ONLY support python 3.6+, pytorch 1.0.0+. Tested in Ubuntu 16.04.

通过更改雷达驱动使驱动使之可以输出通道ring信息，为算法提供支持 使用方法参考rslidar

git clone https://https://github.com/tt-leader/Lidar-detection.git

pip3 install numba pyntcloud pyyaml rospkg pyquaternion protobuf

安装ros_numpy:

git clone https://github.com/eric-wieser/ros_numpy
cd ros_numpy && python setup.py install

LLVM_CONFIG找不到的问题：

sudo apt-get install llvm-8
export LLVM_CONFIG=/usr/bin/llvm-config-8
pip3 install numba

No lapack/blas resources found 问题：

apt-get install gfortran libopenblas-dev liblapack-dev
pip3 install scipy

Follow instructions in spconv to install spconv.

can not find CUDNN问题：添加软连接

sudo ln -s libcudnn.so.7 libcudnn.so

Follow instructions in torch2trt to install torch2trt

apt install ros-melodic-rospy ros-melodic-ros-base ros-melodic-sensor-msgs ros-melodic-jsk-recognition-msgs ros-melodic-visualization-msgs

you need to add following environment variable for numba.cuda, you can add them to ~/.bashrc:

export NUMBAPRO_CUDA_DRIVER=/usr/lib/aarch64-linux-gnu/libcuda.so
export NUMBAPRO_NVVM=/usr/local/cuda/nvvm/lib64/libnvvm.so
export NUMBAPRO_LIBDEVICE=/usr/local/cuda/nvvm/libdevice

1. 进入工程根目录

cd ./Lidar-detection/second

2. 生成tensorrt引擎

python utils/convert2rt.py --model_dir ./weights/pointpillarsBG --bg_filter  

3. 启动ros推理节点

python inference_ros.py --tensorrt  --model_dir ./weights/pointpillarsBG --bg_filter --tensorrt

4. 启动rviz可视化

rviz -d rviz/lidar_detection.rviz

• Kitti lidar box

A kitti lidar box is consist of 7 elements: [x, y, z, w, l, h, rz], see figure.

All training and inference code use kitti box format. So we need to convert other format to KITTI format before training.

• Kitti camera box

A kitti camera box is consist of 7 elements: [x, y, z, l, h, w, ry].