FLVIS-gpu

KO | EN

gitlite — search

#hacktoberfest #react #ai #llm #golang #android #rust #docker #nodejs #windows #kubernetes #openai #ios #open-source

(★ 9)

No description available.

FLVIS-gpu 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# FLVIS-gpu
This repository is a GPU-accelerated version FLVIS. If you want to achieve best performance on your machine, we recommend you to use it without hesitation.
### 1. Prerequisite
The dependency is almost the same as original FLVIS. Besides, you need a cuda version OpenCV. We only test it under OpenCV 4.5.1
### 2. How to Compile
Set the correct path of your OpenCV in ````CMakeLists.txt````:
````
find_package(OpenCV 4 REQUIRED PATHS your_path)
````
### 3. How to Use
Turn on ````use_gpu```` in ````/launch/d435i/sn943222072828_depth.yaml```` to enable fast detector and optical flow tracking.
### 4. Result

|             Jetson Tx2              |       Feature detect        |        LK optical flow tracking       |
|:-----------------------------------:|:---------------------------:|:-------------------------------------:|
|             CPU version             |             15ms            |                 15ms                  |
|             GPU support             |             8ms             |                 11ms                  |

#  FLVIS
## Feedback Loop Based Visual Inertial SLAM

### 1-Video
<a href="https://www.youtube.com/embed/U1GmOICc1Ac" target="_blank"><img src="http://img.youtube.com/vi/ljZWb2x6CRQ/0.jpg" 
alt="cla" width="480" height="300" border="1" /></a>

| EuRoC DataSet MH_05 | Handheld Test in Lab | FlVIS on UAV Platform |
| ---------------------- | ---------------------- |---------------------- |
| <img src="others/mh05_gif.gif" width="250"> | <img src="others/lab_gif.gif" width="250"> | <img src="others/uav_gif.gif" width="250"> |

### 2-Relevent Publication:
Under Review, a pre-print version can be found [here](https://arxiv.org/abs/2007.02250)
### 3-Support Hardware/Dataset:
Intel RealSense D435i Camera 
EuRoC MAV Dataset
### 4-Build The Project
We have tested in the following environment: 
Ubuntu 16.04 + ROS Kinetic 
Ubuntu 18.04 + ROS melodic 
Clone the repository to the catkin work space eg. /catkin_ws/src
````
git clone https://github.com/HKPolyU-UAV/FLVIS.git
````
Install 3rd Part library
````
cd catkin_ws/src/FLVIS/3rdPartLib/
./install3rdPartLib.sh
````
Compile
````
cd ~/catkin_ws
catkin_make
````
### 5-Verification
Support Platform D435i(Depth/Stereo Mode), EuRoC Dataset, KITTI Dataset
#### 5.1 D435i Camera Depth Mode
##### 5.1.1 Use our recorded rosbag 
Download the dataset [Link-melab_sn943222072828.bag](https://drive.google.com/file/d/1kfOkQTt-i-Hd2M0FZa8Dia4_BweE-ttf/view?usp=sharing) to /bag folder 
Decompress the rosbag:
````
rosbag decompress melab_sn943222072828.bag
````
run the following launch files:
````
roslaunch flvis rviz_bag.launch
roslaunch flvis flvis_bag.launch
````
##### 5.1.2 Use your own camera:
Install the [realsense driver](https://github.com/IntelRealSense/librealsense) and its [ros wrapper](https://github.com/IntelRealSense/realsense-ros) 
Boot the d435i camera and echo the camera infomation
````
roslaunch flvis d435i_depth.launch
rostopic echo /camera/infra1/camera_info
````
You will get the camera infomation like:
<img src="others/camera_info.png">
As shown, where the resolution is 640x480 and fx=384.16455078125; fy=384.16455078125; cx=320.2144470214844;cy=238.94403076171875. 
Edit these information in the config yaml file (say: ````/launch/d435i/sn943222072828_depth.yaml````):
```yaml
image_width: 640
image_height: 480
cam0_intrinsics: [384.16455078125, 384.16455078125, 320.2144470214844, 238.94403076171875]#fx fy cx cy
cam0_distortion_coeffs: [0.0, 0.0, 0.0, 0.0]#k1 k2 r1 r2
```
In the launch file "flvis_d435i.launch", make sure "/yamlconfigfile" is point to the edited config file
````
<param name="/yamlconfigfile" type="string" value="$(find flvis)/launch/d435i/sn943222072828_depth.yaml"/>
````
run the following launch files:
````
roslaunch flvis rviz_bag.launch
roslaunch flvis flvis_d435i_depth.launch
````
#### 5.2 D435i Camera Stero Mode
Like what we did in 5.1.2, we need to config the ````/launch/d435i/sn943222072828_stereo.yaml```` 
Note that, by default the two camera share the same intrinsic parameters, and the baseline length is 0.05m:
```yaml
cam0_intrinsics: [384.16455078125, 384.16455078125, 320.2144470214844, 238.94403076171875]#fx fy cx cy
cam0_distortion_coeffs: [0.0, 0.0, 0.0, 0.0]#k1 k2 r1 r2
cam1_intrinsics: [384.16455078125, 384.16455078125, 320.2144470214844, 238.94403076171875]#fx fy cx cy
cam1_distortion_coeffs: [0.0, 0.0, 0.0, 0.0]#k1 k2 r1 r2
T_cam0_cam1:
[ 1.0, 0.0, 0.0, 0.05,
 0.0, 1.0, 0.0, 0.0,
 0.0, 0.0, 1.0, 0.0,
 0.0, 0.0, 0.0, 1.0]

```

#### 5.3 EuRoC MAV Dataset
Download the dataset(say MH_05_difficult) into the bag folder:
````
roscd flvis/bag/
wget http://robotics.ethz.ch/~asl-datasets/ijrr_euroc_mav_dataset/machine_hall/MH_05_difficult/MH_05_difficult.bag
````
Edit the corresponding bag name in flvis_euroc_mav.launch file:
````
<node pkg="rosbag" type="play" name="rosbag" args="$(find flvis)/bag/MH_05_difficult.bag"/>
````
run the following launch files:
````
roslaunch flvis rviz_euroc.launch
roslaunch flvis flvis_euroc_mav.launch
````

#### 5.4 KITTI Dataset
Download the dataset into the bag folder:

Edit the corresponding bag name in flvis_kitti.launch file:
````
<param name="/publish_gt" type="bool" value="true" />
<param name="/dataset_folder_path" type="string" value="$(find flvis)/bag/KITTI/dataset/sequences/00/" />
<param name="/dataset_gt_file" type="string" value="$(find flvis)/bag/KITTI/dataset/poses/00.txt" />
````
run the following launch files:
````
roslaunch flvis rviz_kitti.launch
roslaunch flvis flvis_kitti.launch
````

### Maintainer:
[Shengyang Chen](https://www.polyu.edu.hk/researchgrp/cywen/index.php/en/people/researchstudent.html)(Dept.ME,PolyU): shengyang.chen@connect.polyu.hk 
Yajing Zou(Dept.LSGI,PolyU):rick.zou@connect.polyu.hk

FLVIS-gpu

FLVIS-gpu 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

FLVIS-gpu

FLVIS-gpu 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README