def __init__(self, args): self.args = args use_cuda = bool(strtobool(self.args.use_cuda)) #use_cuda=False if args.display: cv2.namedWindow("test", cv2.WINDOW_NORMAL) cv2.resizeWindow("test", args.display_width, args.display_height) self.vdo = cv2.VideoCapture() self.detectron2 = Detectron2() self.deepsort = DeepSort(args.deepsort_checkpoint, use_cuda=use_cuda)
def __init__(self, args): self.args = args use_cuda = bool(strtobool(self.args.use_cuda)) self.detectron2 = Detectron2(self.args.detectron_cfg, self.args.detectron_ckpt) if self.args.deep_sort: self.deepsort = DeepSort(args.deepsort_checkpoint, use_cuda=use_cuda)
def __init__(self, args): self.args = args use_cuda = bool(strtobool(self.args.use_cuda)) self.vdo = cv2.VideoCapture() self.detectron2 = Detectron2() # Initialize coordinate mapper myCoordMapper = coord_mapper.CoordMapper(coord_mapper.ISSIA_kozep_elorol) self.deepsort = DeepSort(args.deepsort_checkpoint, lambdaParam=1.0, coordMapper=myCoordMapper, max_dist=1.0, min_confidence=0.1, nms_max_overlap=0.7, max_iou_distance=0.7, max_age=75, n_init=3, nn_budget=50, use_cuda=use_cuda)
def __init__(self, args): self.args = args use_cuda = bool(strtobool(self.args.use_cuda)) #self.vdo = cv2.VideoCapture() self.imgList = natsort.natsorted(glob.glob(self.args.imgs_path)) self.detectron2 = Detectron2() # Initialize coordinate mapper self.myCoordMapper = coord_mapper.CoordMapperCSG( match_code='HUN-BEL 2. Half') self.fps = 6 self.deepsort = DeepSort(args.deepsort_checkpoint, lambdaParam=0.6, coordMapper=self.myCoordMapper, max_dist=1.0, min_confidence=0.1, nms_max_overlap=0.7, max_iou_distance=0.7, max_age=self.fps * 3, n_init=3, nn_budget=50, use_cuda=use_cuda)
def __init__(self): self.use_cuda = True self.display = True self.config = yaml.load(open('config.yaml', 'r')) self.dataset = self.config['DATASET']['NAME'] self.detectron2 = Detectron2() if (self.dataset == 'kitti'): self.rgb_path = self.config['DATASET']['KITTI']['DATA_PATH'] self.sequence_name = self.config['DATASET']['KITTI'][ 'SEQUENCE_NAME'] self.sequence_list = os.listdir(self.rgb_path) self.velo2cam = np.array( self.config['DATASET']['KITTI']['TRANSFORMS']['Velo2cam']) self.kitti_timestamps = open( self.config['DATASET']['KITTI']['TIMESTAMPS']).readlines() self.kitti_odom = open( self.config['DATASET']['KITTI']['ODOM_PATH']).readlines() self.matches = sorted(self.sequence_list) self.max_vel = self.config['DATASET']['KITTI']['MAX_VEL'] self.max_iou = self.config['DATASET']['KITTI']['MAX_IOU'] self.max_depth = self.config['DATASET']['KITTI']['MAX_DEPTH'] self.min_depth = self.config['DATASET']['KITTI']['MIN_DEPTH'] self.mask_points = self.config['DATASET']['KITTI']['MASK_POINTS'] self.fx = self.config['DATASET']['KITTI']['CAMERA'][ 'focal_length_x'] self.fy = self.config['DATASET']['KITTI']['CAMERA'][ 'focal_length_y'] self.cx = self.config['DATASET']['KITTI']['CAMERA'][ 'optical_center_x'] self.cy = self.config['DATASET']['KITTI']['CAMERA'][ 'optical_center_y'] elif (self.dataset == 'tum'): self.sequence_name = self.config['DATASET']['TUM']['SEQUENCE_NAME'] self.rgb_path = self.config['DATASET']['TUM']['RGB_PATH'] self.depth_path = self.config['DATASET']['TUM']['DEPTH_PATH'] self.odom_path = self.config['DATASET']['TUM']['ODOM_PATH'] self.first_list = read_file_list(self.rgb_path + '.txt') self.second_list = read_file_list(self.depth_path + '.txt') self.third_list = read_file_list(self.odom_path + '.txt') self.matches = associate(self.first_list, self.second_list, self.third_list, 0.0, 0.02) self.max_vel = self.config['DATASET']['TUM']['MAX_VEL'] self.max_iou = self.config['DATASET']['TUM']['MAX_IOU'] self.max_depth = self.config['DATASET']['TUM']['MAX_DEPTH'] self.min_depth = self.config['DATASET']['TUM']['MIN_DEPTH'] self.depth_factor = self.config['DATASET']['TUM']['DEPTH_FACTOR'] self.mask_points = self.config['DATASET']['TUM']['MASK_POINTS'] self.fx = self.config['DATASET']['TUM']['CAMERA']['focal_length_x'] self.fy = self.config['DATASET']['TUM']['CAMERA']['focal_length_y'] self.cx = self.config['DATASET']['TUM']['CAMERA'][ 'optical_center_x'] self.cy = self.config['DATASET']['TUM']['CAMERA'][ 'optical_center_y'] self.class_names = self.config['CLASSES']['ALL'] self.rigid = self.config['CLASSES']['RIGID'] self.not_rigid = self.config['CLASSES']['NON_RIGID'] self.save_mask = self.config['SAVE_MASK'] del self.config
