tools/voc2coco.py at master · BeingDog/tools · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
import cv2
import json
from tqdm import tqdm
from glob import glob
import xmltodict
import argparse


def coco_format_init(classes):
    '''
        给 anno["categories"] 赋值
    '''
    anno_dict = {}
    anno_dict["licenses"] = [{"id":0, "name":"", "url":""}]
    anno_dict["info"] = {
        "contributor":"",
        "data_created":"",
        "description":"",
        "url":"",
        "version":"",
        "year":""
    }
    anno_dict["categories"] = []
    for i, name in enumerate(classes):
        anno_dict["categories"].append(
            {
                "supercategory": "_",
                "id": i,
                "name": name
            }
        )
    anno_dict["images"] = []
    anno_dict["annotations"] = []
    return anno_dict


def images_info_set(anno_dict, paths):
    '''
        给 anno_dict["images"] 赋值
    '''
    id = 0
    pbar = tqdm(total=len(paths))
    images_shapes = {}
    images_ids = {}
    for i, path in enumerate(paths):
        pbar.update(1)
        try:
            img = cv2.imread(path)
            h, w, c = img.shape
            imginfo = {}
            imginfo["coco_url"] = ""
            imginfo["data_captured"] = ""
            imginfo["file_name"] = path.split('/')[-1]
            imginfo["flickr_url"] = ""
            imginfo["license"] = 0
            imginfo["id"] = id
            images_ids[i] = id
            id += 1
            imginfo["height"] = h
            imginfo["width"] = w
        except Exception as e:
            print("{} is broke! ".format(path), e)
            continue
        anno_dict["images"].append(imginfo)
        images_shapes[i] = [h, w]
    return images_shapes, images_ids

def labels_info_set(anno_dict, paths, images_shapes, images_ids, classes):
    '''
        给 anno_dict["annotations"]赋值
    '''
    pbar = tqdm(total=len(paths))
    cls_map = {}
    for i, name in enumerate(classes):
        cls_map[name] = i
    print(cls_map)

    j = 0
    for i, path in enumerate(paths):
        pbar.update(1)
        try:
            with open(path, 'r', encoding='utf-8') as f:
                xml_anno = xmltodict.parse(f.read())
            bboxes = []
            clses = []
            A = "annotation" if "annotation" in xml_anno else "Annotation"
            if "object" in xml_anno[A].keys():
                objs = xml_anno[A]["object"]
                if not isinstance(objs, list):
                    objs = [objs]
                for obj in objs:
                    if obj["name"] in classes:
                        bboxes.append([obj["bndbox"]["xmin"], obj["bndbox"]["ymin"], obj["bndbox"]["xmax"], obj["bndbox"]["ymax"]])
                        clses.append(cls_map[obj["name"]])
        except Exception as e:
            print(e)
            continue
        for x in range(len(bboxes)):
            try:
                bbox = [float(num) for num in bboxes[x]]
                bbox[2] = bbox[2] - bbox[0]
                bbox[3] = bbox[3] - bbox[1]
                seg = bbox[:2]
                seg.extend([bbox[0], bbox[1]+bbox[3]])
                seg.extend([bbox[0] + bbox[2], bbox[1] + bbox[3]])
                seg.extend([bbox[0] + bbox[2], bbox[1]])
                labelinfo = {}
                labelinfo["area"] = bbox[2] * bbox[3]
                labelinfo["bbox"] = bbox
                labelinfo["segmentation"] = [seg]
                labelinfo["image_id"] = images_ids[i]
                labelinfo["id"] = j
                j += 1
                labelinfo["iscrowd"] = 0
                labelinfo["category_id"] = int(clses[x])
            except Exception as e:
                print("label convert error! ", e)
                continue
            anno_dict["annotations"].append(labelinfo)


def main():
    parser = argparse.ArgumentParser(
        description='This script support converting voc format xmls to coco format json')
    parser.add_argument('--classes', type=list, default=["cloud", "person"])
    parser.add_argument('--img_dir', type=str, default="temp/*.jpg",
                        help='path to image files directory.')
    parser.add_argument('--output', type=str, default='output.json', help='path to output json file')
    args = parser.parse_args()

    anno_dict = coco_format_init(args.classes)
    print("ann_dict Init finised. class is {}".format(anno_dict["categories"]))
    image_paths = glob(args.img_dir)
    print(image_paths)
    label_paths = [p.replace(".jpg", ".xml") for p in image_paths]
    image_shapes, image_ids = images_info_set(anno_dict, image_paths)
    print(image_ids)
    print("ann_dict[images] generated.")
    labels_info_set(anno_dict, label_paths, image_shapes, image_ids, args.classes)
    print("ann_dict[annotation] generated.")

    with open(args.output, 'w') as f:
        json.dump(anno_dict, f, indent=4)

if __name__=="__main__":
    main()