处理BOP数据集，将其和COCO数据集结合

BOP

取消映射关系，并自增80
取消文件名的images前缀

import os
import json
from tqdm import tqdm
import argparse

parser = argparse.ArgumentParser()
parser.add_argument('--json_path', default='H:/Dataset/COCO/train_pbr/000002/coco/annotations/scene_gt_coco.json', type=str,
                    help="input: coco format(json)")
parser.add_argument('--save_path', default='H:/Dataset/COCO/train_pbr/000002/coco/labels', type=str,
                    help="specify where to save the output dir of labels")
arg = parser.parse_args()


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

    x = x * dw
    w = w * dw
    y = y * dh
    h = h * dh
    return (x, y, w, h)


if __name__ == '__main__':
    json_file = arg.json_path  # COCO Object Instance 类型的标注
    ana_txt_save_path = arg.save_path  # 保存的路径

    data = json.load(open(json_file, 'r'))
    if not os.path.exists(ana_txt_save_path):
        os.makedirs(ana_txt_save_path)

    id_map = {}  # coco数据集的id不连续！重新映射一下再输出！
    for i, category in enumerate(data['categories']):
        id_map[category['id']] = i

    # 通过事先建表来降低时间复杂度
    max_id = 0
    for img in data['images']:
        max_id = max(max_id, img['id'])
    # 注意这里不能写作 [[]]*(max_id+1)，否则列表内的空列表共享地址
    img_ann_dict = [[] for i in range(max_id + 1)]
    for i, ann in enumerate(data['annotations']):
        img_ann_dict[ann['image_id']].append(i)

    for img in tqdm(data['images']):
        filename = img["file_name"]
        img_width = img["width"]
        img_height = img["height"]
        img_id = img["id"]
        head, tail = os.path.splitext(filename)
        head2 = head.split("/")
        head3 = head2[1]
        ana_txt_name = head3 + ".txt"  # 对应的txt名字，与jpg一致
        f_txt = open(os.path.join(ana_txt_save_path, ana_txt_name), 'w')
        '''for ann in data['annotations']:
            if ann['image_id'] == img_id:
                box = convert((img_width, img_height), ann["bbox"])
                f_txt.write("%s %s %s %s %s\n" % (id_map[ann["category_id"]], box[0], box[1], box[2], box[3]))'''
        # 这里可以直接查表而无需重复遍历
        for ann_id in img_ann_dict[img_id]:
            ann = data['annotations'][ann_id]
            box = convert((img_width, img_height), ann["bbox"])
            #print(box[0],box[1],box[2],box[3])
            f_txt.write("%s %s %s %s %s\n" % (ann["category_id"]+80, box[0], box[1], box[2], box[3]))
        f_txt.close()