MIT NEET Fall 2018

Multiplex SSH

This is highly recommended for connecting to the drones:

Create or edit a file in the .ssh/ directory inside your home directory, called config:

$ nano ~/.ssh/config or $ gedit ~/.ssh/configor $ vi ~/.ssh/config (or whichever text editor you are comfortable with)

You will need to add two lines to the file, and you will need to replace PATH_TO_HOME with the correct path for your platform:

• on linux, /home/YOURTEAMNAME/.ssh/ (or your username, if you are not using a team laptop)
• on mac, /Users/YOURUSERNAME/.ssh
• windows, (TODO)

Add the following two lines (Remember: change PATH_TO_HOME as specified above):

ControlMaster auto
ControlPath PATH_TO_HOME/.ssh/ssh_mux_%h_%p_%r

Important - remember to start your FIRST SSH connection with -Y if you plan to use xforwarding (e.g., for rqt_image_view)

Directory Setup

These instructions replace this section on the website.

On the SSH window on your team laptop, enter the following commands

cd ~
cd ~/bwsi-uav/catkin_ws/src
git clone https://github.com/BWSI-UAV/aero_control.git
cd aero_control
git remote add upstream https://github.com/BWSI-UAV/aero_control.git 
cd ~/bwsi-uav
git clone https://github.com/BWSI-UAV/laboratory.git
cd laboratory
git remote add upstream https://github.com/BWSI-UAV/laboratory.git 
cd ~/bwsi-uav
git clone https://github.com/BWSI-UAV/documents.git
cd documents
git remote add upstream https://github.com/BWSI-UAV/documents.git 

Compressed Camera Feeds

To find if your drone supports compressed camera feeds:

1. Start roscore
2. Start optical flow: sudo -E ~/bwsi-uav/catkin-ws/src/aero-optical-flow/build/aero-optical-flow
3. $ rostopic list | grep compressed

If you don't see /aero_downward_camera/image/compressed in the results you will need to install compressed transport support:

sudo apt-get install ros-kinetic-image-transport-plugins

then restart your camera feed by restarting the aero-optical-flow binary (step 1 above).

To record a compressed downward camera feed:

$ cd ~/rosbags/ # or wherever you want to store your rosbag
$ time rosbag record -O downward /aero_downward_camera/image/compressed # -O specifies the filename

You can then SCP your rosbag to your team laptop.

To convert compressed camera messages to OpenCV images, you can't use CVBridge. Here is an OpenCV-specific decoding solution. (You could also use CompressedImage from sensor_msgs.msg):

from __future__ import print_function
import cv2
import numpy as np
import roslib
import rospy

from sensor_msgs.msg import CompressedImage
# We do not use cv_bridge since it does not support CompressedImage
# from cv_bridge import CvBridge, CvBridgeError

import rosbag
import os

DEST = "/path/to/folder/to/save/images"
BAG  = "/path/to/rosbag.bag"
#your camera topic:
CAM = '/aero_downward_camera/image/compressed'

def bag2msgs():
    bag = rosbag.Bag(BAG)
    if bag is None:
        raise ValueError("no bag {}".format(BAG))
    msgs = []
    for topic, msg, t in bag.read_messages(topics=[CAM]):
        msgs.append(msg)
    bag.close()
    print("MESSAGES: {}".format(len(msgs)))
    return msgs

def uncompress(msgs):
    imgs = []
    for msg in msgs:
         #### direct conversion to CV2 ####
         np_arr = np.fromstring(msg.data, np.uint8)
         image_np = cv2.imdecode(np_arr, cv2.IMREAD_COLOR) # OpenCV >= 3.0:
         imgs.append(image_np)
    return imgs


if __name__ == "__main__":

    if os.listdir(DEST) != []:
        raise ValueError('need empty directory for dest {}'.format(DEST))

    msgs = bag2msgs()
    imgs = uncompress(msgs)

    for idx,im in enumerate(imgs):
        if idx % 50 == 0:
            print(idx)
        imname = "frame{:05d}.jpg".format(idx)
        cv2.imwrite(DEST + imname, im)