File size: 11,634 Bytes
3217786 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 | import os
import glob
import shutil
import numpy as np
import xml.etree.ElementTree as ET
import cv2
def load_camera_xml(xml_path):
tree = ET.parse(str(xml_path))
root = tree.getroot()
Ks = []
poses = []
dist_coeffs = []
for i in range(4):
name = f'camera_{i}_intrinsic_matrix'
K = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [3, 3])
# K[0, 0] = K[0, 0]
# K[1, 1] = K[1, 1]
# K[0, 2] = 0.5 *
# K[1, 2] = 0.5
# Ks.append([1920., 1080., K[0, 0], K[1, 1], 960., 540.])
Ks.append(K)
name = f'camera_{i}_rotation_matrix'
R = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [3, 3])
name = f'camera_{i}_translation_vector'
T = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [3, 1])/1000
pose = np.concatenate((R, T), axis=1)
eye = np.eye(4, dtype=np.float32)
eye[0:3] = pose
pose = eye
poses.append(pose)
name = f'camera_{i}_distortion_coeffs'
dist_coeff = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [1, 5])
dist_coeffs.append(dist_coeff)
Ks = np.array(Ks)
poses = np.linalg.inv(np.array(poses))
dist_coeffs = np.array(dist_coeffs)
# poses = np.array(poses)
return Ks, poses, dist_coeffs
def get_render_pose(para_ext, render_views=120):
Rs = np.repeat(para_ext[0:1, 0:4, 0:3], render_views, 0)
Ts_yz = np.repeat(para_ext[0:1, 1:4, 3:4], render_views, 0)
interp_x = np.linspace(para_ext[0, 0, 3], para_ext[1, 0, 3], render_views, dtype=np.float32)
Ts_x = interp_x[:, np.newaxis, np.newaxis]
Ts = np.concatenate([Ts_x, Ts_yz], 1)
poses = np.concatenate([Rs, Ts], -1)
return poses[:, :, :]
# #将kinect图像复制到images下
# for k in range(10):
# source_path = f'/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/batch_data/yangkun_new/{str(k).zfill(3)}'
# tgt_path = f'/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_{str(k+1)}'
# for i in range(4):
# folder_path = os.path.join(source_path, str(i))
# png_files = glob.glob(os.path.join(folder_path, '*.png'))
# png_files_names = [os.path.basename(file_path) for file_path in png_files]
# for png_file in png_files_names:
# frame_name = png_file.replace('test', 'frame')[0:-4]
# frame_path = os.path.join(tgt_path, frame_name, 'images')
# os.makedirs(frame_path, exist_ok=True)
# shutil.copy(os.path.join(folder_path,png_file), frame_path + f'/{str(i).zfill(4)}.png')
##写入参数
# para_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/parameter.xml'
# Ks, poses, dist_coeffs = load_camera_xml(para_path)
# print(Ks)
# render_views = get_render_pose(poses)
# np.save('/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/render_path.npy', render_views)
# tgt_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_0'
# for i in range(150):
# mesh_path = os.path.join(tgt_path, f'frame{str(i)}','mesh3D')
# os.makedirs(mesh_path, exist_ok=True)
# np.save(mesh_path + '/cameras.npy', poses)
# np.save(mesh_path + '/images.npy', Ks)
#把mask图像复制到masked文件夹下
# source_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/mask/out'
# tgt_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_0'
# for i in range(0, 150):
# mask_path = os.path.join(source_path,f'frame{str(i)}', 'images')
# tgtmask_path = os.path.join(tgt_path,f'frame{str(i)}', 'masked')
# # os.makedirs(tgtmask_path, exist_ok=True)
# shutil.copytree(mask_path, tgtmask_path, dirs_exist_ok=True)
def modify_obj_file_with_color(input_file, output_file, x_offset, z_offset):
with open(input_file, 'r') as file:
lines = file.readlines()
modified_lines = []
for line in lines:
if line.startswith('v '):
parts = line.split()
# 解析顶点坐标
x = float(parts[1]) - x_offset
y = -float(parts[2])
z = -float(parts[3]) - z_offset
# 重新构建顶点行，保留颜色信息（如果存在）
if len(parts) > 4:
color_info = " ".join(parts[4:])
modified_line = f"v {x} {y} {z} {color_info}\n"
else:
modified_line = f"v {x} {y} {z}\n"
modified_lines.append(modified_line)
else:
modified_lines.append(line)
with open(output_file, 'w') as file:
file.writelines(modified_lines)
##修改obj flame的位置
# # 使用示例
# input_file = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_0/frame0/mesh3D/0003_points3Da_1.obj'
# output_file = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_0/frame0/mesh3D/0003_points3D.obj'
# x_offset = -0.15
# z_offset = -0.3
# modify_obj_file_with_color(input_file, output_file, x_offset, z_offset)
##去畸变
