sanqingqu / tongji_sanitation_car

Tongji sanitation car projet's visual perception code.

Tongji sanitation-car project's visual perception repo

This repo is used to maintain the tongji sanitation-car project's visual perception codes.

Env:

Ros melodic + Ubuntu 18.04 + Python 2.7

Due to the requirements of the Ros, Python 2.7 is required.

Use WR_LiDAR

you should first configure your local static network ip address to make sure your device and the wr_lidar are in the same network segment. (The lidar's default ip is 192.168.0.10)

run the wr_lidar:

roslaunch wr_ls1207de_udp_with_1_lidar.launch

Use Camera

it evolves changing the permission of accessing the cameras, see http://wiki.ros.org/libuvc_camera for details.

read fisheye camera image:

rosrun libuvc_camera camera_node _width:=1920 _height:=1080 _frame_rate:=30 _video_mode:=mjpeg # should not change the args

calibrate it

rosrun camera_calibration camera_calibrator_tj.py # will print params at end

show undistorted

rosrun camera_calibration camera_undistorted_tj_lower.py # should be modified by infomation printed by camera_calibrator_tj.py

Use Image and Lidar Fusion

The lidar_camera_fusion_detection package evolves three scripts.

extract camera_input img.

rosrun camera_lidar_fusion_detectn img_extract.py

extract wr_lidar laser_scan msg.

rosrun camera_lidar_fusion_detectn laser_point_extract.py

fusion wr_lidar and input_undistorted img

rosrun camera_lidar_fusion_detection img_laser_fusion.py --debug_mode # use --help to see more detail configurations.