search

ardymarky / Hadron-640r

Object detection and 3d tracking used thermal cameras and python3
0 stars 0 forks source link

readme

Hadron-640r

Object detection and 3d tracking using thermal cameras and OpenCV

Proces 1 (Timesync):

  1. Plug in GPS to Avermedia board
  2. Wait 30 seconds
  3. Run "sudo python3.8 timesync.py"

Process 2 (Recording Data):

  1. Run "sudo maincamera*.py"

    Commands: "R" - record/stop "C" - reset static mask "Q" - quit

For recording calibration photos:

  1. Run "python3 calibration.py".
  2. Press spacebar to save image.
  3. Cool thermal checkerboard in fridge
  4. Take about 30-40 photos of checkerboard attached to a posterboard at different locations

To correct for lens distortion:

  1. Add captured images to Image Folder
  2. Inside image folder, run "python3 ../camera_calib.py" and wait
  3. Record outputted matrix and dist arrays
  4. Plug matrixs into "maincamera*.py"

LAN Network

Guide 1 Guide 2

To remote into each jetson nano from a laptop, it is recommended to use xrdp and Windows Remote Desktop. Make sure to use a grapical desktop interface like xfce4 (included in first guide). An optional additional step is to disable the "thinclient_drives" mounted drive. This is done by including EnableFuseMount=false in the /etc/xrdp/sesman.ini file.

Beam distance is 62 inches Bracket piece is 55mm Distance between two cameras is 1.5198 meters or 1519.8 millimeters The lens is 76mm along the beam and 16mmn forward of the gnss receiver

Camera 1 (Red power cable): mtx = np.matrix([[1.15337497e+03, 0, 3.18136149e+02], [0.00000000e+00, 1.15271799e+03, 2.59848723e+02], [0.00000000e+00, 0.00000000e+00, 1.00000000e+00]]) dist = np.array([6.28685479e-02, -3.68503422e+00, -1.55064164e-03, -1.41925924e-03, 1.84969914e+01])

Camera 2 (Blue power cable): mtx= np.matrix([[1.19975668e+03, 0.00000000e+00, 2.90733126e+02], [0.00000000e+00, 1.20094533e+03, 2.45528429e+02], [0.00000000e+00, 0.00000000e+00, 1.00000000e+00]]) dist = np.array([[ 1.19948432e-01, -4.15188268e+00, -1.51634226e-03, -3.83780623e-03, 2.28387360e+01]])

Camera 3 (White): mtx= np.matrix([[1.11467677e+03, 0.00000000e+00, 3.25094199e+02], [0.00000000e+00, 1.11898652e+03, 2.57595640e+02], [0.00000000e+00, 0.00000000e+00, 1.00000000e+00]]) dist = np.array([ 2.33737757e-03, -1.88523229e+00, 1.27303159e-03, -2.99893824e-03, 4.57368528e+00])