SimulaMet-HOST/HockeyRink

本数据集是用于冰球场地关键点检测的数据集，包含了来自瑞典冰球联赛（SHL）的661个高分辨率视频帧（1920*1080）。数据集能够准确检测冰球场上的56个关键点地标，支持不同的摄像机角度和光照条件，并能处理球员遮挡和动态比赛情况。该数据集适用于相机校准、同态变换估计、2D/3D场景映射、球员跟踪分析、广播覆盖生成、游戏分析和统计数据以及AR/VR应用。

This dataset is for ice hockey rink keypoint detection, containing 661 high-resolution video frames (1920*1080) from the Swedish Hockey League (SHL). The dataset accurately detects 56 keypoint landmarks on hockey rinks, supports various camera angles and lighting conditions, and handles player occlusions and dynamic game situations. It is suitable for camera calibration, homography estimation, 2D/3D scene mapping, player tracking and analysis, broadcast overlay generation, game analytics and statistics, and AR/VR applications.