Tutorial

This page contains some tutorial tasks that you can solve to geet a better understanding of raillabel and its concepts. Every task is based on the 1_calibration_1.1 scene from OSDaR23. You need to download the dataset and substitute the path to the file as path_to_calibration_json.

You can find the solution here.

import raillabel

path_to_calibration_json = ... # insert path to the downloaded 1_calibration_1.1_labels.json
scene = raillabel.load(path_to_calibration_json)


# Task 1: What is the ID of the first frame in the scene?
first_frame_id = ...

assert first_frame_id == 12


# Task 2: What is the timestamp of the third frame in the scene?
third_frame_timestamp = ...

assert third_frame_timestamp == Decimal("1631441453.499971000")


# Task 3: How many annotations are in the second frame?
second_frame_annotation_number = ...

assert second_frame_annotation_number == 210


# Task 4: How many cuboid annotations are in the fourth frame?
number_of_cuboid_annotations_in_fourth_frame = ...

assert number_of_cuboid_annotations_in_fourth_frame == 15


# Task 5: How many camera sensors are in the scene?
number_of_camera_sensors = ...

assert number_of_camera_sensors == 9


# Task 6: How many unique persons are in the scene?
number_of_persons = ...

assert number_of_persons == 3