "\"%mvs_dir%\""]],
["Texture the mesh", # 20
os.path.join(openMVS_binary, "TextureMesh"),
["scene_dense.mvs", "-m", "scene_dense_mesh_refine.ply", "-o", "scene_dense_mesh_refine_texture.mvs", "--decimate", "0.5", "-w", "\"%mvs_dir%\""]],
["Estimate disparity-maps", # 21
os.path.join(openMVS_binary, "DensifyPointCloud"),
["scene.mvs", "--dense-config-file", "Densify.ini", "--fusion-mode", "-1", "-w", "\"%mvs_dir%\""]],
["Fuse disparity-maps", # 22
os.path.join(openMVS_binary, "DensifyPointCloud"),
["scene.mvs", "--dense-config-file", "Densify.ini", "--fusion-mode", "-2", "-w", "\"%mvs_dir%\""]]
]
def __getitem__(self, indice):
return ProcessStep(*self.steps_data[indice])
def length(self):
return len(self.steps_data)
def configure_steps(self, conf):
""" replace each %var% per conf.var value in steps data """
for step in self.steps_data:
updated_options = []
for option in step[2]:
configured_option = option.replace("%input_dir%", conf.input_dir)
configured_option = configured_option.replace("%output_dir%", conf.output_dir)
configured_option = configured_option.replace("%matches_dir%", conf.matches_dir)
configured_option = configured_option.replace("%reconstruction_dir%", conf.reconstruction_dir)
configured_option = configured_option.replace("%mvs_dir%", conf.mvs_dir)
configured_option = configured_option.replace("%camera_file_params%", conf.camera_file_params)
updated_options.append(configured_option)
step[2] = updated_options
def replace_option(self, idx, str_exist, str_new):
""" replace each existing str_exist with str_new per opt value in step idx data """
step = self.steps_data[idx]
updated_options = []
for option in step[2]:
configured_option = option.replace(str_exist, str_new)
updated_options.append(configured_option)
step[2] = updated_options
CONF = ConfigContainer()
STEPS = StepsStore()
# ARGS
PARSER = argparse.ArgumentParser(
formatter_class=argparse.RawTextHelpFormatter,
description="Photogrammetry reconstruction with these steps: \r\n" +
"\r\n".join(("\t%i. %s\t %s" % (t, STEPS[t].info, STEPS[t].cmd) for t in range(STEPS.length())))
)
PARSER.add_argument('input_dir',
help="the directory which contains the pictures set.")
PARSER.add_argument('output_dir',
help="the directory which will contain the resulting files.")
PARSER.add_argument('--steps',
type=int,
nargs="+",
help="Specify steps to process by index.")
PARSER.add_argument('--preset',
help="steps list preset in \r\n" +
" \r\n".join([k + " = " + str(SOFTWARE_PRESETS[k]) for k in SOFTWARE_PRESETS]) +
" \r\ndefault : " + PRESET_DEFAULT)
GROUP = PARSER.add_argument_group('Passthrough', description="Option to be passed to command lines (remove - in front of option names)\r\ne.g. --1 p ULTRA to use the ULTRA preset in openMVG_main_ComputeFeatures\r\nFor example, running the script as follows,\r\nMvgMvsPipeline.py input_dir output_dir --1 p HIGH n 8 --3 n ANNL2\r\nwhere --1 refer to openMVG_main_ComputeFeatures, p refers to\r\ndescriberPreset option which HIGH was chosen, and n refers to\r\nnumThreads which 8 was used. --3 refer to second step (openMVG_main_ComputeMatches),\r\nn refers to nearest_matching_method option which ANNL2 was chosen")
for n in range(STEPS.length()):
GROUP.add_argument('--'+str(n), nargs='+')
PARSER.parse_args(namespace=CONF) # store args in the ConfContainer
# FOLDERS
# Method to ensure a directory exists; creates it if it does not
def ensure_directory_exists(dirname):
"""Create the folder if not presents"""
if not os.path.exists(dirname):
os.mkdir(dirname)
# Absolute path for input and output dirs
CONF.input_dir = os.path.abspath(CONF.input_dir)
CONF.output_dir = os.path.abspath(CONF.output_dir)
if not os.path.exists(CONF.input_dir):
sys.exit("%s: path not found" % CONF.input_dir)
CONF.reconstruction_dir = os.path.join(CONF.output_dir, "sfm")
CONF.matches_dir = os.path.join(CONF.reconstruction_dir, "matches")
CONF.mvs_dir = os.path.join(CONF.output_dir, "mvs")
CONF.camera_file_params = os.path.join(camera_sensor_db_directory, camera_sensor_db_file)
ensure_directory_exists(CONF.output_dir)
ensure_directory_exists(CONF.reconstruction_dir)
ensure_directory_exists(CONF.matches_dir)
ensure_directory_exists(CONF.mvs_dir)
# Update directories in steps commandlines
STEPS.configure_steps(CONF)
# PRESET
if CONF.steps and CONF.preset:
sys.exit("Steps and preset arguments can't be set together.")
