colmap重建 指定内参和自动去畸变

 

run.sh

参数说明

1指定内参

--ImageReader.camera_params "1451.7946523730436,1438.2609968095967,960,540,0,0,0,0" \
内参数据

Camera.fx: 355.0974745605948
Camera.fy: 355.47832693317105
Camera.cx: 357.7074039567714
Camera.cy: 351.0244037313849
 
Camera.k1: -0.023790306606729556
Camera.k2: -0.0007571494794293715
Camera.p1: 0.00016452517056601848
Camera.p2: -0.0005743824914513448
Camera.k3: 0
Camera.k4: 0

　　2colmap重建代码

calibDataRoot 文件根目录

imagePath 图像地址

ImageReader.camera_model PINHOLE 相机模型选择
相机内参

无畸变模型

ImageReader.camera_params "1329.72479114946,1329.65375225699,961.340052491983,542.650028093206" \

--ImageReader.camera_model PINHOLE  
--ImageReader.camera_params "fx, fy, cx, cy"

有畸变模型
k1, k2, p1, p2, k3：径向和切向畸变系数。

--ImageReader.camera_model OPENCV
--ImageReader.camera_params "fx, fy, cx, cy, k1, k2, p1, p2, k3"

　　脚本指令参考

脚本1 给定内参 真孔相机

#!/bin/bash
calibDataRoot="/home/dongdong/2project/0data/house4/70/colmap/"
imagePath="/home/dongdong/2project/0data/house4/70/colmap/images/"

mkdir -p ${calibDataRoot}/input
colmap feature_extractor --database_path $calibDataRoot/input/database.db \
                                --image_path ${imagePath} \
                                --ImageReader.single_camera true \
                                --ImageReader.camera_model PINHOLE  \
                                --ImageReader.camera_params "1329.72479114946,1329.65375225699,961.340052491983,542.650028093206" \
                                --SiftExtraction.estimate_affine_shape true \
                                --SiftExtraction.domain_size_pooling true \
                                --SiftExtraction.num_threads -1 \
                                --SiftExtraction.gpu_index 0

#sequential_matcher
    colmap  exhaustive_matcher --database_path $calibDataRoot/input/database.db \
                            --SiftMatching.guided_matching true \
                            --SiftMatching.num_threads -1 --SiftMatching.gpu_index 0
    # mkdir -p ${calibDataRoot}/input
    colmap mapper --database_path $calibDataRoot/input/database.db \
                        --image_path ${imagePath} \
                        --output_path $calibDataRoot/input \
                        --Mapper.ba_local_max_refinements 3 \
                        --Mapper.ba_local_max_num_iterations 100 \
                        --Mapper.max_extra_param 99999 \
                        --Mapper.ba_refine_principal_point 0 \
                        --Mapper.ba_refine_focal_length 0 \
                        --Mapper.ba_refine_extra_params 0 \
                        --Mapper.tri_ignore_two_view_tracks false \
                        --Mapper.num_threads -1

    
    colmap model_converter --input_path ${calibDataRoot}/input/0 \
                        --output_type txt --output_path ${calibDataRoot}/input/0

 

畸变矫正参考

### Image undistortion
## We need to undistort our images into ideal pinhole intrinsics.
img_undist_cmd = (colmap_command + " image_undistorter \
    --image_path " + args.source_path + "/input \
    --input_path " + args.source_path + "/distorted/sparse/0 \
    --output_path " + args.source_path + "\
    --output_type COLMAP")
exit_code = os.system(img_undist_cmd)
if exit_code != 0:
    logging.error(f"Mapper failed with code {exit_code}. Exiting.")
    exit(exit_code)

files = os.listdir(args.source_path + "/sparse")
os.makedirs(args.source_path + "/sparse/0", exist_ok=True)
# Copy each file from the source directory to the destination directory
for file in files:
    if file == '0':
        continue
    source_file = os.path.join(args.source_path, "sparse", file)
    destination_file = os.path.join(args.source_path, "sparse", "0", file)
    shutil.move(source_file, destination_file)

if(args.resize):
    print("Copying and resizing...")

