forked from hb0702/Didi_challenge_2017_Python
-
Notifications
You must be signed in to change notification settings - Fork 0
/
cluster_classify_prediction.py
88 lines (62 loc) · 2.56 KB
/
cluster_classify_prediction.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
import numpy as np
import tensorflow as tf
import keras
from keras.models import load_model
import os
import time
from sklearn.cluster import DBSCAN
from keras.optimizers import Adam
from cluster_classify_model import cluster_classify_model
from cluster_classify_util import *
from cluster_classify_train import *
from test_on_udacity_data import *
from keras.utils.generic_utils import get_custom_objects
get_custom_objects().update({"my_loss": my_loss})
def predict(model,lidar, thresh=0.5):
lidar, labels = cluster(lidar)
list_clusters = list(set(labels))
nb_clusters = len(list_clusters)
if nb_clusters == 0:
return np.array([])
list_of_cluster = np.array([lidar[labels == list_clusters[i]] for i in range(nb_clusters)] )
centers = np.zeros((nb_clusters,2))
imgs = np.zeros((nb_clusters, 64,64,2))
for i in range(nb_clusters):
img, center = discretize(list_of_cluster[i])
imgs[i] = img
centers[i] = center
features = model.predict(imgs)
#print('features.shape: ', features.shape)
features_thresh = features[features[:,0] >= thresh]
centers_thresh = centers[features[:,0] >= thresh]
boxes = np.array([gt_box_decode(features_thresh[i], centers_thresh[i], z_min = -1.5) for i in range(len(features_thresh)) ])
return boxes
def predict_test_set(model, test_dir, pred_dir):
if not os.path.exists(pred_dir):
os.mkdir(pred_dir)
print('start predicting ....')
start = time.time()
nb = 0
for bag in os.listdir(test_dir):
bag_dir = os.path.join(test_dir, bag)
pred_bag_dir = os.path.join(pred_dir, bag)
if not os.path.exists(pred_bag_dir):
os.mkdir(pred_bag_dir)
for f in os.listdir(bag_dir):
nb+=1
lidar_file = os.path.join(bag_dir, f)
lidar = np.load(lidar_file)
boxes = predict(model, lidar)
box_file = os.path.join(pred_bag_dir, f.replace('lidar', 'boxes') )
np.save(box_file, boxes)
print('End prediction. Number of frame: {0}. Total time: {1}. Time per frame {2}'.format(nb, int(time.time()-start), (time.time()-start)/nb))
if __name__ == "__main__":
#lidar = np.load('./data/training_didi_data/car_train_edited/bmw_sitting_still/lidar/lidar_100.npy')
#gtbox = np.load('./data/training_didi_data/car_train_gt_box_edited/bmw_sitting_still/gt_boxes3d/gt_boxes3d_100.npy')
#viz_mayavi_with_labels(lidar, gtbox)
model = load_model('./saved_model/last_model.h5')
#boxes = predict(model, lidar)
#viz_mayavi_with_labels(lidar, boxes)
test_dir = './data/test_cars/'
pred_dir = './data/pred_box_cars/'
predict_test_set(model, test_dir, pred_dir)