0基础新手教程：COLMAP3.8将bin转为txt再转为json文件（附代码），其它类型同理

前提：电脑已安装colmap3.8

看一下colmap3.8对类型转换的提供的参数（主要使用 --project_path、--input_path、--output_path、--output_type）：

1.使用命令行将bin文件转为txt文件（红色路径替换为你的实际路劲）：
colmap model_converter --input_path"D:\3DGS\data\yangzl260314\sparse\0"--output_path"D:\3DGS\data\yangzl260314\sparse\0"--output_type TXT

下面是colmap跑出来的，"D:\3DGS\data\yangzl260314\sparse\0"文件夹下的bin文件:

执行后命令行后就得到了points3D.txt、images.txt、cameras.txt：

以points3D.txt为例，查看一下里面的内容（发现内容很杂乱，不易于阅读，因此转为json文件阅读更方便）：

2.使用代码将txt转为json文件，代码如下（红色路径替换为你的实际路劲，直接执行即可）：

pythonD:\3DGS\data\yangzl260314\convert.py-iD:\3DGS\data\yangzl260314\sparse\0

import json import os from argparse import ArgumentParser def parse_colmap_cameras(camera_path): cameras = {} with open(camera_path, 'r') as f: for line in f: line = line.strip() if not line or line.startswith('#'): continue parts = line.split() cam_id = int(parts[0]) model = parts[1] width = int(parts[2]) height = int(parts[3]) params = list(map(float, parts[4:])) cameras[cam_id] = { "camera_id": cam_id, "model": model, "width": width, "height": height, "params": params } return cameras def parse_colmap_images(image_path): images = {} with open(image_path, 'r') as f: lines = [line.strip() for line in f if line.strip() and not line.strip().startswith('#')] for i in range(0, len(lines), 2): img_line = lines[i].split() img_id = int(img_line[0]) qw, qx, qy, qz = map(float, img_line[1:5]) tx, ty, tz = map(float, img_line[5:8]) cam_id = int(img_line[8]) name = img_line[9] pts_line = lines[i+1].split() pts2d = [] for j in range(0, len(pts_line), 3): x = float(pts_line[j]) y = float(pts_line[j+1]) p3d_id = int(pts_line[j+2]) pts2d.append([x, y, p3d_id]) images[img_id] = { "image_id": img_id, "quaternion": [qw, qx, qy, qz], "translation": [tx, ty, tz], "camera_id": cam_id, "name": name, "points2d": pts2d } return images def parse_colmap_points3d(points_path): points3d = {} with open(points_path, 'r') as f: for line in f: line = line.strip() if not line or line.startswith('#'): continue parts = line.split() p3d_id = int(parts[0]) x, y, z = map(float, parts[1:4]) r, g, b = map(int, parts[4:7]) error = float(parts[7]) track = [] for j in range(8, len(parts), 2): img_id = int(parts[j]) p2d_idx = int(parts[j+1]) track.append([img_id, p2d_idx]) points3d[p3d_id] = { "point3d_id": p3d_id, "xyz": [x, y, z], "rgb": [r, g, b], "error": error, "track": track } return points3d if __name__ == "__main__": parser = ArgumentParser("Convert bin to json") parser.add_argument('--input', "-i", required=True, type=str) args = parser.parse_args() camera_file = os.path.join(args.input, "cameras.txt") image_file = os.path.join(args.input, "images.txt") points_file = os.path.join(args.input, "points3D.txt") # 解析 cameras.txt、images.txt、points3D.txt cameras = parse_colmap_cameras(camera_file) images = parse_colmap_images(image_file) points3d = parse_colmap_points3d(points_file) # 分别保存 cameras.json、images.json、points3D.json with open(os.path.join(args.input, "cameras.json"), "w", encoding="utf-8") as f: json.dump(cameras, f, indent=2, ensure_ascii=False) with open(os.path.join(args.input, "images.json"), "w", encoding="utf-8") as f: json.dump(images, f, indent=2, ensure_ascii=False) with open(os.path.join(args.input, "points3d.json"), "w", encoding="utf-8") as f: json.dump(points3d, f, indent=2, ensure_ascii=False) print("转换成功！")

这样就得到最终结果了：

3.看一下points3d.json的内容：