　　

　　

python代码

程序1-1 单纯畸变矫正图像重新保存 图像

import os
import cv2
import yaml
import glob

def load_camera_parameters(yaml_file):
    with open(yaml_file, 'r') as file:
        params = yaml.safe_load(file)
    return params

def undistort_images(input_folder, output_folder, camera_params):
    # 获取所有图像文件，并按时间戳排序
    image_files = sorted(glob.glob(os.path.join(input_folder, '*.jpg')))
    
    # 提取相机内参
    fx = camera_params['Camera.fx']
    fy = camera_params['Camera.fy']
    cx = camera_params['Camera.cx']
    cy = camera_params['Camera.cy']

    # 提取畸变系数
    k1 = camera_params['Camera.k1']
    k2 = camera_params['Camera.k2']
    p1 = camera_params['Camera.p1']
    p2 = camera_params['Camera.p2']
    k3 = camera_params['Camera.k3']
    k4 = camera_params['Camera.k4']

    # 构造相机内参矩阵和畸变系数数组
    camera_matrix = np.array([[fx, 0, cx],
                               [0, fy, cy],
                               [0, 0, 1]], dtype=np.float64)

    dist_coeffs = np.array([k1, k2, p1, p2, k3, k4], dtype=np.float64)

    for image_file in image_files:
        # 读取图像
        image = cv2.imread(image_file)

        # 畸变校正
        undistorted_image = cv2.undistort(image, camera_matrix, dist_coeffs)

        # 构造输出文件名
        output_file = os.path.join(output_folder, os.path.basename(image_file))

        # 保存校正后的图像
        cv2.imwrite(output_file, undistorted_image)
        print(f"Saved undistorted image: {output_file}")

if __name__ == '__main__':
    input_folder = 'path/to/input_folder'  # 输入文件夹路径
    output_folder = 'path/to/output_folder'  # 输出文件夹路径
    yaml_file = 'path/to/camera_parameters.yaml'  # YAML 文件路径

    # 加载相机参数
    camera_params = load_camera_parameters(yaml_file)

    # 创建输出文件夹（如果不存在）
    os.makedirs(output_folder, exist_ok=True)

    # 进行畸变校正
    undistort_images(input_folder, output_folder, camera_params)

程序1-2 单纯畸变矫正参考

import os
import cv2
import yaml
import glob
import numpy as np
 
def load_camera_parameters(yaml_file):
    with open(yaml_file, 'r') as file:
        params = yaml.safe_load(file)
    return params
 
def undistort_images(input_folder, output_folder, camera_params,img_type='png'):
    # 获取所有图像文件，并按时间戳排序
    image_files = sorted(glob.glob(os.path.join(input_folder, '*.'+img_type)))
     
    # 提取相机内参
    fx = camera_params['Camera.fx']
    fy = camera_params['Camera.fy']
    cx = camera_params['Camera.cx']
    cy = camera_params['Camera.cy']
 
    # 提取畸变系数
    k1 = camera_params['Camera.k1']
    k2 = camera_params['Camera.k2']
    p1 = camera_params['Camera.p1']
    p2 = camera_params['Camera.p2']
    k3 = camera_params['Camera.k3']
    k4 = camera_params['Camera.k4']
    #print("fx",fx,'fy',fy,'cx',cx,'cy',cy,'k1',k1,'k2',k2,'p1',p1,'p2',p2)
 
    # 构造相机内参矩阵和畸变系数数组
    camera_matrix = np.array([[fx, 0, cx],
                               [0, fy, cy],
                               [0, 0, 1]], dtype=np.float64)
 
