Drone Traffic Management Project

Dynamic Traffic Management and Vehicle Track and count using Drones

1. Connecting to DJI Tello Drone:

1. Place the batteries inside the drone and press the switch on button on the drone placed on the left side.
2. wait for a second untill the front LED indicator on the drone starts blinking in yellowish color.
3. Locate the Tello Wifi Network in your Wifi List and then connect to it.

2. To see it in action:

1. Open Two Command prompts in the working directory (or just open cmd and then direct it to the folder using cd command)
2. In the first command prompt: type python tello-command.py and hit enter, this will initialise the tello drone. (Your DJI tello drones must be connceted for this step. refer to 1. Connecting to DJI Tello Drone).
3. Type command and hit enter, this will tell the drone to operate in a command prompt way.

4. after which you should see a response: ok.
5. after this you can tell your drone to takeoff etc, simply type takeoff and hit enter.
6. Type streamon and hit enter, this will switch on the camera stream for the next process to follow.
7. Now in the second command prompt copy and paste this command:

python detect_video.py --weights ./checkpoints/yolov4-tiny-416 --size 416 --model yolov4 --video 0 --output ./data/video/output.avi --count --tiny

8. This will run the detect_video.py file and instruct it to use the camera stream as an input to detect the cars and then save the output in an "output.avi" file in the data/video folder.

If you see some red color errors after running the detect_video.py don't worry they just occur because it takes time to listen to the video stream of the drone and until then the program can't find any frame/image to run the detection onto, it takes about 3-4 seconds in the beginning.

3. Sample Videos

Sample 1: Simulated video

data_video_wow.mp4

Sample 2: Real Working

data_video_test2out.mp4