-
Notifications
You must be signed in to change notification settings - Fork 48
/
ArgoverseDataset.py
279 lines (258 loc) · 16.6 KB
/
ArgoverseDataset.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
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
''' ArgoverseForecastDataset继承了torch.utils.data.Dataset，实现三个函数用于初始化和获取地图数据
加载Argoverse HD map 和 Forecast 数据集，并将地图和轨迹数据进行向量化（vector map）归一化等处理
由__getitem__函数将处理过的数据转为tensor并返回 '''
import torch
import torch.utils.data
import torchvision.transforms as T
import numpy as np
import argoverse
from argoverse.map_representation.map_api import ArgoverseMap
from argoverse.data_loading.argoverse_tracking_loader import ArgoverseTrackingLoader
from argoverse.data_loading.argoverse_forecasting_loader import ArgoverseForecastingLoader
from tqdm import tqdm
import matplotlib.pyplot as plt
from common import *
import json
import pickle
import sys
## 根据轨迹点p0(x0,y0), p1(x1,y1)计算它们组成的向量的2x2旋转矩阵
def get_rotate_matrix(trajectory):
x0, y0, x1, y1 = trajectory.flatten()
vec1 = np.array([x1 - x0, y1 - y0])
vec2 = np.array([0, 1])
cosalpha = vec1.dot(vec2) / (np.sqrt(vec1.dot(vec1)) * 1 + 1e-5)
sinalpha = np.sqrt(1 - cosalpha * cosalpha)
if x1 - x0 < 0:
sinalpha = -sinalpha
rotate_matrix = np.array([[cosalpha, -sinalpha], [sinalpha, cosalpha]])
return rotate_matrix
class ArgoverseForecastDataset(torch.utils.data.Dataset):
def __init__(self, cfg):
super().__init__()
self.am = ArgoverseMap() # HD map in argoverse-api/map_files
self.axis_range = self.get_map_range(self.am) # 获取整个城市的坐标范围，用于归一化坐标
self.city_halluc_bbox_table, self.city_halluc_tableidx_to_laneid_map = self.am.build_hallucinated_lane_bbox_index() # 用于快速查询车道
self.laneid_map = self.process_laneid_map() # {'PIT': {9604854: '0'}, 'MIA': {9605252: '0'}}
self.vector_map, self.extra_map = self.generate_vector_map() # get HD map and convert to vector, extra_map includes OBJECT_TYP, turn_direction, lane_id, in_intersection, has_traffic_control
# am.draw_lane(city_halluc_tableidx_to_laneid_map['PIT']['494'], 'PIT')
# self.save_vector_map(self.vector_map)
self.last_observe = cfg['last_observe']
##set root_dir to the correct path to your dataset folder
self.root_dir = cfg['data_locate']
self.device = cfg['device']
self.afl = ArgoverseForecastingLoader(self.root_dir)
# self.map_feature = dict(PIT=[], MIA=[])
self.city_name, self.center_xy, self.rotate_matrix = dict(), dict(), dict()
def __len__(self):
return len(self.afl)
def __getitem__(self, index): # 迭代获取数据函数，在该函数中读取了trajectory数据，同时对坐标进行了一系列预处理，最后转换为归一化的轨迹和地图tensor
# self.am.find_local_lane_polygons()
self.trajectory, city_name, extra_fields = self.get_trajectory(index) # 由索引获取一段轨迹，见图2021-10-11 21-36-01 的屏幕截图.png
traj_id = extra_fields['trajectory_id'][0] # 将xxx.csv中文件名作为scenario id，数据见data.txt，由于是同一段轨迹，所以id是一样的，所以我们取第0个
self.city_name.update({str(traj_id): city_name})
center_xy = self.trajectory[self.last_observe-1][1] # 将第last_observe-1个轨迹点作为中心点
self.center_xy.update({str(traj_id): center_xy}) # 选取一个中心点，用于归一化处理,数据见data.txt, {'425': array([ 186.48895452, 1560.94612336])}
trajectory_feature = (self.trajectory - np.array(center_xy).reshape(1, 1, 2)).reshape(-1, 4) # [[x1,y1,x2,y2],[x3,y3,x4,y4],...]