    #dist_coeffs = np.array([k1, k2, p1, p2, k3, k4], dtype=np.float64)
    dist_coeffs = np.array([k1, k2, p1, p2,k3], dtype=np.float64)
 
    for image_file in image_files:
        # 读取图像
        image = cv2.imread(image_file)
 
        # 畸变校正
        #undistorted_image = cv2.undistort(image, camera_matrix, dist_coeffs)
        h, w = image.shape[:2]
        mapx, mapy = cv2.initUndistortRectifyMap(camera_matrix, dist_coeffs, None, camera_matrix, (w, h), 5)
        undistorted_image = cv2.remap(image, mapx, mapy, cv2.INTER_LINEAR)
        # 构造输出文件名
        output_file = os.path.join(output_folder, os.path.basename(image_file))
 
        # 保存校正后的图像
        cv2.imwrite(output_file, undistorted_image)
        print(f"Saved undistorted image: {output_file}")
 
if __name__ == '__main__':
    data_path = "/home/dongdong/2project/0data/ehang/ehang2_test/"
    input_folder = data_path+'colmap/in_images/'  # 输入文件夹路径
    output_folder = data_path+'colmap/images/'   # 输出文件夹路径
    yaml_file =  data_path+'FHY_config.yaml'  # YAML 文件路径
 
    # 加载相机参数
    camera_params = load_camera_parameters(yaml_file)
 
    # 创建输出文件夹（如果不存在）
    os.makedirs(output_folder, exist_ok=True)
 
    # 进行畸变校正
    undistort_images(input_folder, output_folder, camera_params)

　　

　　

程序2 不指定内参 自动估计重建且畸变矫正图像重新输出文件夹和结果

使用

'''
--source_path 输入文件夹  下面有一个input文件夹 存放原始没有畸变矫正的图片
--camera 相机模型  PINHOLE  OPENCV
--output_path 输出文件夹 images 输出畸变矫正后的图像 spase输出点云结果

python convert.py \
--source_path /home/dongdong/2project/0data/ehang/ehang2_test/ehang2 \
--camera OPENCV \
--output_path /home/dongdong/2project/0data/ehang/ehang2_test/ehang2_output

没有指定 --no_gpu 和 --skip_matching，所以默认会调用 GPU 并进行匹配。--resize 也没有被设置，因此默认值为 False。

'''

　input输入未校准的图像位置

 

　第一次未校准的结果存放在distorted

 输出畸变矫正后结果

 

具体命令

#
# Copyright (C) 2023, Inria
# GRAPHDECO research group, https://team.inria.fr/graphdeco
# All rights reserved.
#
# This software is free for non-commercial, research and evaluation use
# under the terms of the LICENSE.md file.
#
# For inquiries contact  george.drettakis@inria.fr
#
import time
import os
import logging
from argparse import ArgumentParser
import shutil

'''
python convert.py -s /home/dongdong/2project/0data/ehang/ehang2_test/ehang2 --camera PINHOLE

python convert.py -s /home/dongdong/2project/0data/ehang/ehang2_test/npu --camera OPENCV
python convert.py \
--source_path /home/dongdong/2project/0data/ehang/ehang2_test/ehang2 \
--camera OPENCV \
--output_path /home/dongdong/2project/0data/ehang/ehang2_test/ehang2_output

没有指定 --no_gpu 和 --skip_matching，所以默认会调用 GPU 并进行匹配。--resize 也没有被设置，因此默认值为 False。

'''

# This Python script is based on the shell converter script provided in the MipNerF 360 repository.
parser = ArgumentParser("Colmap converter")
parser.add_argument("--no_gpu", action='store_true')
parser.add_argument("--skip_matching", action='store_true')
parser.add_argument("--source_path", "-s", required=True, type=str)
parser.add_argument("--output_path", "-o", required=True, type=str)
parser.add_argument("--camera", default="OPENCV", type=str)
parser.add_argument("--colmap_executable", default="", type=str)
parser.add_argument("--resize", action="store_true")
parser.add_argument("--magick_executable", default="", type=str)
args = parser.parse_args()
colmap_command = '"{}"'.format(args.colmap_executable) if len(args.colmap_executable) > 0 else "colmap"
magick_command = '"{}"'.format(args.magick_executable) if len(args.magick_executable) > 0 else "magick"
use_gpu = 1 if not args.no_gpu else 0  # 默认使用gpu


