BevFusion-NerIPS-2023记录1：nusences数据集

最新推荐文章于 2024-08-31 12:50:18 发布

Ms_WangH

最新推荐文章于 2024-08-31 12:50:18 发布

阅读量350

点赞数 2

文章标签： BEV

本文链接：https://blog.csdn.net/qq_26416195/article/details/136716896

版权

本文详细描述了如何使用mmdetection3d初始化nuscenes数据集，包括所需的完整目录结构、样本文件夹和json文件的分布，以及数据集划分（训练集、验证集和测试集）。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

mmdetection3d在执行下列语句初始化nusences数据集时需要有nuscenes完整目录，现记录：

nuscenes完整目录

data/nuscenes/
├── maps ## 包含nusences数据集中train/test部分
		├── 36092f0b03a857c6a3403e25b4b7aab3.png
		├── 37819e65e09e5547b8a3ceaefba56bb2.png
		├── 53992ee3023e5494b90c316c183be829.png
		├── 93406b464a165eaba6d9de76ca09f5da.png
		├── basemap
		├── expansion
		└── prediction
├── samples  
		## 包含nusences数据集中train/test部分, 包括10个train压缩包和1个test包解压后全部samples，下列12个文件夹
		├── CAM_BACK
		├── CAM_BACK_LEFT
		├── CAM_BACK_RIGHT
		├── CAM_FRONT
		├── CAM_FRONT_LEFT
		├── CAM_FRONT_RIGHT
		├── LIDAR_TOP
		├── RADAR_BACK_LEFT
		├── RADAR_BACK_RIGHT
		├── RADAR_FRONT
		├── RADAR_FRONT_LEFT
		└── RADAR_FRONT_RIGHT
├── sweeps
		## 包含nusences数据集中train/test部分, 包括10个train压缩包和1个test包解压后全部sweeps，下列12个文件夹
		├── CAM_BACK
		├── CAM_BACK_LEFT
		├── CAM_BACK_RIGHT
		├── CAM_FRONT
		├── CAM_FRONT_LEFT
		├── CAM_FRONT_RIGHT
		├── LIDAR_TOP
		├── RADAR_BACK_LEFT
		├── RADAR_BACK_RIGHT
		├── RADAR_FRONT
		├── RADAR_FRONT_LEFT
		└── RADAR_FRONT_RIGHT
├── v1.0-test
		## 包含json文件13个
		├── attribute.json
		├── calibrated_sensor.json
		├── category.json
		├── ego_pose.json
		├── instance.json
		├── log.json
		├── map.json
		├── sample_annotation.json
		├── sample_data.json
		├── sample.json
		├── scene.json
		├── sensor.json
		└── visibility.json
└── v1.0-trainval
	  ## 包含json文件13个
	  	├── attribute.json
		├── calibrated_sensor.json
		├── category.json
		├── ego_pose.json
		├── instance.json
		├── log.json
		├── map.json
		├── sample_annotation.json
		├── sample_data.json
		├── sample.json
		├── scene.json
		├── sensor.json
		└── visibility.json

$ python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./data/nuscenes --extra-tag nuscenes
======
Loading NuScenes tables for version v1.0-trainval...
23 category,
8 attribute,
4 visibility,
64386 instance,
12 sensor,
10200 calibrated_sensor,
2631083 ego_pose,
68 log,
850 scene,
34149 sample,
2631083 sample_data,
1166187 sample_annotation,
4 map,
Done loading in 55.352 seconds.
======
Reverse indexing ...
Done reverse indexing in 6.1 seconds.
======
total scene num: 850
exist scene num: 850
train scene: 700, val scene: 150
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 34149/34149, 17.4 task/s, elapsed: 1959s, ETA:     0s
train sample: 28130, val sample: 6019
======
Loading NuScenes tables for version v1.0-trainval...
23 category,
8 attribute,
4 visibility,
64386 instance,
12 sensor,
10200 calibrated_sensor,
2631083 ego_pose,
68 log,
850 scene,
34149 sample,
2631083 sample_data,
1166187 sample_annotation,
4 map,
Done loading in 64.106 seconds.
======
Reverse indexing ...
Done reverse indexing in 5.8 seconds.
======
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 28130/28130, 3.8 task/s, elapsed: 7381s, ETA:     0s
======
Loading NuScenes tables for version v1.0-trainval...
23 category,
8 attribute,
4 visibility,
64386 instance,
12 sensor,
10200 calibrated_sensor,
2631083 ego_pose,
68 log,
850 scene,
34149 sample,
2631083 sample_data,
1166187 sample_annotation,
4 map,
Done loading in 60.241 seconds.
======
Reverse indexing ...
Done reverse indexing in 6.0 seconds.
======
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 6019/6019, 3.7 task/s, elapsed: 1607s, ETA:     0s
Create GT Database of NuScenesDataset
noise setting:
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 28130/28130, 1.6 task/s, elapsed: 17780s, ETA:     0s
load 65262 truck database infos
load 161928 pedestrian database infos
load 339949 car database infos
load 2120 movable_object.debris database infos
load 62964 traffic_cone database infos
load 8846 motorcycle database infos
load 2259 static_object.bicycle_rack database infos
load 19195 movable_object.pushable_pullable database infos
load 11 vehicle.emergency.ambulance database infos
load 11050 construction_vehicle database infos
load 19202 trailer database infos
load 107507 barrier database infos
load 8185 bicycle database infos
load 12286 bus database infos
load 498 vehicle.emergency.police database infos
load 751 human.pedestrian.stroller database infos
load 619 animal database infos
load 492 human.pedestrian.wheelchair database infos
load 352 human.pedestrian.personal_mobility database infos
======
Loading NuScenes tables for version v1.0-test...
23 category,
8 attribute,
4 visibility,
0 instance,
12 sensor,
1800 calibrated_sensor,
462901 ego_pose,
15 log,
150 scene,
6008 sample,
462901 sample_data,
0 sample_annotation,
4 map,
Done loading in 30.217 seconds.
======
Reverse indexing ...
Done reverse indexing in 0.7 seconds.
======
total scene num: 150
exist scene num: 150
test scene: 150
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 6008/6008, 438.8 task/s, elapsed: 14s, ETA:     0s
test sample: 6008

运行后，nusences数据集目录示例：

data/nuscenes/
├── maps
├── nuscenes_dbinfos_train.pkl
├── nuscenes_gt_database
	## 好多.bin文件
├── nuscenes_infos_test.pkl
├── nuscenes_infos_train.coco.json
├── nuscenes_infos_train.pkl
├── nuscenes_infos_val.coco.json
├── nuscenes_infos_val.pkl
├── samples
├── sweeps
├── v1.0-test
└── v1.0-trainval