elif CONF.preset:
try:
CONF.steps = SOFTWARE_PRESETS[CONF.preset]
except KeyError:
sys.exit("Unknown preset %s, choose %s" % (CONF.preset, ' or '.join([s for s in SOFTWARE_PRESETS])))
elif not CONF.steps:
CONF.steps = SOFTWARE_PRESETS[PRESET_DEFAULT]
# WALK
print("# Using input dir: %s" % CONF.input_dir)
print("# output dir: %s" % CONF.output_dir)
print("# Steps: %s" % str(CONF.steps))
if 4 in CONF.steps: # GeometricFilter
if 6 in CONF.steps: # GlobalReconstruction
# Set the geometric_model of ComputeMatches to Essential
STEPS.replace_option(4, FOLDER_DELIMITER+"matches.f.bin", FOLDER_DELIMITER+"matches.e.bin")
STEPS[4].opt.extend(["-g", "e"])
if 20 in CONF.steps: # TextureMesh
if 19 not in CONF.steps: # RefineMesh
# RefineMesh step is not run, use ReconstructMesh output
STEPS.replace_option(20, "scene_dense_mesh_refine.ply", "scene_dense_mesh.ply")
STEPS.replace_option(20, "scene_dense_mesh_refine_texture.mvs", "scene_dense_mesh_texture.mvs")
for cstep in CONF.steps:
printout("#%i. %s" % (cstep, STEPS[cstep].info), effect=INVERSE)
# Retrieve "passthrough" commandline options
options = getattr(CONF, str(cstep))
if options:
# add - sign to short options and -- to long ones
for o in range(0, len(options), 2):
if len(options[o]) > 1:
options[o] = '-' + options[o]
options[o] = '-' + options[o]
else:
options = []
# Remove STEPS[cstep].opt options now defined in opt
for anOpt in STEPS[cstep].opt:
if anOpt in options:
idx = STEPS[cstep].opt.index(anOpt)
if DEBUGDEBUG_MODE_ENABLED:
print('#\tRemove ' + str(anOpt) + ' from defaults options at id ' + str(idx))
del STEPS[cstep].opt[idx:idx+2]
# create a commandline for the current step
cmdline = [STEPS[cstep].cmd] + STEPS[cstep].opt + options
print('CMD: ' + ' '.join(cmdline))
if not DEBUGDEBUG_MODE_ENABLED:
# Launch the current step
try:
pStep = subprocess.Popen(cmdline)
pStep.wait()
if pStep.returncode != 0:
break
except KeyboardInterrupt:
sys.exit('\r\nProcess canceled by user, all files remains')
else:
print('\t'.join(cmdline))
printout("# Pipeline end #", effect=INVERSE)

77
scripts/python/MvgOptimizeSfM.py
Unescape View File

@ -0,0 +1,77 @@ @@ -0,0 +1,77 @@
#!/usr/bin/python3
# -*- encoding: utf-8 -*-
"""
This script is for comparing the poses stored in an OpenMVS project to the poses optimized by OpenMVG
usage: run 'MvgOptimizeSfM.py' in a sub-folder to the OpenMVS project folder containing
'scene.mvs' and images stored in 'images' folder; structure ex:
-OpenMVS_project
-images
-scene.mvs
-mvg
-run script here
"""
import os
import sys
import subprocess
if sys.platform.startswith('win'):
PATH_DELIM = ';'
else:
PATH_DELIM = ':'
# add this script's directory to PATH
os.environ['PATH'] += PATH_DELIM + os.path.dirname(os.path.abspath(__file__))
# add current directory to PATH
os.environ['PATH'] += PATH_DELIM + os.getcwd()
def whereis(afile):
"""
return directory in which afile is, None if not found. Look in PATH
"""
if sys.platform.startswith('win'):
cmd = "where"
else:
cmd = "which"
try:
ret = subprocess.run([cmd, afile], stdout=subprocess.PIPE, stderr=subprocess.STDOUT, check=True)
return os.path.split(ret.stdout.decode())[0]
except subprocess.CalledProcessError:
return None
def launch(cmdline):
# Launch the current step
print('Cmd: ' + ' '.join(cmdline))
try:
pStep = subprocess.Popen(cmdline)
pStep.wait()
if pStep.returncode != 0:
return