# 默认不跳过skip_matching
if not args.skip_matching:
    os.makedirs(args.source_path + "/distorted/sparse", exist_ok=True)

    ## Feature extraction 特征提取 ===================
    feat_extracton_cmd = colmap_command + " feature_extractor "\
        "--database_path " + args.source_path + "/distorted/database.db \
        --image_path " + args.source_path + "/input \
        --ImageReader.single_camera 1 \
        --ImageReader.camera_model " + args.camera + " \
        --SiftExtraction.use_gpu " + str(use_gpu) 
    exit_code = os.system(feat_extracton_cmd)
    if exit_code != 0:
        logging.error(f"Feature extraction failed with code {exit_code}. Exiting.")
        exit(exit_code)

    ## Feature matching 特征匹配 ===================
    feat_matching_cmd = colmap_command + " exhaustive_matcher \
        --database_path " + args.source_path + "/distorted/database.db \
        --SiftMatching.use_gpu " + str(use_gpu) +" \
        --SiftMatching.guided_matching true"
    exit_code = os.system(feat_matching_cmd)
    if exit_code != 0:
        logging.error(f"Feature matching failed with code {exit_code}. Exiting.")
        exit(exit_code)

    ### Bundle adjustment 优化重建 ==================
    # The default Mapper tolerance is unnecessarily large,
    # decreasing it speeds up bundle adjustment steps.
    mapper_cmd = (colmap_command + " mapper \
        --database_path " + args.source_path + "/distorted/database.db \
        --image_path "  + args.source_path + "/input \
        --output_path "  + args.source_path + "/distorted/sparse \
        --Mapper.ba_global_function_tolerance=0.000001")
    exit_code = os.system(mapper_cmd)
    if exit_code != 0:
        logging.error(f"Mapper failed with code {exit_code}. Exiting.")
        exit(exit_code)

    # 输出转换为txt
    mapper_cmd = (colmap_command + " model_converter \
        --input_path " + args.source_path + "/distorted/sparse/0 \
        --output_type "  +  "txt \
        --output_path "  + args.source_path + "/distorted/sparse/0"
        )
    exit_code = os.system(mapper_cmd)
    if exit_code != 0:
        logging.error(f"bin转换txt失败{exit_code}. Exiting.")
        exit(exit_code)

### Image undistortion
## We need to undistort our images into ideal pinhole intrinsics.
'''
--image_path:
输入图像的路径。可以是单个图像文件或包含多个图像的目录。
--output_path: 默认images
去畸变后的图像输出路径。可以是一个目录，COLMAP 将在该目录下生成去畸变后的图像。
--input_path:
输入相机参数和姿态文件的路径，通常是 COLMAP 数据库或导出的模型文件。
--output_type:
指定输出图像的格式，如 PNG、JPEG 等。
--max_image_size:
指定输出图像的最大尺寸。COLMAP 会根据原图像的比例缩放图像。
--remove_uncovered:
选择是否移除未被覆盖的区域，这在某些情况下可以用来减少输出图像的数量。
--image_list_path:
包含要去畸变的图像列表的文件路径。
'''
print("===============准备开始矫正================")
time.sleep(5)



img_undist_cmd = (colmap_command + " image_undistorter \
    --image_path " + args.source_path + "/input \
    --input_path " + args.source_path + "/distorted/sparse/0 \
    --output_path " + args.output_path + "\
    --output_type COLMAP")
exit_code = os.system(img_undist_cmd)
if exit_code != 0:
    logging.error(f"Mapper failed with code {exit_code}. Exiting.")
    exit(exit_code)

