nexar-ai/nexar_collision_prediction
收藏资源简介:
Nexar碰撞预测数据集是由Nexar行车记录仪收集的视频组成的数据集。这些视频的分辨率为1280x720,每秒30帧,通常持续约40秒。数据集包含1500个视频,其中一半展示的是发生碰撞或即将发生碰撞的事件(正面案例),另一半展示的是正常驾驶情况(负面案例)。对于正面案例,提供了事件发生的时间。该数据集的目的是帮助构建能够预测碰撞时间的模型,其中碰撞和险些碰撞都被视为正面案例。
The Nexar Collision Prediction Dataset consists of videos captured from Nexar dashcams. These videos have a resolution of 1280x720 at 30 frames per second and typically last about 40 seconds. The dataset includes 1500 videos, half of which show events where a collision occurred or was imminent (positive cases), and the other half show regular driving (negative cases). The time of the event (collision or near-miss) is provided for positive cases. The goal of this dataset is to assist in building models capable of predicting the time of collision, with both collisions and near-misses treated as positive cases.




