-
Notifications
You must be signed in to change notification settings - Fork 3
/
dataset_loc.py
187 lines (160 loc) · 7.13 KB
/
dataset_loc.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
from __future__ import print_function, division
import os
import torch
import numpy as np
import matplotlib.pyplot as plt
from torch.utils.data import Dataset, DataLoader
from torchvision import transforms, utils
import pickle
import cv2
import torchvision
import torch.nn.functional as F
from torch.autograd import Variable
import random
from scipy.spatial.distance import pdist, squareform
class OneHopDataset(Dataset):
def __init__(self, f_name=None, K=3, R=1.5):
self.root_dir = './data'
if f_name:
print('f name = {}'.format(f_name))
self.data = pickle.load(open(f_name, "rb" ))
else:
self.data = pickle.load(open('./airsim_dataset_F3.pkl', "rb" ))
self.K = K
self.R = R
def __len__(self):
return len(self.data)
def __getitem__(self, index):
'''
x_img_paths = self.data[index]['x_img_paths']
fovs = []
x_img = []
for j in range(self.K):
fovs = []
for k in range(1, 51):
img_files = x_img_paths['time-{}_drone_{}'.format(j, k)]
f_img_file = img_files[0]
l_img_file = img_files[1]
r_img_file = img_files[2]
b_img_file = img_files[3]
f_image = cv2.imread(f_img_file)
l_image = cv2.imread(l_img_file)
r_image = cv2.imread(r_img_file)
b_image = cv2.imread(b_img_file)
if type(f_image) == type(None):
if j == 0:
img_files = x_img_paths['time-{}_drone_{}'.format(j+1, k)]
f_img_file = img_files[0]
l_img_file = img_files[1]
r_img_file = img_files[2]
b_img_file = img_files[3]
f_image = cv2.imread(f_img_file)
l_image = cv2.imread(l_img_file)
r_image = cv2.imread(r_img_file)
b_image = cv2.imread(b_img_file)
else:
img_files = x_img_paths['time-{}_drone_{}'.format(j-1, k)]
f_img_file = img_files[0]
l_img_file = img_files[1]
r_img_file = img_files[2]
b_img_file = img_files[3]
f_image = cv2.imread(f_img_file)
l_image = cv2.imread(l_img_file)
r_image = cv2.imread(r_img_file)
b_image = cv2.imread(b_img_file)
fov = np.concatenate([f_image, l_image, r_image, b_image], axis=1)
fov = fov.swapaxes(0, 2).swapaxes(1, 2)
fovs.append(np.expand_dims(fov, axis=0))
fovs = np.expand_dims(np.concatenate(fovs, axis=0), axis=4)
x_img.append(fovs)
x_img = x_img[::-1]
x_img = np.concatenate(x_img, axis=4)
'''
#x_agg = np.clip(self.data[index]['x_agg'], -30, 30)
#a_nets = self.data[index]['a_nets']
#mylist = ['0', '1', '2']
#choice = random.choice(mylist)
#if choice == '0':
# a_nets = self.data[index]['a_nets_com15']
# #print('choose com 1.5')
#elif choice == '1':
# a_nets = self.data[index]['a_nets_com25']
# #print('choose com 2.5')
#else:
# a_nets = self.data[index]['a_nets_com35']
x_locs = self.data[index]['x_locs'] # 50 x 4 x K
n_drone = x_locs.shape[0]
a_nets = np.zeros((n_drone, n_drone, self.K))
for j in range(self.K):
#R = np.random.uniform(low=1.0, high=3.5, size=None)
a_net = get_connectivity(x_locs[:, :, j], self.R)
a_nets[:, :, j] = a_net
a_nets[a_nets != a_nets] = 0
actions = np.clip(self.data[index]['actions'], -1, 1)
aggs = self.data[index]['x_aggs']
sample = {'anets': a_nets, 'actions': actions, 'x_agg': aggs}
return sample
def imread(path, device):
img = cv2.imread(path, cv2.IMREAD_COLOR)
img = img.transpose(2,0,1)
img = torch.from_numpy(img).float().to(device)
img, _ = pad_to_square(img, 0)
img = resize(img, 416).unsqueeze(0)
Tensor = torch.cuda.FloatTensor if torch.cuda.is_available() else torch.FloatTensor
return Variable(img.type(Tensor))
def resize(image, size):
image = F.interpolate(image.unsqueeze(0), size=size, mode="nearest").squeeze(0)
return image
def pad_to_square(img, pad_value):
c, h, w = img.shape
dim_diff = np.abs(h - w)
# (upper / left) padding and (lower / right) padding
pad1, pad2 = dim_diff // 2, dim_diff - dim_diff // 2
# Determine padding
pad = (0, 0, pad1, pad2) if h <= w else (pad1, pad2, 0, 0)
# Add padding
img = F.pad(img, pad, "constant", value=pad_value)
return img, pad
def rescale_boxes(boxes, current_dim, original_shape):
""" Rescales bounding boxes to the original shape """
orig_h, orig_w = original_shape
# The amount of padding that was added
pad_x = max(orig_h - orig_w, 0) * (current_dim / max(original_shape))
pad_y = max(orig_w - orig_h, 0) * (current_dim / max(original_shape))
# Image height and width after padding is removed
unpad_h = current_dim - pad_y
unpad_w = current_dim - pad_x
# Rescale bounding boxes to dimension of original image
boxes[:, 0] = ((boxes[:, 0] - pad_x // 2) / unpad_w) * orig_w
boxes[:, 1] = ((boxes[:, 1] - pad_y // 2) / unpad_h) * orig_h
boxes[:, 2] = ((boxes[:, 2] - pad_x // 2) / unpad_w) * orig_w
boxes[:, 3] = ((boxes[:, 3] - pad_y // 2) / unpad_h) * orig_h
return boxes
def get_connectivity(x, comm_radius):
"""
Get the adjacency matrix of the network based on agent locations by computing pairwise distances using pdist
Args:
x (): current states of all agents
Returns: adjacency matrix of network
"""
n_nodes = x.shape[0]
x_t_loc = x[:, 0:2] # x,y location determines connectivity
a_net = squareform(pdist(x_t_loc.reshape((n_nodes, 2)), 'euclidean'))
a_net = (a_net < comm_radius).astype(float)
np.fill_diagonal(a_net, 0)
return a_net
def main():
print('test dataloader')
drone_dataset = OneHopDataset(f_name='./optimal_K_3_n_vis_24_R_1.5_vinit_3.0_comm_model_disk.pkl')
print(len(drone_dataset))
droneTrainLoader = torch.utils.data.DataLoader(drone_dataset,batch_size=1,shuffle=True, num_workers=4)
max_value = -float("Inf")
for i_batch,sample_batched in enumerate(droneTrainLoader,0):
#print('sample index = {}, agg shape = {}'.format(i_batch,sample_batched['x_agg'].shape))
print('sample index = {}, action shape = {}'.format(i_batch,sample_batched['actions'].shape))
print('sample index = {}, anet shape = {}'.format(i_batch,sample_batched['anets'].shape))
#print('sample index = {}, img shape = {}'.format(i_batch,sample_batched['x_img'].shape))
print('sample index = {}, action = {}'.format(i_batch,sample_batched['actions'].shape))
print('sample index = {}, x aggs = {}'.format(i_batch,sample_batched['x_agg'].shape))
if __name__ == "__main__":
main()