-
Notifications
You must be signed in to change notification settings - Fork 9
/
Copy pathiou_timestam.py
133 lines (118 loc) · 5.23 KB
/
iou_timestam.py
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
# -*- coding: utf-8 -*
'''
遍历标注文件，按照时间戳对聚类结果进行查找计算IOU
'''
import numpy as np
import os
def getbox():
'''
提取存在标注文件的聚类结果数据
'''
file_num = 0
name = np.array([])
label_file_all = os.listdir(label_file_path)
label_file_all.sort(key=lambda x:float(x[:-4]))
for filename in label_file_all:
star = filename.replace('.txt', '')
star = float(star) # 取出时间戳
if dataset == 'eult':
star = star / 1000000
name = np.append(name, star)
file_num += 1
row_data = np.loadtxt(cluster_file_path, delimiter=' ', usecols=(1, 2, 3, 4, 5, 6, 7)) # 取出数据
if dataset == 'kitti':
row_data[:, 0] = np.trunc(row_data[:, 0] * 1000000)
index = np.array([])
for i in range(file_num):
ind = np.where(row_data[:, 0] == name[i]) #以标注文件为准把时间戳相等的聚类结果索引提取出来
index = np.append(index,ind[0]) #把索引ind存入索引数组index
data_labeled = row_data[index.astype(np.int), :]
return data_labeled
def iou(box1,box2):
'''
3D IoU计算
box表示形式：[x1,y1,z1,x2,y2,z2] 分别是两对角点的坐标
'''
in_w = min(box1[3],box2[3]) - max(box1[0],box2[0])
in_l = min(box1[4],box2[4]) - max(box1[1],box2[1])
in_h = min(box1[5],box2[5]) - max(box1[2],box2[2])
inter = 0 if in_w < 0 or in_l < 0 or in_h < 0 else in_w * in_l * in_h
union = (box1[3] - box1[0]) * (box1[4] - box1[1]) * (box1[5] - box1[2]) + (box2[3] - box2[0]) * (box2[4] - box2[1]) * (box2[5] - box2[2]) - inter
iou = inter / union
return iou
def result():
data_labeled = getbox()
iou_list = np.array([])
skip_frame = np.array([])
name_zero = np.array([]) # 存储出现 IOU = 0 的时间戳
# 按照标注文件进行遍历计算IOU
label_file_all = os.listdir(label_file_path)
label_file_all.sort(key=lambda x:float(x[:-4]))
for filename in label_file_all:
labels = np.loadtxt(label_file_path + filename, delimiter=' ', usecols=(4, 5, 6, 7, 8, 9)) # 取出标注
star = filename.replace('.txt','')
star = float(star)
if dataset == 'eult':
star = star / 1000000
skip_flag = 0
iou1 = np.array([]) # 存放各组标注对应的IOU
if np.size(labels) == 6: #如果该时间戳只有1行
flag1 = 0
max1 = 0
for j in data_labeled:
if j[0] == star:
flag1 = 1
j_data = np.delete(j, 0, axis=0) #删除时间戳行
iou_res = iou(labels, j_data)
iou1 = np.append(iou1, iou_res) #这一行所有IOU值表
if flag1 == 1:
skip_flag = 1
max1 = np.max(iou1)
if max1 == 0:
name_zero = np.append(name_zero, star)
iou_list = np.append(iou_list, max1)
else:
iou_list = np.append(iou_list, max1)
else:
name_zero = np.append(name_zero, star)
iou_list = np.append(iou_list, 0)
else:
for row in labels:
flag2 = 0
max2 = 0
iou2 = np.array([]) # 存储每一行标注数据与算法中的众多框的IOU值
for j in data_labeled:
if j[0]==star:
flag2 = 1
j_data = np.delete(j, 0, axis=0) # 删除时间戳行
iou_res = iou(row, j_data) # 计算每一行与data_labeled中的IOU值
iou2 = np.append(iou2, iou_res)
if flag2 ==1:
skip_flag = 1
max2 = np.max(iou2) #取该行IOU最大值
if max2 == 0:
name_zero = np.append(name_zero, star)
iou1 = np.append(iou1, max2) # 该时间戳getbox每行的IOU值表
else:
iou1 = np.append(iou1, max2) # 该时间戳getbox每行的IOU值表
else:
name_zero = np.append(name_zero, star)
iou1 = np.append(iou1, 0)
iou_list = np.append(iou_list, iou1)
if skip_flag == 0:
skip_frame = np.append(skip_frame, star)
iou_mean = sum(iou_list)/len(iou_list)
print(iou_list)
print('平均IOU为：', iou_mean)
print('保存的IOU数：', len(iou_list))
print('IOU为0的标签数：', len(name_zero))
print('丢失的帧数：', len(skip_frame))
if __name__ =='__main__':
dataset = 'eult' # lcas, eult, kitti
# label_file_path = './iou/lcas_labels/' #标注数据的文件夹地址
# cluster_file_path = './iou/lcas_results/run_clustering_lcas.txt'#算法输出文件的地址
label_file_path = './iou/utbm_labels_car/' #标注数据的文件夹地址
cluster_file_path = './iou/utbm_results/euclidean_clustering_utbm.txt'#算法输出文件的地址
# label_file_path = './iou/kitti_labels_new/' #标注数据的文件夹地址
# cluster_file_path = './iou/kitti_results/adaptive_clustering_kitti.txt'#算法输出文件的地址
IOU_3D = result()