rotate_matrix = get_rotate_matrix(trajectory_feature[self.last_observe, :]) # get rotate coordinate from last_observe vector
self.rotate_matrix.update({str(traj_id): rotate_matrix})
# 如果所有轨迹点都在一条直线上，那么旋转后的点都在y轴上
trajectory_feature = ((trajectory_feature.reshape(-1, 2)).dot(rotate_matrix.T)).reshape(-1, 4) # 轨迹特征旋转并reshape
# print('trajectory_feature before normalize :')
# print(trajectory_feature)
trajectory_feature = self.normalize_coordinate(trajectory_feature, city_name) #
# np.savetxt('traj.txt',trajectory_feature,fmt='%0.8f')
# 轨迹特征为6维[x1,y1,x2,y2,TIMESTAMP,trajectory_id]
# self.traj_feature = torch.from_numpy(np.hstack((trajectory_feature,
# extra_fields['TIMESTAMP'].reshape(-1, 1),
# # extra_fields['OBJECT_TYPE'].reshape(-1, 1),
# extra_fields['trajectory_id'].reshape(-1, 1)))).float()
self.traj_feature = torch.from_numpy(trajectory_feature).float()
# map_feature_dict = dict(PIT=[], MIA=[])
# 地图特征为8维[v0x,v0y,v1x,v1y,turn_direction,in_intersection,has_traffic_control,lane_id]
# 上面得到了self.center_xy和self.rotate_matrix，下面对每个点地图也需要做相应的去中心化和旋转
self.map_feature = []
# mf = []
lane_ids = self.am.get_lane_ids_in_xy_bbox(center_xy[0], center_xy[1], city_name, 20)
for id in lane_ids:
index_str = self.laneid_map[city_name][id]
i = int(index_str)
vecmap_feature = (self.vector_map[city_name][i] - np.array(center_xy).reshape(1, 1, 2)).reshape(-1, 2) # 地图点去中心点
vecmap_feature = (vecmap_feature.dot(rotate_matrix.T)).reshape(-1, 4) # 旋转并reshape
vecmap_feature = self.normalize_coordinate(vecmap_feature, city_name) # 再归一化
# mf.append(vecmap_feature)
tmp_tensor = torch.from_numpy(np.hstack((vecmap_feature,
self.extra_map[city_name]['turn_direction'][i],
self.extra_map[city_name]['in_intersection'][i],
self.extra_map[city_name]['has_traffic_control'][i],
# self.extra_map[city_name]['OBJECT_TYPE'][i],
self.extra_map[city_name]['lane_id'][i])))
self.map_feature.append(tmp_tensor)
# map_length = len(self.map_feature)
# if map_length > 32:
# self.map_feature = self.map_feature[:32]
# elif map_length < 32:
# need_align = True
# while need_align:
# for i in range(map_length):
# self.map_feature.append(self.map_feature[i])
# if len(self.map_feature) == 32:
# need_align = False
# break
# for city in ['PIT', 'MIA']:
# for i in range(len(self.vector_map[city])):
# map_feature = (self.vector_map[city][i] - np.array(center_xy).reshape(1, 1, 2)).reshape(-1, 2) # 地图点减去中心点作为map_feature
# map_feature = (map_feature.dot(rotate_matrix.T)).reshape(-1, 4) # 地图特征旋转并reshape
# map_feature = self.normalize_coordinate(map_feature, city)
# tmp_tensor = torch.from_numpy(np.hstack((map_feature,
# self.extra_map[city]['turn_direction'][i],
# self.extra_map[city]['in_intersection'][i],
# self.extra_map[city]['has_traffic_control'][i],
# # self.extra_map[city]['OBJECT_TYPE'][i],
# self.extra_map[city]['lane_id'][i])))
# map_feature_dict[city].append(tmp_tensor.float())
# # self.map_feature[city] = np.array(self.map_feature[city])
# self.map_feature[city] = map_feature_dict[city]
# self.map_feature['city_name'] = city_name
# mapfeature = np.vstack(mf)
# np.savetxt('map.txt',mapfeature,fmt='%0.8f')
# sys.exit()
return self.traj_feature, self.map_feature # 返回的是一条5s轨迹向量(49,6)和中心点周围的n个地图向量(n,18,8)
