车辆外观损伤识别

针对常见的小汽车车型，识别车辆外观受损部件及损伤类型，支持32种车辆部件、5大类外观损伤。同时可输出损伤的数值化结果（长宽、面积、部件占比），支持单图多种损伤的识别。

Aimed at common passenger car models, this dataset is designed to identify damaged vehicle exterior components and damage types, supporting recognition for 32 types of vehicle components and 5 major categories of exterior damages. Meanwhile, it can output quantified damage results including length, width, area and component proportion, and supports recognition of multiple damages in a single image.

搜集汇总

背景与挑战

背景概述

该数据集针对常见小汽车车型，识别车辆外观受损部件及损伤类型，支持32种车辆部件和5大类外观损伤，并可输出损伤的数值化结果。数据集发布于2021年1月，更新频度为一年，采用增量更新方式。