# 创建sparse文件夹
files = os.listdir(args.output_path + "/sparse")
os.makedirs(args.output_path + "/sparse/0", exist_ok=True)
# Copy each file from the source directory to the destination directory
for file in files:
    if file == '0':
        continue
    source_file = os.path.join(args.output_path, "sparse", file)
    destination_file = os.path.join(args.output_path, "sparse", "0", file)
    shutil.move(source_file, destination_file)

# 输出转换为txt
mapper_cmd = (colmap_command + " model_converter \
    --input_path " + args.output_path + "/sparse/0 \
    --output_type "  +  "txt \
    --output_path "  + args.output_path + "/sparse/0"
    )
exit_code = os.system(mapper_cmd)
if exit_code != 0:
    logging.error(f"bin转换txt失败{exit_code}. Exiting.")
    exit(exit_code)


if(args.resize):
    print("Copying and resizing...")

    # Resize images.
    os.makedirs(args.source_path + "/images_2", exist_ok=True)
    os.makedirs(args.source_path + "/images_4", exist_ok=True)
    os.makedirs(args.source_path + "/images_8", exist_ok=True)
    # Get the list of files in the source directory
    files = os.listdir(args.source_path + "/images")
    # Copy each file from the source directory to the destination directory
    for file in files:
        source_file = os.path.join(args.source_path, "images", file)

        destination_file = os.path.join(args.source_path, "images_2", file)
        shutil.copy2(source_file, destination_file)
        exit_code = os.system(magick_command + " mogrify -resize 50% " + destination_file)
        if exit_code != 0:
            logging.error(f"50% resize failed with code {exit_code}. Exiting.")
            exit(exit_code)

        destination_file = os.path.join(args.source_path, "images_4", file)
        shutil.copy2(source_file, destination_file)
        exit_code = os.system(magick_command + " mogrify -resize 25% " + destination_file)
        if exit_code != 0:
            logging.error(f"25% resize failed with code {exit_code}. Exiting.")
            exit(exit_code)

        destination_file = os.path.join(args.source_path, "images_8", file)
        shutil.copy2(source_file, destination_file)
        exit_code = os.system(magick_command + " mogrify -resize 12.5% " + destination_file)
        if exit_code != 0:
            logging.error(f"12.5% resize failed with code {exit_code}. Exiting.")
            exit(exit_code)

print("Done.")

　　

 

程序3 指定内参 重建且畸变矫正图像重新输出文件夹和结果

 

 

 

 

 

# 教程地址  https://blog.csdn.net/a40850273/article/details/127976111
#
import time
import os
import logging
from argparse import ArgumentParser
import shutil
import yaml
import numpy as np


def load_camera_parameters(yaml_file):
    with open(yaml_file, 'r') as file:
        params = yaml.safe_load(file)
    return params

'''
输入参数：
--source_path 输入文件夹  下面有一个input文件夹 存放原始没有畸变矫正的图片
--camera 相机模型  PINHOLE  OPENCV
--output_path 输出文件夹 images 输出畸变矫正后的图像 spase输出点云结果
--yaml_file 相机内参
格式 
    Camera.fx: 355.0974745605948
    Camera.fy: 355.47832693317105
    Camera.cx: 357.7074039567714
    Camera.cy: 351.0244037313849
    
    Camera.k1: -0.023790306606729556
    Camera.k2: -0.0007571494794293715
    Camera.p1: 0.00016452517056601848
    Camera.p2: -0.0005743824914513448
    Camera.k3: 0
    Camera.k4: 0

执行命令：
python convert_yaml.py \
--source_path /media/dongdong/Elements1/项目/fjx/2TestData/2TestData/ehang2 \
--output_path /home/dongdong/2project/0data/ehang/ehang2/ehang2_output_yaml \
--camera OPENCV \
--yaml_file /home/dongdong/2project/0data/ehang/ehang2/FHY_config.yaml

