文章目錄

• KITTI數(shù)據(jù)集
• • 數(shù)據(jù)集結(jié)構(gòu)
  • 數(shù)據(jù)集內(nèi)容
  • data_object_calib 樣本標(biāo)定數(shù)據(jù)
  • data_object_label_2 3D點云標(biāo)注文件

KITTI數(shù)據(jù)集

數(shù)據(jù)集結(jié)構(gòu)

• KITTI數(shù)據(jù)集網(wǎng)盤 提取碼：0bjl

KITTI ├── devkit_object | ├── cpp | ├── mapping | ├── matlab | └── readme.txt │—— data_object_calib │ ├── testing │ │ └── calib │ └── training │ └── calib │—— data_object_label_2 │ ├── testing │ └── label_2 │—— data_object_image_2 │ ├── testing │ │ └── image_2 │ └──training │ └── image_2 │—— data_object_velodyne │ ├── testing │ │ └── velodyne │ └── training │ └── velodyne │—— data_road │ ├── testing │ │ ├── calib │ │ └── image_2 │ └── training │ ├── calib │ ├── gt_image_2 │ └── image_2 │—— depth_2

數(shù)據(jù)集內(nèi)容

- data_object_calib 每個樣本的標(biāo)定數(shù)據(jù) （測試集7518個，訓(xùn)練集7481個） - data_object_label_2 3D點云標(biāo)注文件（只有訓(xùn)練集7481個） - data_object_image_2 RGB圖像 （測試集7518個，訓(xùn)練集7481個） - data_object_velodyne 激光雷達(dá)數(shù)據(jù)（測試集7518個，訓(xùn)練集7481個） - data_road 包含測試集和訓(xùn)練集 每集中包含標(biāo)定參數(shù)以及RGB圖像 訓(xùn)練集中還多個gt_image_2 - depth_2 7481張深度圖像

data_object_calib 樣本標(biāo)定數(shù)據(jù)

P0：校對后的 camera0 投影矩陣，3x4 數(shù)組 P1：校對后的 camera1 投影矩陣，3x4 數(shù)組 P2：校對后的 camera2 投影矩陣，3x4 數(shù)組 P3：校對后的 camera3 投影矩陣，3x4 數(shù)組 R0_rect：校準(zhǔn)旋轉(zhuǎn)矩陣，4x4 數(shù)組 Tr_velo_to_cam：從 Velodyne 坐標(biāo)到相機(jī)坐標(biāo)的變換矩陣，4x4 數(shù)組 Tr_imu_to_velo：從 IMU 坐標(biāo)到 Velodyne 坐標(biāo)的變換矩陣，4x4 數(shù)組

data_object_label_2 3D點云標(biāo)注文件

#Values Name Description ----------------------------------------------------------------------------1 type (類別信息:車 貨車 卡車...)Describes the type of object: 'Car', 'Van', 'Truck','Pedestrian', 'Person_sitting', 'Cyclist', 'Tram','Misc' or 'DontCare'1 truncated (是否被截斷:0-1截斷程度）Float from 0 (non-truncated) to 1 (truncated), wheretruncated refers to the object leaving image boundaries1 occluded (是否被遮擋:0完全可見 1部分遮擋 2大部分遮擋 3未知)Integer (0,1,2,3) indicating occlusion state:0 = fully visible, 1 = partly occluded2 = largely occluded, 3 = unknown1 alpha (物體觀察角度:[-pi, pi](弧度))Observation angle of object, ranging [-pi..pi]4 bbox (2D圖像標(biāo)注框:左上角右下角坐標(biāo))2D bounding box of object in the image (0-based index):contains left, top, right, bottom pixel coordinates3 dimensions (3D點云標(biāo)注框:高 寬 長(m))3D object dimensions: height, width, length (in meters)3 location (3D點云標(biāo)注框:相機(jī)坐標(biāo)系底面中心點坐標(biāo)(m))3D object location x,y,z in camera coordinates (in meters)1 rotation_y (物體正前方向與相機(jī)水平方向角度)Rotation ry around Y-axis in camera coordinates [-pi..pi]1 score (得分)Only for results: Float, indicating confidence indetection, needed for p/r curves, higher is better.

總結(jié)

以上是生活随笔為你收集整理的KITTI数据集简析的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。