Intersection Monitoring: A dataset for vehicle detection using an infrastructure camera including ground truth vehicle localization

Intersection Monitoring: A dataset for vehicle detection using an infrastructure camera including ground truth vehicle localization.

This dataset contains images from a infrastructure camera monitoring an intersection and the ground-truth information for a vehicle crossing it from different directions. The data is aimed to develop and improve image-based vehicle detection algorithms.

The dataset includes two different recordings with the same structure. For each of them, the sequence of images is provided in PNG format, along with the ground truth of the vehicle. The ground truth data was captured using a high-precision GNSS receiver installed on the test vehicle, fused with in-vehicle sensors using an Extended Kalman Filter (EKF) .

Time considerations

The camera and GNSS receiver clocks were synchronized before each test to have the same time base.

Each test last about 140 seconds.

Image files

Image files can be found on the img/ folder within for each test. The camera was configured to record images at 25 fps:

Test 1: 3704 files

Test 2: 3426 files

Vehicle localization ground truth

The test vehicle is a prototype of Autonomous Vehicle developed by the AUTOPIA research group at the Centre for Automation and Robotics in Spain.

Vehicle location mainly depends on a Trimble BX982 GNSS receiver using RTK inputs from a local station. However, the location algorithm applies an EKF for combining GNSS measurements with different onboard sensors providing yaw rate, longitudinal acceleration and speed, steering wheel position and speed, etc.

For each test, the vehicle.csv file contains the vehicle information recorded at 20Hz. This file includes:

UTM Time: Time using the format HHMMSSss:

HH: Hours

MM: Minutes

SS: Seconds

ss: Fraction of seconds

UTM East: East coordinate of the GNSS antenna in the UTM frame, in meters.

UTMNorth: North coordinate of the GNSS antenna in the UTM frame, in meters.

Orientation: Yaw angle of the vehicle, measured from the East axis (x-axis).

Speed: Vehicle speed in m/s

Acceleration: Vehicle acceleration in m/s^2

Camera info

The file camera_parameters.json includes all the information about the intrinsic and extrinsic parameters for the camera. The configuration stored on this file appplies to all tests.

The camera used is an AXIS M1125 with variable focal length. It was installed in a communication tower near the intersection.

Vehicle info

The file vehicle_parameters.json includes information about vehicle dimensions and antenna location. The GNSS antenna is installed near the rear axle of the vehicle, in the middle part of the vehicle. The configuration stored on this file appplies to all tests.

Tools

Some MATLAB tools can be found in https://github.com/autopia-car/datasets-tools-intersection-monitoring

Disclaimer

That the dataset comes "AS IS", without express or implied warranty and/or any liability exceeding mandatory statutory obligations. This especially applies to any obligations of care or indemnification in connection with the dataset. The dataset was created for our research purposes only and no quality assessment was done for the usage in products of any kind. We can therefore not guarantee for the correctness, completeness or reliability of the provided data set.