没有指定 --no_gpu 和 --skip_matching，所以默认会调用 GPU 并进行匹配。--resize 也没有被设置，因此默认值为 False。

'''

# This Python script is based on the shell converter script provided in the MipNerF 360 repository.
parser = ArgumentParser("Colmap converter")
parser.add_argument("--no_gpu", action='store_true')
parser.add_argument("--skip_matching", action='store_true')
parser.add_argument("--source_path", "-s", required=True, type=str)
parser.add_argument("--output_path", "-o", required=True, type=str)
parser.add_argument("--yaml_file", "-y", required=True, type=str)
parser.add_argument("--camera", default="OPENCV", type=str)
parser.add_argument("--colmap_executable", default="", type=str)
parser.add_argument("--resize", action="store_true")
parser.add_argument("--magick_executable", default="", type=str)
args = parser.parse_args()
colmap_command = '"{}"'.format(args.colmap_executable) if len(args.colmap_executable) > 0 else "colmap"
magick_command = '"{}"'.format(args.magick_executable) if len(args.magick_executable) > 0 else "magick"
use_gpu = 1 if not args.no_gpu else 0  # 默认使用gpu


# 加载相机参数
camera_params = load_camera_parameters(args.yaml_file)
# 提取相机内参
fx = camera_params['Camera.fx']
fy = camera_params['Camera.fy']
cx = camera_params['Camera.cx']
cy = camera_params['Camera.cy']

# 提取畸变系数
k1 = camera_params['Camera.k1']
k2 = camera_params['Camera.k2']
p1 = camera_params['Camera.p1']
p2 = camera_params['Camera.p2']
# k3 = camera_params['Camera.k3']
# k4 = camera_params['Camera.k4']
camera_matrix=str(fx)+","+str(fy)+","+str(cx)+","+str(cx)
dist_coeffs=str(k1)+","+str(k2)+","+str(p1)+","+str(p2)
all_camInfo="\""+camera_matrix+","+dist_coeffs+"\""

print("相机内参","fx",fx,'fy',fy,'cx',cx,'cy',cy,'k1',k1,'k2',k2,'p1',p1,'p2',p2)
print("相机内参",all_camInfo)

# # 构造相机内参矩阵和畸变系数数组
# camera_matrix = np.array([[fx, 0, cx],
#                             [0, fy, cy],
#                             [0, 0, 1]], dtype=np.float64)

# #dist_coeffs = np.array([k1, k2, p1, p2, k3, k4], dtype=np.float64)
# dist_coeffs = np.array([k1, k2, p1, p2,k3], dtype=np.float64)


# 默认不跳过skip_matching
if not args.skip_matching:
    os.makedirs(args.source_path + "/distorted/sparse", exist_ok=True)

    ## Feature extraction 特征提取 ===================
    # --ImageReader.camera_params "fx, fy, cx, cy, k1, k2, p1, p2, k3"
    feat_extracton_cmd = colmap_command + " feature_extractor "\
        "--database_path " + args.source_path + "/distorted/database.db \
        --image_path " + args.source_path + "/input \
        --ImageReader.single_camera true \
        --ImageReader.camera_params "+ all_camInfo + " \
        --ImageReader.camera_model " + args.camera + " \
        --SiftExtraction.use_gpu " + str(use_gpu) + " \
        --SiftExtraction.estimate_affine_shape true \
        --SiftExtraction.domain_size_pooling true \
        --SiftExtraction.num_threads -1 " 
        
    exit_code = os.system(feat_extracton_cmd)
    if exit_code != 0:
        logging.error(f"Feature extraction failed with code {exit_code}. Exiting.")
        exit(exit_code)

