Yolov5_DeepSort_Pytorch_ros

Connect Yolov5 detection module and DeepSort tracking module via ROS

Prepare the environment

• Add a .pth in /usr/lib/python3/dist-packages/
  • User can take environment.sh as reference to vim a deepsort.pth
• Build catkin workspace first time In Yolov5_DeepSort_Pytorch_ros root folder:

catkin_make

Commands

roscore
source devel/setup.bash

roslaunch realseanse camera:

roslaunch realsense2_camera rs_camera.launch color_width:=1280 color_height:=720 color_fps:=30

To rosrun Yolov5 detection node:

rosrun detection_only det_only_v3.py 

To rosrun DeepSort tracking node:

rosrun detection_only track_only_final.py 
rosrun detection_only track_only_final.py --save-img

To rosrun slim version DeepSort tracking node:

rosrun detection_only track_only_slim.py 
rosrun detection_only track_only_slim.py --save-img

Topics

Topics	Publisher	Subscriber
/camera/color/image_raw	realsense_node	detection_node
/det_result	detection_node	tracking_node
/track_result	tracking_node