except KeyboardInterrupt:
sys.exit('\r\nProcess canceled by user, all files remains')
# Try to find openMVG and openMVS binaries in PATH
OPENMVG_BIN = whereis("openMVG_main_SfMInit_ImageListing")
OPENMVS_BIN = whereis("ReconstructMesh")
# Ask user for openMVG and openMVS directories if not found
if not OPENMVG_BIN:
OPENMVG_BIN = input("openMVG binary folder?\n")
if not OPENMVS_BIN:
OPENMVS_BIN = input("openMVS binary folder?\n")
launch([os.path.join(OPENMVS_BIN, 'InterfaceCOLMAP'), '../scene.mvs', '-o', '../gt_dense_cameras.camera'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_SfMInit_ImageListingFromKnownPoses'), '-i', '../images', '-g', '../gt_dense_cameras', '-t', '1', '-o', '.'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_ComputeFeatures'), '-i', 'sfm_data.json', '-o', '.'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_ComputeMatches'), '-i', 'sfm_data.json', '-o', '.', '-m', '1'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_ComputeStructureFromKnownPoses'), '-i', 'sfm_data.json', '-m', '.', '-o', 'sfm_data_struct.bin', '-b'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_ComputeSfM_DataColor'), '-i', 'sfm_data_struct.bin', '-o', 'scene.ply'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_openMVG2openMVS'), '-i', 'sfm_data_struct.bin', '-o', 'scene.mvs', '-d', 'images'])
launch([os.path.join(OPENMVS_BIN, 'InterfaceCOLMAP'), 'scene.mvs', '-o', 'cameras.camera'])
launch([os.path.join(OPENMVG_BIN, 'openMVG_main_evalQuality'), '-i', '..', '-c', 'sfm_data_struct.bin', '-o', 'compare'])

42
scripts/python/MvsReadDMAP.py
Unescape View File

@ -0,0 +1,42 @@ @@ -0,0 +1,42 @@
'''
Example usage of MvsUtils.py for reading DMAP file content.
usage: MvsReadDMAP.py [-h] [--input INPUT] [--output OUTPUT]
'''
from argparse import ArgumentParser
from concurrent.futures import ProcessPoolExecutor
from glob import glob
from MvsUtils import loadDMAP
import numpy as np
import os
import pyvips
def exportDMAPContent(dmap_path):
dmap = loadDMAP(dmap_path)
basename = os.path.splitext(os.path.basename(dmap['file_name']))[0]
pyvips.Image.new_from_array(np.uint8(dmap['depth_map'] * (1 / dmap['depth_max']) * 255)).write_to_file('%s_depth_map.png' % basename)
if dmap['has_normal']:
pyvips.Image.new_from_array(np.uint8((dmap['normal_map'] @ -dmap['R'] + 1) * 0.5 * 255)).write_to_file('%s_normal_map.png' % basename)
if dmap['has_conf']:
pyvips.Image.new_from_array(np.uint8(dmap['confidence_map'] * (1 / dmap['confidence_map'].max()) * 255)).write_to_file('%s_confidence_map.png' % basename)
def main():
parser = ArgumentParser()
parser.add_argument('-i', '--input', type=str, required=True, help='Path to the DMAP file directory')
parser.add_argument('-t', '--threads', type=int, default=int(os.cpu_count() * 0.5) - 1, help='Number of parallel computations')
parser.add_argument('-o', '--output', type=str, required=True, help='Path to the output directory')
args = parser.parse_args()
dmap_paths = glob(os.path.join(args.input, '*.dmap'))
os.makedirs(args.output, exist_ok = True)
os.chdir(args.output)
with ProcessPoolExecutor(max_workers=args.threads) as executor:
executor.map(exportDMAPContent, dmap_paths)
if __name__ == '__main__':
main()

35
scripts/python/MvsReadMVS.py
Unescape View File

@ -0,0 +1,35 @@ @@ -0,0 +1,35 @@
'''
Example usage of MvsUtils.py for reading MVS interface archive content.
