Labelme转Roboflow

Labelme：json文件
Roboflow：YoloV5格式txt文件

import json
import os
from glob import glob
from tqdm import tqdm
from PIL import Image
import cv2


def Json2Txt(json_path, txt_path):
    label_json = json.load(open(json_path, 'r'))
    img_size = (label_json['imageWidth'], label_json['imageHeight'])
    msg_annotation = label_json['shapes']
    with open(txt_path, 'w') as f:
        for i in msg_annotation:
            # （x1,y1）,（x2,y2） 左上角，右下角
            x1 = float(i['points'][0][0])
            y1 = float(i['points'][0][1])
            x2 = float(i['points'][1][0])
            y2 = float(i['points'][1][1])
            # 转换并归一化
            center_x = (x1 + x2) * 0.5 / img_size[0]
            center_y = (y1 + y2) * 0.5 / img_size[1]
            w = abs((x2 - x1)) * 1.0 / img_size[0]
            h = abs((y2 - y1)) * 1.0 / img_size[1]

            bbox = (center_x, center_y, w, h)  # 仅适用矩形框标注

            cls_id = names.index(i['label'])  # 位于定义类别索引位置

            f.write(str(cls_id) + ' ' +" ".join([str(a) for a in bbox])+"n")  # 生成格式0 cx,cy,w,h


names = ['1', '2', '3', '4', '5', '6', '7']

json_dir = glob('error-detection-v3-yolov5pytorch/train/images/*.json')

for i in tqdm(json_dir):
    
    json_path = i
    
    txt_path = i.replace('.json', '.txt')
    
    Json2Txt(json_path, txt_path)