def DetectInit():
global sess, model, mc
detect_net = 'squeezeDet'
checkpoint = '/home/ubuntu/catkin_ws/src/robo_perception/scripts/weights/model.ckpt-99999'
assert detect_net == 'squeezeDet' or detect_net == 'squeezeDet+', \
'Selected nueral net architecture not supported'
tf.Graph().as_default()
# Load model
if detect_net == 'squeezeDet':
mc = kitti_squeezeDet_config()
mc.BATCH_SIZE = 1
# model parameters will be restored from checkpoint
mc.LOAD_PRETRAINED_MODEL = False
model = SqueezeDet(mc, '0')
elif detect_net == 'squeezeDet+':
mc = kitti_squeezeDetPlus_config()
mc.BATCH_SIZE = 1
mc.LOAD_PRETRAINED_MODEL = False
model = SqueezeDetPlus(mc, '0')
saver = tf.train.Saver(model.model_params)
# Use jit xla
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.5)
config = tf.ConfigProto(allow_soft_placement=True, gpu_options=gpu_options)
config.graph_options.optimizer_options.global_jit_level = tf.OptimizerOptions.ON_1
#with tf.Session(config=config) as sess:
sess = tf.Session(config=config)
saver.restore(sess, checkpoint)
points = rs.points()
pipeline= rs.pipeline()
config = rs.config()
config.enable_stream(rs.stream.color, 1280, 720, rs.format.bgr8, 30)
config.enable_stream(rs.stream.depth, 1280, 720, rs.format.z16, 30)
profile = pipeline.start(config)
depth_sensor = profile.get_device().first_depth_sensor()
depth_scale = depth_sensor.get_depth_scale()
align_to = rs.stream.color
align = rs.align(align_to)
###=====================================================================================###
#SSD_detection
# TensorFlow session: grow memory when needed. TF, DO NOT USE ALL MY GPU MEMORY!!!
gpu_options = tf.GPUOptions(allow_growth=True)
config = tf.ConfigProto(log_device_placement=False, gpu_options=gpu_options)
isess = tf.InteractiveSession(config=config)
# Input placeholder.
net_shape = (300, 300)
data_format = 'NHWC'
img_input = tf.placeholder(tf.uint8, shape=(None, None, 3))
# Evaluation pre-processing: resize to SSD net shape.
image_pre, labels_pre, bboxes_pre, bbox_img = ssd_vgg_preprocessing.preprocess_for_eval(
img_input, None, None, net_shape, data_format, resize=ssd_vgg_preprocessing.Resize.WARP_RESIZE)
image_4d = tf.expand_dims(image_pre, 0)
# Define the SSD model.
reuse = True if 'ssd_net' in locals() else None
ssd_net = ssd_vgg_300.SSDNet()
with slim.arg_scope(ssd_net.arg_scope(data_format=data_format)):
def get_session():
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
return tf.Session(config=config)
with tf.gfile.GFile(PATH_TO_CKPT, 'rb') as fid:
serialized_graph = fid.read()
od_graph_def.ParseFromString(serialized_graph)
tf.import_graph_def(od_graph_def, name='')
## Loading label map
# Label maps map indices to category names, so that when our convolution network predicts `5`,
# we know that this corresponds to `airplane`. Here we use internal utility functions,
# but anything that returns a dictionary mapping integers to appropriate string labels would be fine
label_map = label_map_util.load_labelmap(PATH_TO_LABELS)
categories = label_map_util.convert_label_map_to_categories(
label_map, max_num_classes=NUM_CLASSES, use_display_name=True)
category_index = label_map_util.create_category_index(categories)
# Setting the GPU options to use fraction of gpu that has been set
config = tf.ConfigProto()
config.gpu_options.per_process_gpu_memory_fraction = GPU_FRACTION
# Detection
class Detector:
def __init__(self):
self.image_pub = rospy.Publisher("rcnn/debug_image",
Image,
queue_size=1)
self.object_pub = rospy.Publisher("rcnn/objects",
Detection2DArray,
queue_size=1)
# Create a supscriber from topic "image_raw"
self.bridge = CvBridge()
import tf import os #specify the gpu device id os.environ["CUDA_VISIBLE_DEVICES"] = "2" #specify the fraction of all the gpu memory gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.7) sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options)) #specify the allow_growth option gpu_options = tf.GPUOptions(allow_growth=True) sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))