usage: MvsReadMVS.py [-h] [--input INPUT] [--output OUTPUT]
'''
from argparse import ArgumentParser
import json
from MvsUtils import loadMVSInterface
import os
def main():
parser = ArgumentParser()
parser.add_argument('-i', '--input', type=str, required=True, help='Path to the MVS interface archive')
parser.add_argument('-o', '--output', type=str, required=True, help='Path to the output json file')
args = parser.parse_args()
mvs = loadMVSInterface(args.input)
for platform_index in range(len(mvs['platforms'])):
for camera_index in range(len(mvs['platforms'][platform_index]['cameras'])):
camera = mvs['platforms'][platform_index]['cameras'][camera_index]
image_max = max(camera['width'], camera['height'])
fx = camera['K'][0][0] / image_max
fy = camera['K'][1][1] / image_max
poses_size = len(camera['poses'])
print('Camera model loaded: platform {}; camera {}; f {:.3f}x{:.3f}; poses {}'.format(platform_index, camera_index, fx, fy, poses_size))
os.makedirs(os.path.dirname(args.output), exist_ok = True)
with open(args.output, 'w') as file:
json.dump(mvs, file, indent=2)
if __name__ == '__main__':
main()

180
scripts/python/MvsScalablePipeline.py
Unescape View File

@ -0,0 +1,180 @@ @@ -0,0 +1,180 @@
#!/usr/bin/python3
# -*- encoding: utf-8 -*-
"""
This script helps with OpenMVS scalable pipeline.
Starting from a SfM solution stored into a MVS project accompanied by the undistorted images,
the fist step is to compute all depth maps without fusion:
DensifyPointCloud scene.mvs --fusion-mode 1
Next split the scene in sub-scenes using the area parameter, which is related to the inverse of GSD;
it is a bit non intuitive, but normally it should remain constant for a desired memory limit;
for example you can use the bellow value to limit memory usage to ~16GB:
DensifyPointCloud scene.mvs --sub-scene-area 660000
disable depth-maps re-filtering by creating a file Densify.ini with just this line:
Optimize = 0
and call fusion on each of the sub-scenes like:
DensifyPointCloud scene_0000.mvs --dense-config-file Densify.ini
............
DensifyPointCloud scene_000n.mvs --dense-config-file Densify.ini
This script helps to automate the process of calling DensifyPointCloud/ReconstructMesh on all sub-scenes.
usage: MvsScalablePipeline.py openMVS_module input_scene <options>
ex: MvsScalablePipeline.py DensifyPointCloud scene_XXXX.mvs --number-views-fuse 2
"""
import os
import subprocess
import sys
import argparse
import glob
DEBUG = False
if sys.platform.startswith('win'):
PATH_DELIM = ';'
FOLDER_DELIM = '\\'
else:
PATH_DELIM = ':'
FOLDER_DELIM = '/'
def whereis(afile):
"""
return directory in which afile is, None if not found. Look in PATH
"""
if sys.platform.startswith('win'):
cmd = "where"
else:
cmd = "which"
try:
ret = subprocess.run([cmd, afile], stdout=subprocess.PIPE, stderr=subprocess.STDOUT, check=True)
return os.path.split(ret.stdout.decode())[0]
except subprocess.CalledProcessError:
return None
def find(afile):
"""
As whereis look only for executable on linux, this find look for all file type
"""
for d in os.environ['PATH'].split(PATH_DELIM):
if os.path.isfile(os.path.join(d, afile)):
return d
return None
# Try to find openMVS binaries in PATH
OPENMVS_BIN = whereis("ReconstructMesh")
# Ask user for openMVS directory if not found
if not OPENMVS_BIN:
OPENMVS_BIN = input("openMVS binary folder?\n")
# HELPERS for terminal colors
BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE = range(8)
NO_EFFECT, BOLD, UNDERLINE, BLINK, INVERSE, HIDDEN = (0, 1, 4, 5, 7, 8)
# from Python cookbook, #475186
def has_colours(stream):
'''
Return stream colours capability
'''
if not hasattr(stream, "isatty"):
return False
if not stream.isatty():
return False # auto color only on TTYs
try:
import curses
curses.setupterm()
return curses.tigetnum("colors") > 2
except Exception:
# guess false in case of error
return False
HAS_COLOURS = has_colours(sys.stdout)
def printout(text, colour=WHITE, background=BLACK, effect=NO_EFFECT):
"""
print() with colour
"""
if HAS_COLOURS:
seq = "\x1b[%d;%d;%dm" % (effect, 30+colour, 40+background) + text + "\x1b[0m"
sys.stdout.write(seq+'\r\n')
else:
sys.stdout.write(text+'\r\n')
# store config and data in
class ConfContainer:
"""
Container for all the config variables
"""
def __init__(self):
pass
CONF = ConfContainer()
# ARGS
PARSER = argparse.ArgumentParser(
formatter_class=argparse.RawTextHelpFormatter,
description="Scalable MVS reconstruction with these steps: \r\n" +
"MvsScalablePipeline.py openMVS_module input_scene <options>\r\n"
)
PARSER.add_argument('openMVS_module',
help="the OpenMVS module to use: DensifyPointCloud, ReconstructMesh, etc.")
