基于RDK S100的BEV多任务和CenterPoint 3D算法部署流程

应用开发 机器人开发
,
1

1. 环境部署

下载tros安装包，拷贝到RDK S100，解压缩：

mkdir ~/tros
tar -zxvf install-s100.tar.gz -C ~/tros

2. 运行

2.1 BEV算法

cd ~/tros/ 
source install/local_setup.bash  

if [ -L qat ]; then rm qat; fi 
ln -s `ros2 pkg prefix hobot_bev`/lib/qat/ qat 
ln -s ~/hobot_bev_data/nuscenes_bev_val nuscenes_bev_val  

ros2 launch hobot_bev hobot_bev.launch.py

2.2 CenterPoint算法

cd ~/tros/ 
source install/local_setup.bash  

if [ -L qat ]; then rm qat; fi 
ln -s `ros2 pkg prefix hobot_centerpoint`/lib/qat/ qat 
ln -s ~/centerpoint_data centerpoint_data  

ros2 launch hobot_centerpoint hobot_centerpoint.launch.py

3. 展示

启动成功后，打开同一网络电脑的浏览器，访问RDK的IP地址http://IP:8000（IP为RDK的IP地址），即可看到算法可视化的实时效果。

提示：右键->在新标签页中打开视频。

BEV:

Centerpoint:

2

尝试使用BEV和centerpoint程序，报错如下：
[INFO] [hobot_bev-1]: process started with pid [20017]

[INFO] [websocket-2]: process started with pid [20019]

[hobot_bev-1] [UCP]: log level = 3

[hobot_bev-1] [UCP]: UCP version = 3.7.3

[hobot_bev-1] [VP]: log level = 3

[hobot_bev-1] [DNN]: log level = 3

[hobot_bev-1] [HPL]: log level = 3

[hobot_bev-1] [UCPT]: log level = 6

[websocket-2] [WARN] [1750432899.136972325] [websocket]:

[websocket-2] Parameter:

[websocket-2] image_topic: /image_jpeg

[websocket-2] image_type: mjpeg

[websocket-2] only_show_image: 1

[websocket-2] output_fps: 0

[websocket-2] [INFO] [1750432899.137261141] [websocket]: Websocket using image mjpeg

[hobot_bev-1] [INFO] [1750432899.177184222] [bev_node]: BevNode init

[hobot_bev-1] [WARN] [1750432899.177522110] [bev_node]:

[hobot_bev-1] topic_name: image_jpeg

[hobot_bev-1] save_image: false

[hobot_bev-1] config_file: /opt/tros/humble/lib/hobot_bev/config/bev_gkt_mixvargenet_multitask_nuscenes/workflow_la

[hobot_bev-1] glog_level: 1

[hobot_bev-1] [WARN] [1750432899.182547313] [ai_wrapper]:

[hobot_bev-1] Set glog level in cmd line with ‘–glog_level=$num’

[hobot_bev-1] EXAMPLE_SYSTEM = 0, EXAMPLE_REPORT = 1, EXAMPLE_DETAIL = 2, EXAMPLE_DEBUG = 3

[hobot_bev-1] bin --config_file=/opt/tros/humble/lib/hobot_bev/config/bev_gkt_mixvargenet_multitask_nuscenes/workflo

–glog_level=1

[hobot_bev-1] I0000 00:00:00.000000 20017 vlog_is_on.cc:197] RAW: Set VLOG level for “*” to 1

[hobot_bev-1] I0620 23:21:39.182687 20017 ai_wrapper.cpp:85] EXAMPLE_SYSTEM

[hobot_bev-1] I0620 23:21:39.182710 20017 ai_wrapper.cpp:86] EXAMPLE_REPORT

[hobot_bev-1] I0620 23:21:39.182714 20017 ai_wrapper.cpp:92] config_fname: /opt/tros/humble/lib/hobot_bev/config/bev

t_multitask_nuscenes/workflow_latency.json

[hobot_bev-1] [BPU][[BPU_MONITOR]][281473108190720][INFO]BPULib verison(2, 1, 2)[0d3f195]!

[hobot_bev-1] [DNN] HBTL_EXT_DNN log level:6

[hobot_bev-1] [DNN]: 3.7.3_(4.2.11 HBRT)

[hobot_bev-1] malloc(): invalid size (unsorted)

[websocket-2] [ERROR] [1750432904.142624919] [websocket]: Websocket did not receive image data! Please check whether

isher still exists by ‘ros2 topic info /image_jpeg’!

[websocket-2] [ERROR] [1750432909.142438662] [websocket]: Websocket did not receive image data! Please check whether

isher still exists by ‘ros2 topic info /image_jpeg’!

[ERROR] [hobot_bev-1]: process has died [pid 20017, exit code -6, cmd '/opt/tros/humble/lib/hobot_bev/hobot_bev --ro

vel info --ros-args --params-file /tmp/launch_params_bwg_1xfd --params-file /tmp/launch_params_btswo435 --params-fil

arams_f4dvf1ef’].

[websocket-2] [ERROR] [1750432914.142277330] [websocket]: Websocket did not receive image data! Please check whether

isher still exists by ‘ros2 topic info /image_jpeg’!

[websocket-2] [ERROR] [1750432919.142100573] [websocket]: Websocket did not receive image data! Please check whether

isher still exists by ‘ros2 topic info /image_jpeg’!

[websocket-2] [ERROR] [1750432924.141926066] [websocket]: Websocket did not receive i

mage data! Please check whether the image publisher still exists by 'ros2 topic info

/image_jpeg’!