KITTI-flow-2015

The flow 2015 benchmark consists of 200 training scenes and 200 test scenes (4 color images per scene, saved in loss less png format). Compared to the stereo 2012 and flow 2012 benchmarks, it comprises dynamic scenes for which the ground truth has been established in a semi-automatic process. Our evaluation server computes the percentage of bad pixels averaged over all ground truth pixels of all 200 test images. For this benchmark, we consider a pixel to be correctly estimated if the disparity or flow end-point error is <3px or <5% (for scene flow this criterion needs to be fulfilled for both disparity maps and the flow map). We require that all methods use the same parameter set for all test pairs. Our development kit provides details about the data format as well as MATLAB / C++ utility functions for reading and writing disparity maps and flow fields. More details can be found in Object Scene Flow for Autonomous Vehicles (CVPR 2015). We equipped a standard station wagon with two high-resolution color and grayscale video cameras. Accurate ground truth is provided by a Velodyne laser scanner and a GPS localization system. Our datsets are captured by driving around the mid-size city of Karlsruhe, in rural areas and on highways. Up to 15 cars and 30 pedestrians are visible per image.

Flow 2015基准测试集（Flow 2015 benchmark）包含200个训练场景与200个测试场景，每个场景配有4张彩色图像，以无损PNG（lossless PNG）格式存储。相较于立体视觉2012（Stereo 2012）与光流2012（Flow 2012）基准测试集，该数据集涵盖动态场景，其真值标签（ground truth）通过半自动流程生成。本评估服务器会对全部200张测试图像的所有真值像素进行平均，计算错误像素（bad pixels）占比。针对该基准测试集，当视差（disparity）或光流端点误差小于3像素（px）或小于5%时，即认为该像素被正确估计；对于场景流（scene flow）任务，则需同时满足视差图（disparity maps）与光流场（flow fields）的该判定标准。我们要求所有参赛方法在所有测试图像对上使用完全一致的参数配置。本项目提供的开发套件（development kit）详细说明了数据格式，并附带用于读写视差图与光流场的MATLAB及C++工具函数。更多细节可参阅论文《面向自动驾驶车辆的目标场景流》（Object Scene Flow for Autonomous Vehicles，CVPR 2015）。我们在一台标准旅行轿车上搭载了两台高分辨率彩色与灰度摄像机。该数据集的高精度真值标签由Velodyne激光扫描仪与GPS定位系统采集得到。本数据集通过在卡尔斯鲁厄中等城市周边、乡村区域及高速公路上行车采集获取，每张图像中最多可同时出现15辆汽车与30名行人。

搜集汇总

背景与挑战

背景概述

KITTI-flow-2015是一个用于自动驾驶领域的光流估计基准数据集，包含200个训练场景和200个测试场景，每个场景提供4张彩色图像。数据集通过配备摄像头、激光扫描仪和GPS的车辆在真实道路环境中采集，专注于动态场景，用于评估光流和场景流算法的性能，误差标准为视差或流端点误差<3像素或<5%。