PARSER.add_argument('input_scene',
help="the scene name reg to process: scene_XXXX.mvs")
PARSER.add_argument('passthrough', nargs=argparse.REMAINDER, help="Option to be passed to command lines")
PARSER.parse_args(namespace=CONF) # store args in the ConfContainer
suffix = os.path.basename(CONF.input_scene).replace('scene_XXXX','')
CONF.input_scene = CONF.input_scene.replace('_dense','').replace('_mesh','').replace('_refine','').replace('_texture','')
# Absolute path for input directory
if len(CONF.input_scene) < 10 or CONF.input_scene[-9:] != '_XXXX.mvs':
sys.exit("%s: invalid scene name" % CONF.input_scene)
match CONF.openMVS_module:
case 'ReconstructMesh':
moduleSuffix = '_mesh.mvs'
case 'RefineMesh':
moduleSuffix = '_refine.mvs'
case 'TextureMesh':
moduleSuffix = '_texture.mvs'
case _:
moduleSuffix = '_dense.mvs'
printout("# Module {} start #".format(CONF.openMVS_module), colour=RED, effect=BOLD)
for scene_name in glob.glob(os.path.abspath(os.path.join(os.path.dirname(CONF.input_scene), 'scene_[0-9][0-9][0-9][0-9]'+suffix))):
if os.path.exists(os.path.splitext(scene_name)[0] + moduleSuffix) == False:
printout("# Process: %s" % os.path.basename(scene_name), colour=GREEN, effect=NO_EFFECT)
# create a commandline for the current step
cmdline = [os.path.join(OPENMVS_BIN, CONF.openMVS_module), scene_name] + CONF.passthrough
print('Cmd: ' + ' '.join(cmdline))
if not DEBUG:
# Launch the current step
try:
pStep = subprocess.Popen(cmdline)
pStep.wait()
if pStep.returncode != 0:
printout("# Warning: step failed", colour=RED, effect=BOLD)
except KeyboardInterrupt:
sys.exit('\r\nProcess canceled by user, all files remains')
else:
print('\t'.join(cmdline))
printout("# Module {} end #".format(CONF.openMVS_module), colour=RED, effect=BOLD)

190
scripts/python/MvsUtils.py
Unescape View File

@ -0,0 +1,190 @@ @@ -0,0 +1,190 @@
'''
OpenMVS python utilities.
E.g., from MvsUtils import loadDMAP, loadMVSInterface
'''
import numpy as np
def loadDMAP(dmap_path):
with open(dmap_path, 'rb') as dmap:
file_type = dmap.read(2).decode()
content_type = np.frombuffer(dmap.read(1), dtype=np.dtype('B'))
reserve = np.frombuffer(dmap.read(1), dtype=np.dtype('B'))
has_depth = content_type > 0
has_normal = content_type in [3, 7, 11, 15]
has_conf = content_type in [5, 7, 13, 15]
has_views = content_type in [9, 11, 13, 15]
image_width, image_height = np.frombuffer(dmap.read(8), dtype=np.dtype('I'))
depth_width, depth_height = np.frombuffer(dmap.read(8), dtype=np.dtype('I'))
if (file_type != 'DR' or has_depth == False or depth_width <= 0 or depth_height <= 0 or image_width < depth_width or image_height < depth_height):
print('error: opening file \'{}\' for reading depth-data'.format(dmap_path))
return
depth_min, depth_max = np.frombuffer(dmap.read(8), dtype=np.dtype('f'))
file_name_size = np.frombuffer(dmap.read(2), dtype=np.dtype('H'))[0]
file_name = dmap.read(file_name_size).decode()
view_ids_size = np.frombuffer(dmap.read(4), dtype=np.dtype('I'))[0]
reference_view_id, *neighbor_view_ids = np.frombuffer(dmap.read(4 * view_ids_size), dtype=np.dtype('I'))
K = np.frombuffer(dmap.read(72), dtype=np.dtype('d')).reshape(3, 3)
R = np.frombuffer(dmap.read(72), dtype=np.dtype('d')).reshape(3, 3)
C = np.frombuffer(dmap.read(24), dtype=np.dtype('d'))
data = {
'has_normal': has_normal,
'has_conf': has_conf,
'has_views': has_views,
'image_width': image_width,
'image_height': image_height,
'depth_width': depth_width,
'depth_height': depth_height,
'depth_min': depth_min,
'depth_max': depth_max,
'file_name': file_name,
'reference_view_id': reference_view_id,
'neighbor_view_ids': neighbor_view_ids,
'K': K,
'R': R,
'C': C
}
map_size = depth_width * depth_height
depth_map = np.frombuffer(dmap.read(4 * map_size), dtype=np.dtype('f')).reshape(depth_height, depth_width)
data.update({'depth_map': depth_map})
if has_normal:
normal_map = np.frombuffer(dmap.read(4 * map_size * 3), dtype=np.dtype('f')).reshape(depth_height, depth_width, 3)
data.update({'normal_map': normal_map})
if has_conf:
confidence_map = np.frombuffer(dmap.read(4 * map_size), dtype=np.dtype('f')).reshape(depth_height, depth_width)
data.update({'confidence_map': confidence_map})
