INSTALL.md

Dynamic-VINS

1. Prerequisites

ROS

sudo apt-get install ros-melodic-cv-bridge ros-melodic-tf ros-melodic-message-filters ros-melodic-image-transport ros-melodic-nav-msgs ros-melodic-visualization-msgs

Ceres-Solver

# CMake
sudo apt-get install cmake
# google-glog + gflags
sudo apt-get install libgoogle-glog-dev libgflags-dev
# BLAS & LAPACK
sudo apt-get install libatlas-base-dev
# Eigen3
sudo apt-get install libeigen3-dev
# SuiteSparse and CXSparse (optional)
sudo apt-get install libsuitesparse-dev

git clone https://ceres-solver.googlesource.com/ceres-solver
cd ceres-solver
git checkout 2.0.0
mkdir ceres-bin
cd ceres-bin
cmake ..
make -j3
sudo make install

Sophus

git clone https://github.com/strasdat/Sophus.git
cd Sophus
git checkout a621ff  #版本回溯

gedit sophus/so2.cpp modify sophus/so2.cpp as

SO2::SO2()
{
  unit_complex_.real(1.0);
  unit_complex_.imag(0.0);
}

build

mkdir build && cd build && cmake .. && sudo make install

2. Prerequisites for object detection

We offer two kinds of device for tests, please follow the instruction for your match device.

2.1. NVIDIA devices

Clone the repository and catkin_make:

cd {YOUR_WORKSPACE}/src
git clone https://github.com/HITSZ-NRSL/Dynamic-VINS.git

# yolo_ros
git clone https://github.com/jianhengLiu/yolo_ros.git
# install python dependencies
sudo apt install ros-melodic-ros-numpy
sudo apt install python3-pip
pip3 install --upgrade pip
# conda create -n dvins python=3.6
# conda activate dvins
pip3 install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu113
# if you are running on xavier, please not directly install torch by pip and refer to `Possible Problems on NVIDIA Jetson AGX Xavier` section
pip3 install -r yolo_ros/requirements.txt


# build
cd ..
catkin_make

If the output frame rate of yolo is lower than your video frame rate may result in bad performance of Dynamic-VINS.

try TensorRT acceleration version of yolo_ros: https://github.com/jianhengLiu/yolo_ros/tree/tensorrt

2.2. HUAWEI Atlas200

1. prequisities

  sudo apt install ros-melodic-image-transport-plugins

2. Clone the repository:

  cd {YOUR_WORKSPACE}/src
  git clone https://github.com/HITSZ-NRSL/Dynamic-VINS.git

  git clone https://github.com/jianhengLiu/compressedimg2img.git

  # yolo_ros
  git clone https://github.com/jianhengLiu/yolo_ros.git
  cd yolo_ros
  git checkout atlas200

2. Download pre-trained model and put it in {YOUR_WORKSPACE}/src/yolo_ros/model directory.

   • 百度网盘： https://pan.baidu.com/s/1m0lapSFk8KG5Z1Jo5T2VFQ 密码: wgs5
   • gitee: https://modelzoo-train-atc.obs.cn-north-4.myhuaweicloud.com/003_Atc_Models/AE/ATC%20Model/Yolov3/yolov3_framework_caffe_aipp_1_batch_1_input_int8_output_FP32.om
     • After you download the model, change name yolov3_framework_caffe_aipp_1_batch_1_input_int8_output_FP32.om to yolov3.om.

3. Allocate more cpu cores for VINS

  sudo npu-smi set -t aicpu-config -i 0 -c 0 -d 2
  sudo reboot

4. build

  cd ../..
  catkin_make