46,498 Images - Vehicle Damage Images Collection Data

multiple vehicle types

multiple outdoor scenes

multiple types of vehicle damage

multiple collecting angles

different photographic distances

different resolutions

46,498 Images - Vehicle Damage Images Collection Data. The dataset diversity includes multiple vehicle types, multiple outdoor scenes, multiple types of vehicle damage, multiple collecting angles, different photographic distances, and different resolutions. The types of vehicle damage include bump, scratch, paint loss and other vehicle damage. The locations of vehicle damage include the front hood, left and right headlights, door, body and trunk of the vehicle. This dataset can be used for tasks such as automatic vehicle damage detection.

This is a paid datasets for commercial use, research purpose and more. Licensed ready made datasets help jump-start AI projects.

Specifications

Data size

46,498 images, each image contains only one damaged car

Collecting environment

outdoor scenes (including street intersections, urban and rural roads, urban traffic crossings, etc.)

Data diversity

including multiple vehicle types, multiple outdoor scenes, multiple types of vehicle damage, multiple collecting angles, different photographic distances, and different resolutions

Device

cellphone

Collecting angles

eye-level angle, looking down angle

Collecting Conditions

collecting location: China; collecting time: Daytime; collecting weather: Sunny day; collecting season: Autumn

Vehicle Type Distribution

including car, SUV, MPV, minibus, small trucks, big trucks, etc.

Vehicle Damage Distribution

types of vehicle damage: including bump, scratch, paint loss and other vehicle damage; locations of vehicle damage: including the front hood, left and right headlights, door, body and trunk of the vehicle

Data format

the image data format is .jpg or .png, the annotation file format is .metadata

Collecting content

collecting the images of damaged part of vehicle

Annotation content

collecting location, scenes, season, weather, time, device, image data format and image resolution were labeled in the metadata

Accuracy

the accuracy of labels of collecting location, scenes, season, weather, time, device, image data format, image resolution is not less than 97%

Recommended Dataset

89,747 Images Vehicle Attributes Annotation Data

89,747 Images Vehicle Attributes Annotation Data. The data includes outdoor roads (highway, crossroad) scenes. The data covers multiple vehicle types, multiple vehicle colors, multiple license plate colors, multiple vehicle brands, different time, different vehicle orientations. For a vehicle, 2 rectangular bounding boxes and 5 labels were annotated, the rectangular bounding boxes include the front end or rear end of vehicle, and the whole vehicle; the labels included vehicle color, vehicle type, vehicle brand, vehicle orientation and shooting time; For a vehicle plate, a rectangular bounding box of license plate, the number of license plate and the color of license plate were annotated. This data can be used for tasks such as vehicle attributes analysis and license plate recognition.

Surveillance camera multiple vehicle types multiple vehicle colors multiple license plate colors multiple vehicle brands different time different vehicle orientations vehicle attributes analysis