if has_views:
views_map = np.frombuffer(dmap.read(map_size * 4), dtype=np.dtype('B')).reshape(depth_height, depth_width, 4)
data.update({'views_map': views_map})
return data
def loadMVSInterface(archive_path):
with open(archive_path, 'rb') as mvs:
archive_type = mvs.read(4).decode()
version = np.frombuffer(mvs.read(4), dtype=np.dtype('I')).tolist()[0]
reserve = np.frombuffer(mvs.read(4), dtype=np.dtype('I'))
if archive_type != 'MVSI':
print('error: opening file \'{}\''.format(archive_path))
return
data = {
'project_stream': archive_type,
'project_stream_version': version,
'platforms': [],
'images': [],
'vertices': [],
'vertices_normal': [],
'vertices_color': []
}
platforms_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for platform_index in range(platforms_size):
platform_name_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
platform_name = mvs.read(platform_name_size).decode()
data['platforms'].append({'name': platform_name, 'cameras': []})
cameras_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for camera_index in range(cameras_size):
camera_name_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
camera_name = mvs.read(camera_name_size).decode()
data['platforms'][platform_index]['cameras'].append({'name': camera_name})
if version > 3:
band_name_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
band_name = mvs.read(band_name_size).decode()
data['platforms'][platform_index]['cameras'][camera_index].update({'band_name': band_name})
if version > 0:
width, height = np.frombuffer(mvs.read(8), dtype=np.dtype('I')).tolist()
data['platforms'][platform_index]['cameras'][camera_index].update({'width': width, 'height': height})
K = np.asarray(np.frombuffer(mvs.read(72), dtype=np.dtype('d'))).reshape(3, 3).tolist()
data['platforms'][platform_index]['cameras'][camera_index].update({'K': K, 'poses': []})
identity_matrix = np.asarray(np.frombuffer(mvs.read(96), dtype=np.dtype('d'))).reshape(4, 3)
poses_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(poses_size):
R = np.asarray(np.frombuffer(mvs.read(72), dtype=np.dtype('d'))).reshape(3, 3).tolist()
C = np.asarray(np.frombuffer(mvs.read(24), dtype=np.dtype('d'))).tolist()
data['platforms'][platform_index]['cameras'][camera_index]['poses'].append({'R': R, 'C': C})
images_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for image_index in range(images_size):
name_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
name = mvs.read(name_size).decode()
data['images'].append({'name': name})
if version > 4:
mask_name_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
mask_name = mvs.read(mask_name_size).decode()
data['images'][image_index].update({'mask_name': mask_name})
platform_id, camera_id, pose_id = np.frombuffer(mvs.read(12), dtype=np.dtype('I')).tolist()
data['images'][image_index].update({'platform_id': platform_id, 'camera_id': camera_id, 'pose_id': pose_id})
if version > 2:
id = np.frombuffer(mvs.read(4), dtype=np.dtype('I')).tolist()[0]
data['images'][image_index].update({'id': id})
if version > 6:
min_depth, avg_depth, max_depth = np.frombuffer(mvs.read(12), dtype=np.dtype('f')).tolist()
data['images'][image_index].update({'min_depth': min_depth, 'avg_depth': avg_depth, 'max_depth': max_depth, 'view_scores': []})
view_score_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(view_score_size):
id, points = np.frombuffer(mvs.read(8), dtype=np.dtype('I')).tolist()
scale, angle, area, score = np.frombuffer(mvs.read(16), dtype=np.dtype('f')).tolist()
data['images'][image_index]['view_scores'].append({'id': id, 'points': points, 'scale': scale, 'angle': angle, 'area': area, 'score': score})
vertices_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for vertex_index in range(vertices_size):
X = np.frombuffer(mvs.read(12), dtype=np.dtype('f')).tolist()
data['vertices'].append({'X': X, 'views': []})
views_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(views_size):
image_id = np.frombuffer(mvs.read(4), dtype=np.dtype('I')).tolist()[0]
confidence = np.frombuffer(mvs.read(4), dtype=np.dtype('f')).tolist()[0]
data['vertices'][vertex_index]['views'].append({'image_id': image_id, 'confidence': confidence})