def get_trajectory(self, index):
seq_path = self.afl.seq_list[index]
data = self.afl.get(seq_path).seq_df # Get the dataframe for the current sequence. 见docs/data.txt
data = data[data['OBJECT_TYPE'] == 'AGENT'] # will get AGENT traject, 取出所有agent的轨迹
extra_fields = dict(TIMESTAMP=[], OBJECT_TYPE=[], trajectory_id=[])
polyline = []
j = int(str(seq_path).split('/')[-1].split('.')[0]) # forecating sequence 123.cvs文件名去掉后缀(一串数字)
flag = True
city_name = ''
for _, row in data.iterrows():
if flag:
xlast = row['X']
ylast = row['Y']
tlast = row['TIMESTAMP']
city_name = row['CITY_NAME']
flag = False
continue
startpoint = np.array([xlast, ylast]) # 相邻点组成向量
endpoint = np.array([row['X'], row['Y']])
# plt.annotate('', xy=(endpoint[0],endpoint[1]),xytext=(startpoint[0],startpoint[1]),arrowprops=dict(arrowstyle="->",connectionstyle="arc3"))
xlast = row['X']
ylast = row['Y']
extra_fields['TIMESTAMP'].append(tlast)
extra_fields['OBJECT_TYPE'].append(0) # 'AGENT'
extra_fields['trajectory_id'].append(j) # 'AGENT'
tlast = row['TIMESTAMP']
polyline.append([startpoint, endpoint])
extra_fields['TIMESTAMP'] = np.array(extra_fields['TIMESTAMP'])
extra_fields['TIMESTAMP'] -= np.min(extra_fields['TIMESTAMP']) # adjust time stamp
extra_fields['OBJECT_TYPE'] = np.array(extra_fields['OBJECT_TYPE'])
extra_fields['trajectory_id'] = np.array(extra_fields['trajectory_id'])
# plt.show()
return np.array(polyline), city_name, extra_fields
def generate_vector_map(self): # 读取HD map并转换成vector，返回vector map和由其他信息组成的extra_map
vector_map = {'PIT': [], 'MIA': []}
extra_map = {'PIT': dict(OBJECT_TYPE=[], turn_direction=[], lane_id=[], in_intersection=[],
has_traffic_control=[]),
'MIA': dict(OBJECT_TYPE=[], turn_direction=[], lane_id=[], in_intersection=[],
has_traffic_control=[])}
polyline = []
# index = 1
pbar = tqdm(total=17326) # 进度条
pbar.set_description("Generating Vector Map")
# city_name = 'MIA'
# for i in range(1):
# key = 9624155 + i
# pts = self.am.get_lane_segment_polygon(key, city_name)
# pts = pts[:,:2]
# print(pts)
# x1 = pts[:,0]
# y1 = pts[:,1]
# plt.plot(x1, y1,'ro')
# pts_len = pts.shape[0] // 2 # 21 // 2 返回10
# positive_pts = pts[:pts_len, :2] # 车道左边界(x,y)坐标
# negative_pts = pts[pts_len:2 * pts_len, :2] # 右边界
# for i in range(pts_len - 1):
# v1 = np.array([positive_pts[i], positive_pts[i + 1]]) # 车道左边界向量
# v2 = np.array([negative_pts[pts_len - 1 - i], negative_pts[pts_len - i - 2]]) # 右边界向量
# plt.annotate('', xy=(positive_pts[i+1][0],positive_pts[i+1][1]),xytext=(positive_pts[i][0],positive_pts[i][1]),arrowprops=dict(arrowstyle="->",connectionstyle="arc3"))
# plt.annotate('', xy=(negative_pts[pts_len - i - 2][0],negative_pts[pts_len - i - 2][1]),xytext=(negative_pts[pts_len - 1 - i][0],negative_pts[pts_len - 1 - i][1]),arrowprops=dict(arrowstyle="->",connectionstyle="arc3"))
# plt.show()
for city_name in ['PIT', 'MIA']:
for key in self.laneid_map[city_name]: # lane id
# 由lane_id (key) 和 city_name 返回的pts是由21个三维坐标点(x,y,z)组成的一个闭合车道（第一个点和最后一个点重合）
pts = self.am.get_lane_segment_polygon(key, city_name) # get lane boundries sample points, stitch them as vector (specified in the paper)
