ROS-Object-Detection-2Dto3D-RealsenseD435

Source code Version

Object-Detection-and-location-RealsenseD435

Object Dection with SLAM

With SLAM(SC-Lego-LOAM) ：
Perception-of-Autonomous-mobile-robot

Requirements

Ubuntu18.04 OR 16.04
ROS melodic(Has been Tested) Kinetic C++ 11_std At least,I used the C++ 17 std
Eigen3 :in absolutely Path /usr/local/eigen3
PCL:I used ros-melodic's default config PCL(1.8.1)
Intel Realsense SDK >=2.0
TensorRT: Go to Object-Detection-and-location-RealsenseD435 install the requirments of TensorRT.

How to use

git clone https://github.com/Mazhichaoruya/ROS-Object-Detection-2Dto3D-RealsenseD435.git

Next creat your ROS woekspace to us it.

cd ~
mkdir -p catkin_ws/src
cd ..
catkin_make
cd ~/ROS-Object-Detection-2Dto3D-RealsenseD435/
mv realsense_d435/ realsensen-ros/  ~/catkin_ws/src
catkin_make
mv engine/ ~/catkin_ws/src/realsense_d435

Test

roslaunch realsense_d435 run.launch

Then You can see the Pointcloud By Rviz(Add Display By topic,Choose The Pointcloud2)

Example

ROS version on 9-26: The different color Point means different classes,such ad Person,mouse,book,Tv and soon on,We can see the 3D Point in the camera coordinate system. With SLAM：