label = labels[idx] # Default label else: label = "human" shape_jsons_l.append(_get_shape_j(point, label)) return _get_labelme_template(im, shape_jsons_l, imp, h, w) if __name__ == "__main__": from detectron2_detection import Detectron2 import json w_p = "/nfs/gpu14_datasets/surveillance_weights/visdrone_t1/model_0111599.pth" cfg_p = "/nfs/gpu14_datasets/surveillance_weights/visdrone_t1/test.yaml" det = Detectron2(cfg_path=cfg_p, weights_path=w_p) ### Params to be changed process_freq = 100 root = "/data/client_datasets/idea_forge/videos/29_may/" # vs = ["03April202017_42_05.mp4", "07April202012_40_03.mp4", "10April202019_02_52.mp4", "11April202016_23_55.mp4", "13April202017_22_58.mp4", "14April202009_16_45.mp4"] vs = ["06April202012_06_22.mp4"] vids = [root + i for i in vs] save_root = "/data/client_datasets/idea_forge/annotations/model_preds" for v in vids: print(v) base_name = os.path.basename(v).replace(".mp4", "_") save_f = os.path.join(save_root, base_name) os.mkdir(save_f) cap = cv2.VideoCapture(v)
def __init__(self): self.use_cuda = True self.display = True self.config = yaml.load(open('config.yaml', 'r')) self.dataset = self.config['DATASET']['NAME'] self.detectron2 = Detectron2() # ros self.image_pub = rospy.Publisher("/segmentation_mask", Image_ros, queue_size=10, latch=True) self.pose_sub = rospy.Subscriber("/camera_pose", Image_ros, self.pose_callback) # self.pose_true = False self.pose = np.eye(4) self.bridge = CvBridge() if (self.dataset == 'kitti'): self.sequence_list = os.listdir( self.config['DATASET']['KITTI']['DATA_PATH']) self.velo2cam = np.array( self.config['DATASET']['KITTI']['TRANSFORMS']['Velo2cam']) self.kitti_timestamps = open( self.config['DATASET']['KITTI']['TIMESTAMPS']).readlines() self.kitti_odom = open( self.config['DATASET']['KITTI']['ODOM_PATH']).readlines() self.matches = sorted(self.sequence_list) self.max_vel = self.config['DATASET']['KITTI']['MAX_VEL'] self.fx = self.config['DATASET']['KITTI']['CAMERA'][ 'focal_length_x'] self.fy = self.config['DATASET']['KITTI']['CAMERA'][ 'focal_length_y'] self.cx = self.config['DATASET']['KITTI']['CAMERA'][ 'optical_center_x'] self.cy = self.config['DATASET']['KITTI']['CAMERA'][ 'optical_center_y'] elif (self.dataset == 'tum'): #self.deepsort = DeepSort(args.deepsort_checkpoint, use_cuda=use_cuda) self.first_list = read_file_list( self.config['DATASET']['TUM']['RGB_PATH'] + '.txt') self.second_list = read_file_list( self.config['DATASET']['TUM']['DEPTH_PATH'] + '.txt') self.third_list = read_file_list( self.config['DATASET']['TUM']['ODOM_PATH'] + '.txt') self.matches = associate(self.first_list, self.second_list, self.third_list, 0.0, 0.02) self.max_vel = self.config['DATASET']['TUM']['MAX_VEL'] self.fx = self.config['DATASET']['TUM']['CAMERA']['focal_length_x'] self.fy = self.config['DATASET']['TUM']['CAMERA']['focal_length_y'] self.cx = self.config['DATASET']['TUM']['CAMERA'][ 'optical_center_x'] self.cy = self.config['DATASET']['TUM']['CAMERA'][ 'optical_center_y'] self.class_names = self.config['CLASSES']['ALL'] self.rigid = self.config['CLASSES']['RIGID'] self.not_rigid = self.config['CLASSES']['NON_RIGID']
def __init__(self, args): self.args = args self.vdo = cv2.VideoCapture() self.detectron2 = Detectron2()