labelme的json文件转换成txt文件（可防止出现负值）

import json
import os

name2id = {"block":0,'can':1,'Charger':2,'cup':3,'mouse':4,'nail clippers':5,'pen':6,'screw':7,'tissue box':8,'toothbrush':9}  # Mapping from label names to IDs

def convert(img_size, box):
    dw = 1. / img_size[0]
    dh = 1. / img_size[1]
    x = (box[0] + box[2]) / 2.0 - 1
    y = (box[1] + box[3]) / 2.0 - 1
    w = box[2] - box[0]
    h = box[3] - box[1]
    x = abs(x * dw)
    w = abs(w * dw)
    y = abs(y * dh)
    h = abs(h * dh)
    return (x, y, w, h)

def decode_json(json_path, txt_output_folder):
    txt_name = os.path.join(txt_output_folder, os.path.basename(json_path)[:-5] + '.txt')
    os.makedirs(os.path.dirname(txt_name), exist_ok=True)
    
    with open(txt_name, 'w') as txt_file:
        try:
            data = json.load(open(json_path, 'r', encoding='gb2312'))
        except UnicodeDecodeError:
            data = json.load(open(json_path, 'r', encoding='utf-8'))

        img_w = data['imageWidth']
        img_h = data['imageHeight']

        for i in data['shapes']:
            label_name = i['label']
            if label_name not in name2id:
                print(f"Skipping label: {label_name}")
                continue
            
            if i['shape_type'] == 'rectangle':
                x1 = int(i['points'][0][0])
                y1 = int(i['points'][0][1])
                x2 = int(i['points'][1][0])
                y2 = int(i['points'][1][1])

                bb = (x1, y1, x2, y2)
                bbox = convert((img_w, img_h), bb)
                txt_file.write(str(name2id[label_name]) + " " + " ".join([str(a) for a in bbox]) + '\n')

if __name__ == "__main__":
    json_folder_path = 'C:\\Users\\周\\Desktop\\原始数据集\\'
    txt_output_folder = 'labels'
    
    for root, _, files in os.walk(json_folder_path):
        for file in files:
            if file.endswith('.json'):
                json_path = os.path.join(root, file)
                print(f'Processing {json_path}')
                decode_json(json_path, txt_output_folder)

def convert(img_size, box):
“”“”“”“”“”“”“”“”“
    x = abs(x * dw)
    w = abs(w * dw)
    y = abs(y * dh)
    h = abs(h * dh)
    return (x, y, w, h)
————————————————

返回时加入绝对值
参考原文链接：https://blog.csdn.net/m0_69331214/article/details/134029719

可以实现在路径下找到json文件，排除标签之外的类别转txt

可以划分yolo数据集

import os
import random
import shutil
from math import floor

def create_directory(dir_path):
    if not os.path.exists(dir_path):
        os.makedirs(dir_path)

def copy_files(files, source_image_folder, source_label_folder, target_folder):
    for file in files:
        base_name = os.path.splitext(file)[0]
        image_file = file
        label_file = base_name + '.txt'
        
        # Copy image file
        shutil.copy(os.path.join(source_image_folder, image_file), os.path.join(target_folder, 'images', image_file))
        
        # Copy corresponding label file if it exists
        if os.path.exists(os.path.join(source_label_folder, label_file)):
            shutil.copy(os.path.join(source_label_folder, label_file), os.path.join(target_folder, 'labels', label_file))

def split_dataset(image_folder, label_folder, output_folder, train_ratio=0.7, val_ratio=0.15, test_ratio=0.15, seed=None):
    if seed is not None:
        random.seed(seed)
    
    # Create output directories
    train_folder = os.path.join(output_folder, 'train')
    val_folder = os.path.join(output_folder, 'val')
    test_folder = os.path.join(output_folder, 'test')

    for folder in [train_folder, val_folder, test_folder]:
        create_directory(folder)
        create_directory(os.path.join(folder, 'images'))
        create_directory(os.path.join(folder, 'labels'))

    # Get list of image files
    image_extensions = ['.jpg', '.jpeg', '.png']
    image_files = [f for f in os.listdir(image_folder) if os.path.splitext(f)[1].lower() in image_extensions]
    
    # Shuffle the list
    random.shuffle(image_files)
    
    # Calculate split sizes
    total_files = len(image_files)
    train_size = floor(total_files * train_ratio)
    val_size = floor(total_files * val_ratio)
    test_size = total_files - train_size - val_size  # Ensure all files are used

    # Split the list
    train_files = image_files[:train_size]
    val_files = image_files[train_size:train_size + val_size]
    test_files = image_files[train_size + val_size:]

    # Copy files to respective folders
    copy_files(train_files, image_folder, label_folder, train_folder)
    copy_files(val_files, image_folder, label_folder, val_folder)
    copy_files(test_files, image_folder, label_folder, test_folder)

    print("Dataset split completed!")
    
if __name__ == "__main__":
    image_folder = 'images'  # Replace with your source images folder
    label_folder = 'labels'  # Replace with your source labels folder
    output_folder = 'data'  # Replace with your target output folder
    
    split_dataset(image_folder, label_folder, output_folder, seed=42)