vertices_normal_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(vertices_normal_size):
normal = np.frombuffer(mvs.read(12), dtype=np.dtype('f')).tolist()
data['vertices_normal'].append(normal)
vertices_color_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(vertices_color_size):
color = np.frombuffer(mvs.read(3), dtype=np.dtype('B')).tolist()
data['vertices_color'].append(color)
if version > 0:
data.update({'lines': [], 'lines_normal': [], 'lines_color': []})
lines_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for line_index in range(lines_size):
pt1 = np.frombuffer(mvs.read(12), dtype=np.dtype('f')).tolist()
pt2 = np.frombuffer(mvs.read(12), dtype=np.dtype('f')).tolist()
data['lines'].append({'pt1': pt1, 'pt2': pt2, 'views': []})
views_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(views_size):
image_id = np.frombuffer(mvs.read(4), dtype=np.dtype('I')).tolist()[0]
confidence = np.frombuffer(mvs.read(4), dtype=np.dtype('f')).tolist()[0]
data['lines'][line_index]['views'].append({'image_id': image_id, 'confidence': confidence})
lines_normal_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(lines_normal_size):
normal = np.frombuffer(mvs.read(12), dtype=np.dtype('f')).tolist()
data['lines_normal'].append(normal)
lines_color_size = np.frombuffer(mvs.read(8), dtype=np.dtype('Q'))[0]
for _ in range(lines_color_size):
color = np.frombuffer(mvs.read(3), dtype=np.dtype('B')).tolist()
data['lines_color'].append(color)
if version > 1:
transform = np.frombuffer(mvs.read(128), dtype=np.dtype('d')).reshape(4, 4).tolist()
data.update({'transform': transform})
if version > 5:
rot = np.frombuffer(mvs.read(72), dtype=np.dtype('d')).reshape(3, 3).tolist()
pt_min = np.frombuffer(mvs.read(24), dtype=np.dtype('d')).tolist()
pt_max = np.frombuffer(mvs.read(24), dtype=np.dtype('d')).tolist()
data.update({'obb': {'rot': rot, 'pt_min': pt_min, 'pt_max': pt_max}})
return data

126
scripts/python/blender_decimate.py
Unescape View File

@ -0,0 +1,126 @@ @@ -0,0 +1,126 @@
import bpy, sys, time, math
def clear_default_objects():
for name in ["Cube", "Light", "Camera"]:
if obj := bpy.data.objects.get(name):
bpy.data.objects.remove(obj)
def mark_sharp_edges(obj, angle=30):
bpy.context.view_layer.objects.active = obj
bpy.ops.object.mode_set(mode='EDIT')
bpy.ops.mesh.select_all(action='SELECT')
# 标记30°以上的边为锐边
bpy.ops.mesh.edges_select_sharp(sharpness=math.radians(angle))
bpy.ops.mesh.mark_sharp()
bpy.ops.object.mode_set(mode='OBJECT')
def planar_decimate(obj, angle_limit=35.0, ratio=0.5, protect_uv=True):
"""使用Decimate修改器+智能参数配置"""
print(f"启动高级减面流程 | 模式: 平面优化+权重控制")
#"""
# 添加Decimate修改器（平面模式）
mod = obj.modifiers.new(name="PlanarDecimate", type='DECIMATE')
mod.decimate_type = 'DISSOLVE' # 平面溶解模式
mod.angle_limit = math.radians(angle_limit) # 角度转弧度
mod.use_dissolve_boundaries = False # 溶解边界防止破面[5](@ref)
#"""
"""
mod = obj.modifiers.new(name="CollapseDecimate", type='DECIMATE')
mod.decimate_type = 'COLLAPSE'
mod.ratio = 0.1 # 保留30%面数
mod.use_symmetry = True
mod.vertex_group = "high_detail" # 绑定顶点组保护细节区域
#"""
# 保护关键属性
mod.delimit = {'NORMAL', 'UV'} if protect_uv else {'NORMAL'} # UV保护[1](@ref)
mod.use_symmetry = True # 对称保护[1](@ref)
# 保护边界：创建顶点组标记轮廓边
if "boundary_edges" not in obj.vertex_groups:
obj.vertex_groups.new(name="boundary_edges")
mod.vertex_group = "boundary_edges"
mod.invert_vertex_group = True # 仅减非边界区域
# 应用修改器
bpy.context.view_layer.objects.active = obj
bpy.ops.object.modifier_apply(modifier=mod.name)
print(f"平面优化完成 | 容差角: {angle_limit}°")
def collapse_decimate(obj, ratio=0.3, protect_groups=True):
"""核心减面：Collapse模式+细节保护"""
mod = obj.modifiers.new(name="CollapseDecimate", type='DECIMATE')
mod.decimate_type = 'COLLAPSE'
mod.ratio = ratio
# 顶点组保护（需提前在Blender中标记高细节区域）[1](@ref)
if protect_groups and "high_detail" in obj.vertex_groups:
mod.vertex_group = "high_detail"
mod.invert_vertex_group = True # 仅减非保护区域
print("启用顶点组保护 | 保留高细节区域")
# 关键保护设置
mod.use_symmetry = True
bpy.ops.object.modifier_apply(modifier=mod.name)
print(f"塌陷减面完成 | 保留比例: {ratio*100}%")
def triangulate_mesh(obj, quad_method='BEAUTY', ngon_method='BEAUTY'):
"""将模型三角化[1,2](@ref)"""
print("启动三角化处理...")
