COCO格式json标注数据转标准yolo的txt格式

你可以使用Python编写脚本来将COCO格式的JSON标注数据转换为标准的YOLO格式的TXT文件。下面是一个示例代码，帮助你完成这个转换过程：

import json

def coco_to_yolo(coco_path, yolo_path):
    with open(coco_path, 'r') as f:
        data = json.load(f)
    
    images = data['images']
    annotations = data['annotations']
    categories = data['categories']
    
    # 创建类别索引字典
    category_dict = {}
    for category in categories:
        category_dict[category['id']] = category['name']
    
    with open(yolo_path, 'w') as f:
        for annotation in annotations:
            image_id = annotation['image_id']
            image_info = next((image for image in images if image['id'] == image_id), None)
            if image_info is None:
                continue
            
            width = image_info['width']
            height = image_info['height']
            
            bbox = annotation['bbox']
            x_center = bbox[0] + bbox[2] / 2
            y_center = bbox[1] + bbox[3] / 2
            x_center /= width
            y_center /= height
            bbox_width = bbox[2] / width
            bbox_height = bbox[3] / height
            
            class_id = annotation['category_id']
            class_name = category_dict[class_id]
            
            line = f"{class_name} {x_center} {y_center} {bbox_width} {bbox_height}\n"
            f.write(line)

# 使用示例
coco_to_yolo('coco.json', 'yolo.txt')

你需要将`coco.json`替换为你的COCO格式的JSON文件路径，将`yolo.txt`替换为你想要保存的YOLO格式的TXT文件路径。这段代码将会遍历COCO数据集中的每个标注框，并将其转换为YOLO格式的文本行，每行包含类别名称、边界框相对于图像宽度和高度的归一化坐标以及边界框相对于图像宽度和高度的归一化宽度和高度。

希望这可以帮助到你！如果还有其他问题，请随时提问。