Description:

This dataset was meticulously curated using dashcam footage collected from various driving sessions. It primarily focuses on detecting and identifying a range of vehicle-related signs, vehicles, and pedestrians in diverse road conditions. The dataset is regularly updated to include more classes and further refine the existing categories. It serves as a valuable resource for developing machine learning models aimed at improving autonomous driving, traffic management systems, and advanced driver assistance systems (ADAS).

Download Dataset

Dataset Composition:

The dataset consists of thousands of labeled images, drawn from hours of footage captured by a dashcam installed in the vehicle. The footage spans various driving scenarios, including city streets, highways, and suburban roads. The data includes images captured under different weather and lighting conditions, including clear daylight, nighttime, overcast skies, and rainy weather.

Categories: The dataset covers the following classes:

Vehicles: Cars, bikes, trucks, buses, and more.

Pedestrians: Individuals crossing streets, walking on sidewalks, etc.

Traffic Lights: Labeled images of red, yellow, and green traffic lights in various locations and positions.

Road Signs: A wide array of road signs including yield signs, stop signs, speed limits, roundabouts, and other traffic control signals.

This dataset is sourced from Kaggle.