def test_kitti():
# load net
num_classes = len(KITTI_CLASSES) + 1 # +1 background
net = build_ssd('test', 300, num_classes) # initialize SSD
net.load_state_dict(torch.load(args.trained_model))
net.eval()
print('Finished loading model!')
# load data
testset = KITTIDetection(KITTI_ROOT,['val'], None, KITTIAnnotationTransform())
if args.cuda:
net = net.cuda()
print('test')
cudnn.benchmark = True
# evaluation
test_net(args.save_folder, net, args.cuda, testset, BaseTransform(net.size, (104, 117, 123)),thresh=args.visual_threshold)
if torch.cuda.is_available():
torch.set_default_tensor_type('torch.cuda.FloatTensor')
import time
from ssd import build_ssd
net = build_ssd('test', 300, 4) # initialize SSD
#net.load_weights('../weights/ssd300_mAP_77.43_v2.pth')
#net.load_weights('../weights/VOC.pth')
net.load_weights('../weights/ssd300_Resz_KITTI_80000.pth')
#FOR TESTING WITH ONLY ONE IMAGE, WITHOUT THE DATASET FOLDER, COMMENT THE LIGHT ABOVE AND UNCOMMENT LINE BELOW. ALSO COMMENT THE LINE BELOW image = testset.pull_image(img_id)
#image = cv2.imread('/home/emeka/Downloads/pp.jpeg', cv2.IMREAD_COLOR) # uncomment if dataset not downloaded
from matplotlib import pyplot as plt
from data import VOCDetection, VOC_ROOT, VOCAnnotationTransform, KITTIDetection, KITTI_ROOT, KITTIAnnotationTransform
# here we specify year (07 or 12) and dataset ('test', 'val', 'train')
#testset = VOCDetection(VOC_ROOT, [('2007', 'val')], None, VOCAnnotationTransform()) #CHANGED
testset = KITTIDetection(KITTI_ROOT, ['val'], None, KITTIAnnotationTransform())
for img_id in range(0, 150):
image = testset.pull_image(img_id)
rgb_image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# View the sampled input image before transform
#plt.figure(figsize=(10,10))
#plt.imshow(rgb_image)
#plt.show()
x = cv2.resize(image, (300, 300)).astype(np.float32)
#plt.imshow(x)
# plt.show()
x -= (104.0, 117.0, 123.0)
x = x.astype(np.float32)
x = x[:, :, ::-1].copy()
# extract 2d bounding box in 0-based coordinates
self.xmin = float(data[0]) # left
self.xmax = float(data[1]) # right
self.ymin = float(data[2]) # top
self.ymax = float(data[3]) # bottom
self.zmin = float(data[4]) # top
self.zmax = float(data[5]) # bottom
self.box3d = np.array([self.xmin,self.xmax,self.ymin,self.ymax,self.zmin,self.zmax])
if __name__ == "__main__":
datatype = 'val' #'train'
KITTI_CLASSES = ( 'Index0', 'Background','Cyclist', 'Car', 'Pedestrian')
class_to_ind = dict(zip(KITTI_CLASSES, range(len(KITTI_CLASSES))))
# Load the Dataset
testset = KITTIDetection(KITTI_ROOT,[datatype], None, KITTIAnnotationTransform)
# Necessary Directories
root_dir="/home/dllab/kitti_object/data_object_image_2"
pcd_dir="/home/dllab/kitti_object/data_object_velodyne/pcl"
segmented_pcd_dir="/home/emeka/Schreibtisch/AIS/ais3d/PCD_Files1"
data_set = "training"
images_dir = os.path.join(root_dir,data_set, "image_{0}".format(2))
detection_dir = '/home/emeka/Schreibtisch/AIS/ais3d/Detections'
calib_dir = '/home/dllab/kitti_object/data_object_velodyne/data_object_calib/training/calib'
box_dir = '/home/emeka/Schreibtisch/AIS/ais3d/bbox_labels_new'
label_dir = os.path.join(root_dir,data_set, "label_{0}".format(2))
# Assign 1 to visualize the images and PCD
show_images=0
# Iterate for every image
if torch.cuda.is_available():
torch.set_default_tensor_type('torch.cuda.FloatTensor')
import time
from ssd import build_ssd
net = build_ssd('test', 300, 4) # initialize SSD
#net.load_weights('../weights/ssd300_mAP_77.43_v2.pth')
#net.load_weights('../weights/VOC.pth')
net.load_weights('../weights/ssd300_Resz_KITTI_80000.pth')
#FOR TESTING WITH ONLY ONE IMAGE, WITHOUT THE DATASET FOLDER, COMMENT THE LIGHT ABOVE AND UNCOMMENT LINE BELOW. ALSO COMMENT THE LINE BELOW image = testset.pull_image(img_id)
#image = cv2.imread('/home/emeka/Downloads/pp.jpeg', cv2.IMREAD_COLOR) # uncomment if dataset not downloaded
from matplotlib import pyplot as plt
from data import VOCDetection, VOC_ROOT, VOCAnnotationTransform, KITTIDetection, KITTI_ROOT, KITTIAnnotationTransform
# here we specify year (07 or 12) and dataset ('test', 'val', 'train')
#testset = VOCDetection(VOC_ROOT, [('2007', 'val')], None, VOCAnnotationTransform()) #CHANGED
testset = KITTIDetection(KITTI_ROOT,['train'], None, KITTIAnnotationTransform())
save_folder = '/home/emeka/Schreibtisch/AIS/ais3d/Detections/'
for img_id in range(0,len(testset)):
image = testset.pull_image(img_id)
rgb_image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# View the sampled input image before transform
#plt.figure(figsize=(10,10))
#plt.imshow(rgb_image)
#plt.show()
x = cv2.resize(image, (300, 300)).astype(np.float32)
#plt.imshow(x)
# plt.show()
x -= (104.0, 117.0, 123.0)
x = x.astype(np.float32)
x = x[:, :, ::-1].copy()
self.xmax = float(data[1]) # right
self.ymin = float(data[2]) # top
self.ymax = float(data[3]) # bottom
self.zmin = float(data[4]) # top
self.zmax = float(data[5]) # bottom
self.box3d = np.array(
[self.xmin, self.xmax, self.ymin, self.ymax, self.zmin, self.zmax])
if __name__ == "__main__":
datatype = 'val' #'train' #val
KITTI_CLASSES = ('Index0', 'Background', 'Cyclist', 'Car', 'Pedestrian')
class_to_ind = dict(zip(KITTI_CLASSES, range(len(KITTI_CLASSES))))
# Load the Dataset
testset = KITTIDetection(KITTI_ROOT, [datatype], None,
KITTIAnnotationTransform)
# Necessary Directories
root_dir = "/home/dllab/kitti_object/data_object_image_2"
pcd_dir = "/home/dllab/kitti_object/data_object_velodyne/pcl"
segmented_pcd_dir = "/home/emeka/Schreibtisch/AIS/ais3d/PCD_Files1"
data_set = "training"
images_dir = os.path.join(root_dir, data_set, "image_{0}".format(2))
detection_dir = '/home/emeka/Schreibtisch/AIS/ais3d/Detections'
calib_dir = '/home/dllab/kitti_object/data_object_velodyne/data_object_calib/training/calib'
box_dir = '/home/emeka/Schreibtisch/AIS/ais3d/bbox_labels_new'
points_dir = "/home/emeka/Schreibtisch/AIS/ais3d/PCD_Files2/DetectionLocations"
label_dir = os.path.join(root_dir, data_set, "label_{0}".format(2))
# Assign 1 to visualize the images and PCD
show_images = 1