ROS Lane Detection and Darknet_ros Object Detection

  1. Object Detection
  • Recieve Video from Vehicle Camera(using usb_cam pkg)
  • using Darknet_ros, use networks yoloV3-tiny, for Bus Stop Detection
  • is Detected, Publish 'Speed 0' Command Wait 10sec Vehicle is Drive with lane detection
  1. Lane Detection
  • Recieve Video from usb_cam pkg
  • Video Convert Bird Eyes View Image and Binary
  • using Sliding Windows and Calculate Vehicle's Steer and Publish Steer command
  • Speed is fixed 1000
  • subscribe message like 'Bus Stop Detect!' -> Speed is 0 and Wait 10sec
  • Over 10sec Drive using lane detection

Command

  1. run darknet_ros and usb_cam roslaunch darknet_ros darknet_ros2.launch

  2. run lane detection rosrun lane_detection_example hanho_lane_detector.py

  3. check binary image is right?

  4. binary image is not right -> image shape is not right?

  • move 'control penal' top x,y bottom x,y point
  • this point is BEV matrix point. this change -> BEV image change
  1. binary image is not right -> image binarization is not right?
  • in 'hanho_lane_detector.py' image_process_otsu function code change
  • i think change cv2.inRange(src, lower, upper)
  1. all of image is right
  • run vesc_driver_node
  • this is vesc_node. if it is not running. vehicle don't Drive Anyway roslaunch vesc_driver vesc_driver_node.launch