# para_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/parameter.xml'
# Ks, poses, dist_coeffs = load_camera_xml(para_path)
# source_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_0'
# for i in range(1, 150):
# for j in range(4):
# pic_path = source_path + f'/frame{str(i)}/masked/{str(j).zfill(4)}.png'
# dist_coeff = dist_coeffs[j, 0, :] # 示例径向畸变系数
# camera_matrix = Ks[j, :, :]
# image = cv2.imread(pic_path)
# undistorted_image = cv2.undistort(image, camera_matrix, dist_coeff)
# cv2.imwrite(source_path + f'/frame{str(i)}/masked/{str(j).zfill(4)}.png', undistorted_image)
# 把初始flame复制到每一个mesh3D文件夹里
def copy_obj_files(source_dir, destination_dir):
try:
# 遍历源文件夹中的所有文件和文件夹
for file_name in os.listdir(source_dir):
# 构建源文件的完整路径
source_file = os.path.join(source_dir, file_name)
# 检查文件是否是 .obj 文件
if file_name.endswith(('.obj', '.npy')):
# 构建目标文件的完整路径
destination_file = os.path.join(destination_dir, file_name)
# 复制 .obj 文件到目标文件夹
shutil.copy(source_file, destination_file)
print(f"已复制文件 {source_file} 到 {destination_file}")
except OSError as e:
print(f"复制文件时发生错误：{e}")
source_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_10'
for i in range(1, 150):
for j in range(4):
obj_path = source_path + f'/frame{str(i)}/mesh3D'
os.makedirs(source_path + f'/frame{str(i)}/mesh3D', exist_ok=True)
copy_obj_files('/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/person_0/frame0/mesh3D', source_path + f'/frame{str(i)}/mesh3D')
# ##写入华为外参
# def load_camera_xml(xml_path):
# tree = ET.parse(str(xml_path))
# root = tree.getroot()
# Ks = []
# poses = []
# dist_coeffs = []
# for i in range(2):
# name = f'camera_{i}_intrinsic_matrix'
# K = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [3, 3])
# # K[0, 0] = K[0, 0]
# # K[1, 1] = K[1, 1]
# # K[0, 2] = 0.5 *
# # K[1, 2] = 0.5
# Ks.append([1080., 1920., 1495.07, 1495.07, 540., 960.])
# # Ks.append(K)
# name = f'camera_{i}_rotation_matrix'
# R = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [3, 3])
# name = f'camera_{i}_translation_vector'
# T = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [3, 1])/1000
# pose = np.concatenate((R, T), axis=1)
# eye = np.eye(4, dtype=np.float32)
# if i == 0:
# eye[0:3] = [[0.061394, 0.869224, -0.490592, 113.487], [0.998099, -0.050835, 0.034837, 13.0695], [0.005341, -0.491798, -0.870693, 269.432]]
# elif i == 1:
# eye[0:3] = [[-0.103505, -0.991412, 0.079936, 15.4573], [-0.992358, 0.108363, 0.059015, 3.2613], [-0.06717, -0.073217, -0.995051, 299.719]]
# pose = eye
# poses.append(pose)
# name = f'camera_{i}_distortion_coeffs'
# dist_coeff = np.reshape(np.array(root.find(name).find('data').text.split(), dtype=np.float32), [1, 5])
# dist_coeffs.append(dist_coeff)
# Ks = np.array(Ks)
# poses = np.linalg.inv(np.array(poses))
# dist_coeffs = np.array(dist_coeffs)
# # poses = np.array(poses)
# return Ks, poses, dist_coeffs
# para_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/parameter.xml'
# Ks, poses, dist_coeffs = load_camera_xml(para_path)
# # print(Ks)
# # render_views = get_render_pose(poses)
# # np.save('/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/kinect_data/render_path.npy', render_views)
# tgt_path = '/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/huaweiperson_0'
# for i in range(5):
# mesh_path = os.path.join(tgt_path, f'frame{str(i)}','mesh3D')
# os.makedirs(mesh_path, exist_ok=True)
# np.save(mesh_path + '/cameras.npy', poses)
# np.save(mesh_path + '/images.npy', Ks)
# from pywavefront import Wavefront
# def ply_to_obj(input_file, output_file):
# try:
# # 读取 PLY 文件
# with open(input_file, 'r') as ply_file:
# lines = ply_file.readlines()
# vertices = []
# # 提取点的坐标和 RGB 值
# for line in lines:
# if line.startswith("element vertex"):
# num_vertices = int(line.split()[2])
# elif line.startswith("end_header"):
# break
# for line in lines[-num_vertices:]:
# x, y, z, r, g, b = line.split()[:6]
# vertices.append((float(x), float(y), float(z), int(r), int(g), int(b)))
# # 保存为 OBJ 文件
# with open(output_file, 'w') as obj_file:
# for vertex in vertices:
# obj_file.write(f"v {vertex[0]} {vertex[1]} {vertex[2]} {vertex[3]} {vertex[4]} {vertex[5]}\n")
# print(f"已将 PLY 文件 {input_file} 转换为 OBJ，并保存为 {output_file}")
# except Exception as e:
# print(f"转换文件时发生错误：{e}")
# # # 例子：将名为 "input.ply" 的 PLY 文件转换为 OBJ，并保存为 "output.obj"
# # ply_to_obj("/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/huaweiperson_0/frame0/mesh3D/0001_points3D.ply", "/data_160TB/2021/yangkun/gaussian-splatting-main-copy/eg3d_human/test_real/poeples/huaweiperson_0/frame0/mesh3D/0001_points3D.obj") |