    '''
    exhaustive_matcher：针对少量图像（几百张量级），可以获得足够快且最好的重建结果。它将每张图像与其余所有图像进行匹配，不过 block size 可能限制同时加载到内存中的图像数量。
    sequential_matcher：针对顺序采集的视频图像，由于相邻帧存在视觉上的重叠且没有必要进行完全匹配，它只匹配视频流中的相邻帧。同时，这种匹配方式能够基于 vocabulary tree 进行回环检测。最后，帧之间的前后关系由图像文件名给定，与数据集中的存储顺序无关。
    vocab_tree_matcher：针对大量图像（几千帧量级），可以通过提供 vocabulary tree 从而快速检索视觉上最相近的图像进行匹配。
    spatial_matcher：针对能够提供准确定位信息的图像，可以通过对应图像采集时的 GPS 信息从而仅匹配空间位置上相近的图像。
    transitive_matcher：基于传递规则使用已有的特征匹配关系确定更完全的匹配图，即 A 与 B 匹配，B 与 C 匹配，那将直接匹配 A 和 C。
    Custom Matching：通过 text 文件指定图像的匹配关系，如果是导入的特征可以进一步指定两张图像之间特征的匹配关系。
    '''
    ## Feature matching 特征匹配 ===================
    feat_matching_cmd = colmap_command + " sequential_matcher \
        --database_path " + args.source_path + "/distorted/database.db \
        --SiftMatching.use_gpu " + str(use_gpu) +" \
        --SiftMatching.guided_matching true \
        --SiftMatching.num_threads -1 " 
    exit_code = os.system(feat_matching_cmd)
    if exit_code != 0:
        logging.error(f"Feature matching failed with code {exit_code}. Exiting.")
        exit(exit_code)

    ### Bundle adjustment 优化重建 ==================
    #在执行特征提取与匹配后通过 sfM 进行稀疏三维重建与构图。
    # The default Mapper tolerance is unnecessarily large,
    # decreasing it speeds up bundle adjustment steps.
    '''
    --database_path: 包含匹配特征的数据库路径。
    --image_path: 用于建图的输入图像路径。
    --output_path: 存储重建结果的目录。
    --Mapper.ba_local_max_refinements: 本地束调整的最大精细调整次数。
    --Mapper.ba_local_max_num_iterations: 本地束调整的最大迭代次数。
    --Mapper.max_extra_param: 优化中额外参数的最大值。
    --Mapper.ba_refine_principal_point: 是否对主点进行精细调整（0 表示不调整）。
    --Mapper.ba_refine_focal_length: 是否对焦距进行精细调整（0 表示不调整）。
    --Mapper.ba_refine_extra_params: 是否对其他相机参数进行精细调整（0 表示不调整）。
    --Mapper.tri_ignore_two_view_tracks: 在三角测量中忽略两视图轨迹（false/true）。
    --Mapper.num_threads: 建图时使用的线程数（-1 表示使用所有可用线程）。
    '''
    mapper_cmd = colmap_command + " mapper " \
        "--database_path " + args.source_path + "/distorted/database.db " \
        "--image_path " + args.source_path + "/input " \
        "--output_path " + args.source_path + "/distorted/sparse " + \
        "--Mapper.ba_global_function_tolerance=0.000001 " + \
        "--Mapper.ba_local_max_num_iterations 100 " + \
        "--Mapper.max_extra_param 99999 " + \
        "--Mapper.ba_refine_principal_point 0 " + \
        "--Mapper.ba_refine_focal_length 0 " + \
        "--Mapper.ba_refine_extra_params 0 " + \
        "--Mapper.tri_ignore_two_view_tracks false " + \
        "--Mapper.num_threads -1 "

        
    exit_code = os.system(mapper_cmd)
    if exit_code != 0:
        logging.error(f"Mapper failed with code {exit_code}. Exiting.")
        exit(exit_code)