turn_str = self.am.get_lane_turn_direction(key, city_name)
if turn_str == 'LEFT':
turn = -1
elif turn_str == 'RIGHT':
turn = 1
else:
turn = 0
pts_len = pts.shape[0] // 2 # 21 // 2 返回10
positive_pts = pts[:pts_len, :2] # 车道左边界(x,y)坐标
negative_pts = pts[pts_len:2 * pts_len, :2] # 右边界
# if city_name == 'PIT':
# plt.plot(pts[:pts_len, 0], pts[:pts_len, 1])
# plt.plot(pts[pts_len:2 * pts_len, 0], pts[pts_len:2 * pts_len, 1])
polyline.clear()
for i in range(pts_len - 1):
v1 = np.array([positive_pts[i], positive_pts[i + 1]]) # 车道左边界向量，二维向量只用xy坐标
v2 = np.array([negative_pts[pts_len - 1 - i], negative_pts[pts_len - i - 2]]) # 右边界向量
polyline.append(v1)
polyline.append(v2)
# extra_field['table_index'] = self.laneid_map[city_name][key]
repeat_t = 2*(pts_len-1)
# 最后得到的polyline是18维的，每一维是用两个点表示的向量，转成np.array再加到vector_map
vector_map[city_name].append(np.array(polyline).copy())
extra_map[city_name]['turn_direction'].append(np.repeat(turn, repeat_t, axis=0).reshape(-1, 1))
extra_map[city_name]['OBJECT_TYPE'].append(np.repeat(-1, repeat_t, axis=0).reshape(-1, 1)) #HD Map
extra_map[city_name]['lane_id'].append(np.repeat(int(key), repeat_t, axis=0).reshape(-1, 1))
extra_map[city_name]['in_intersection'].append(np.repeat(
1 * self.am.lane_is_in_intersection(key, city_name), repeat_t, axis=0).reshape(-1, 1))
extra_map[city_name]['has_traffic_control'].append(np.repeat(
1 * self.am.lane_has_traffic_control_measure(key, city_name), repeat_t, axis=0).reshape(-1, 1))
# if index > 10:
# break
# index = index + 1
pbar.update(1)
pbar.close()
# plt.show()
# mylog = open('extra_map.txt', mode = 'a',encoding='utf-8')
# print(extra_map, file=mylog)
print("Generate Vector Map Successfully!")
return vector_map, extra_map #vector_map:list
def process_laneid_map(self):
laneid_map = {}
tmp_map = {}
tmp1_map = {}
for key in self.city_halluc_tableidx_to_laneid_map['PIT']:
tmp_map[self.city_halluc_tableidx_to_laneid_map['PIT'][key]] = key
laneid_map['PIT'] = tmp_map
for key in self.city_halluc_tableidx_to_laneid_map['MIA']:
tmp1_map[self.city_halluc_tableidx_to_laneid_map['MIA'][key]] = key
laneid_map['MIA'] = tmp1_map
return laneid_map
def get_map_range(self, am):
map_range = dict(PIT={}, MIA={})
for city_name in ['PIT', 'MIA']: # 匹兹堡，迈阿密
poly = am.get_vector_map_lane_polygons(city_name) # Get list of lane polygons for a specified city
poly_modified = (np.vstack(poly))[:, :2] # 所有地图数据垂直排列，取前两列xy
max_coordinate = np.max(poly_modified, axis=0) # xy轴的最大值和最小值
min_coordinate = np.min(poly_modified, axis=0)
map_range[city_name].update({'max': max_coordinate})
map_range[city_name].update({'min': min_coordinate})
print(city_name + ' map range :')
print(max_coordinate)
print(min_coordinate)
return map_range
def normalize_coordinate(self, array, city_name):
max_coordinate = self.axis_range[city_name]['max']
min_coordinate = self.axis_range[city_name]['min']
array = (100.*(array.reshape(-1, 2)) / (max_coordinate - min_coordinate)).reshape(-1,4)
return array
def save_vector_map(self, vector_map):
save_path = "./data/vector_map/"
for city_name in ['PIT', 'MIA']:
tmp_map = np.vstack(vector_map[city_name]).reshape(-1, 4)
np.save(save_path+city_name+"_vectormap", tmp_map)