# 添加三角化修改器
mod = obj.modifiers.new(name="Triangulate", type='TRIANGULATE')
mod.quad_method = quad_method # 四边形处理方法
mod.ngon_method = ngon_method # 多边形处理方法
mod.min_vertices = 4 # 对四边形及以上面进行三角化
# 应用修改器
bpy.context.view_layer.objects.active = obj
bpy.ops.object.modifier_apply(modifier=mod.name)
# 验证三角化结果
tris = sum(1 for poly in obj.data.polygons if len(poly.vertices) == 3)
total_faces = len(obj.data.polygons)
print(f"三角化完成 | 三角面比例: {tris/total_faces*100:.1f}% ({tris}/{total_faces})")
def optimize_geometry(obj):
"""后处理：清理拓扑+法线修复[1,6](@ref)"""
bpy.context.view_layer.objects.active = obj
bpy.ops.object.mode_set(mode='EDIT')
bpy.ops.mesh.remove_doubles(threshold=0.001) # 合并重复顶点
bpy.ops.mesh.normals_make_consistent() # 统一法线
bpy.ops.object.mode_set(mode='OBJECT')
print("拓扑清理完成 | 法线已重计算")
if __name__ == "__main__":
argv = sys.argv[sys.argv.index("--") + 1:]
input_path = argv[0]
angle_limit = float(argv[1]) if len(argv) > 1 else 5.0
output_path = argv[2] if len(argv) > 2 else input_path.replace(".ply", "_decimated.ply")
# 硬件加速配置
# bpy.context.scene.cycles.device = 'GPU'
# bpy.context.preferences.system.memory_cache_limit = 4096
clear_default_objects()
# 导入PLY模型
bpy.ops.wm.ply_import(filepath=input_path)
obj = bpy.context.selected_objects[0]
print(f"开始减面操作：{obj.name}")
# 执行减面
mark_sharp_edges(obj)
planar_decimate(obj, angle_limit) # 先平面优化
collapse_decimate(obj, ratio=0.2) # 再智能塌陷 // 0.1-3效果好，0.2-3综合不错
optimize_geometry(obj) # 最终清理
triangulate_mesh(obj)
print("减面完成")
bpy.ops.wm.ply_export(
filepath=output_path,
apply_modifiers=False
)
print(f"模型已导出至：{output_path}")

10
scripts/python/run.sh
Unescape View File

@ -0,0 +1,10 @@ @@ -0,0 +1,10 @@
cd ~/code/openMVS/make
cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH=/lib --build ..
make -j20 && make install
#cmake --build . -j4
#cmake --install .
cd ~/code/openMVS/scripts/python
#InterfaceCOLMAP -i /data/datasets/scan1/output/sfm/dense -o scene.mvs --image-folder /data/datasets/scan1/output/sfm/dense/images -w "/data/datasets/scan1/output/mvs"
#TextureMesh scene_dense.mvs -m scene_dense_mesh.ply -o scene_dense_mesh_texture.mvs --decimate 0.5 -w "/data/datasets/scan1/output2"
# TextureMesh /data/datasets/66536 --image-folder /data/datasets/66536/images/ -m /data/datasets/66536/gsmodel/train/ours_15000/waterobj/final.ply -o mesh.mvs --decimate 0.1 -w "/data/datasets/66536" --virtual-face-images 3
# TextureMesh /data/datasets/work_dir/227124 --image-folder /data/datasets/work_dir/227124/images/ -m /data/datasets/work_dir/227124/gsmodel/train/ours_16000/waterobj/final.ply -o mesh.mvs --outlier-threshold 0.1 --decimate 0.05 --sharpness-weight 0.2 -w /data/datasets/work_dir/227124/gsmodel/train/ours_16000/mesh/ --virtual-face-images 3 --patch-packing-heuristic 0 --max-texture-size 0
Write Preview
Loading…
Cancel
Save