    # 输出转换为txt
    mapper_cmd = (colmap_command + " model_converter \
        --input_path " + args.source_path + "/distorted/sparse/0 \
        --output_type "  +  "txt \
        --output_path "  + args.source_path + "/distorted/sparse/0"
        )
    exit_code = os.system(mapper_cmd)
    if exit_code != 0:
        logging.error(f"bin转换txt失败{exit_code}. Exiting.")
        exit(exit_code)

### Image undistortion
## We need to undistort our images into ideal pinhole intrinsics.
'''
--image_path:
输入图像的路径。可以是单个图像文件或包含多个图像的目录。
--output_path: 默认images
去畸变后的图像输出路径。可以是一个目录，COLMAP 将在该目录下生成去畸变后的图像。
--input_path:
输入相机参数和姿态文件的路径，通常是 COLMAP 数据库或导出的模型文件。
--output_type:
指定输出图像的格式，如 PNG、JPEG 等。
--max_image_size:
指定输出图像的最大尺寸。COLMAP 会根据原图像的比例缩放图像。
--remove_uncovered:
选择是否移除未被覆盖的区域，这在某些情况下可以用来减少输出图像的数量。
--image_list_path:
包含要去畸变的图像列表的文件路径。
'''
print("===============准备开始矫正================")
time.sleep(5)



img_undist_cmd = (colmap_command + " image_undistorter \
    --image_path " + args.source_path + "/input \
    --input_path " + args.source_path + "/distorted/sparse/0 \
    --output_path " + args.output_path + "\
    --output_type COLMAP")
exit_code = os.system(img_undist_cmd)
if exit_code != 0:
    logging.error(f"Mapper failed with code {exit_code}. Exiting.")
    exit(exit_code)

# 创建sparse文件夹
files = os.listdir(args.output_path + "/sparse")
os.makedirs(args.output_path + "/sparse/0", exist_ok=True)
# Copy each file from the source directory to the destination directory
for file in files:
    if file == '0':
        continue
    source_file = os.path.join(args.output_path, "sparse", file)
    destination_file = os.path.join(args.output_path, "sparse", "0", file)
    shutil.move(source_file, destination_file)

# 输出转换为txt
mapper_cmd = (colmap_command + " model_converter \
    --input_path " + args.output_path + "/sparse/0 \
    --output_type "  +  "txt \
    --output_path "  + args.output_path + "/sparse/0"
    )
exit_code = os.system(mapper_cmd)
if exit_code != 0:
    logging.error(f"bin转换txt失败{exit_code}. Exiting.")
    exit(exit_code)


if(args.resize):
    print("Copying and resizing...")

    # Resize images.
    os.makedirs(args.source_path + "/images_2", exist_ok=True)
    os.makedirs(args.source_path + "/images_4", exist_ok=True)
    os.makedirs(args.source_path + "/images_8", exist_ok=True)
    # Get the list of files in the source directory
    files = os.listdir(args.source_path + "/images")
    # Copy each file from the source directory to the destination directory
    for file in files:
        source_file = os.path.join(args.source_path, "images", file)

        destination_file = os.path.join(args.source_path, "images_2", file)
        shutil.copy2(source_file, destination_file)
        exit_code = os.system(magick_command + " mogrify -resize 50% " + destination_file)
        if exit_code != 0:
            logging.error(f"50% resize failed with code {exit_code}. Exiting.")
            exit(exit_code)

        destination_file = os.path.join(args.source_path, "images_4", file)
        shutil.copy2(source_file, destination_file)
        exit_code = os.system(magick_command + " mogrify -resize 25% " + destination_file)
        if exit_code != 0:
            logging.error(f"25% resize failed with code {exit_code}. Exiting.")
            exit(exit_code)

        destination_file = os.path.join(args.source_path, "images_8", file)
        shutil.copy2(source_file, destination_file)
        exit_code = os.system(magick_command + " mogrify -resize 12.5% " + destination_file)
        if exit_code != 0:
            logging.error(f"12.5% resize failed with code {exit_code}. Exiting.")
            exit(exit_code)

print("Done.")