Excellent video! I know this is a 3 y/o video now, but I've been doin a lot with Tello EDU swarms lately. You really nailed the CV capabilities. Nice clean code. Very impressive.
Glad you enjoy it! Sure, I've planned to make video in details on how the installation and training for yolov3 since I'm limited by my GPU at the moment. But the process is very similar.
Hello, what did you do in the code for making the drone follow you? You referenced that, but never showed the implementation in the code. Your explanation or source code would be very helpful. Thanks!
Hello. I have a problem: My code don't open a cv2 window, so I can't see any image from my Tello. I used this code: import socket import cv2 tello_video = cv2.VideoCapture('udp://@0.0.0.0:11111') while True: try: ret,frame = tello_video.read() if ret: cv2.imread(frame) cv2.waitKey(1) except Exception as err: telllo_video.release() cv2.destroyAllWindows() print(err)
@@robotandcode1474 There is no error message, the program simply crashes and nothing happens. I realized that the program sttoped on line: "tello_video = cv2.VideoCapture('udp://@0.0.0.0:11111')" for a while and I also checked that the 'ret' variable is false.
Hello I respect the fact that you try as much as you can to respond to people's comment. My own is, I wrote a OpenCV code for detecting mined and unminned areas for for gold. How do I infuse the code for a Dji phantom to use the code to detect the Mined and unminned areas in the air ? If possible how do I contact connect with you? Thanks.
Hi and thanks for a great video, I'm looking to learn python and wanted to play with Tello, Don't I need the EDU version to do Python? is something missing on the non EDU version that I can do with the EDU version?
Hi, thanks you find it great 😊. Both Tello version can be controlled by SDK over UDP protocol which you can use Python for it. But Tello EDU has more functionality such as swarming.
Level up idea... Once the drone sees your face and classifies you as you, it locks onto your body so you no longer have to be looking at the drone for it to follow you? Then, the drone could be an action camera? Follow you snowboarding down a mountain? Or doing a line of tricks on a skateboard? Dunno how feasible this actually is, but I think I'm gonna have to have a play.... :D
@@robotandcode1474 Not sure if you're interested, but I completed this idea for my undergraduate. Video's not great but: th-cam.com/video/QLc8DUyQ5z8/w-d-xo.html I used Yolov4t and deepsort. It works perfectly!
Hi there! It's a nice job. In the source code, I found only several independent files. I don't know how do you do to send the yolo data to the drone. Is that missing, I am really interested in realize the same things on my own:) Thx a lot!
when i run the"./darknet detector demo training/tello/tello.data training/tello/yolov3-tiny.cfg training/tello/yolov3-tiny_68000.weights" the video stream of tello cannot open . it always shows "Video-stream stopped! " Please help me!!
Developer 1) connect your laptop to Tello’s wifi. 2) send UDP command to IP 192.168.10.1 at port 8889 as in SDK document. dl-cdn.ryzerobotics.com/downloads/Tello/Tello%20SDK%202.0%20User%20Guide.pdf
@@robotandcode1474 Hey! Wonderful video, loved how well you explained everything. Can you please share the source code for linking the face detection with drone controls? Thanks in advance!
Finally got my tello to work for my laptop! The battery really is a pain lol... One question for your face tracking, did you use PID control to make the drone follow you?
Hi Bazli, I have a question, I'm using Yolov3 as my pre train model to train my object detection model, I want to know how do you eliminate or reduce the latency of video streaming while running a model through tello? What I'm facing now, due to big latency and skipping some frames, my model hardly to detect the object and give faulty detection. I'm hoping you could reply my comment thank you..
If you face yolo give false-positive object detection usually thats mean there are not enough trained image for the model. I’ve never skip frames when using yolo, worth to mention yolo required decent gpu to work for real time application. I think when you skip some frame, the image maybe blurry hence trained model is not prepared for that. Maybe..
@@luqmanrasyidi1514 I've tried both of them, yolo-tiny have lower performance than yolo, but yolo need decent gpu to run for good fps (can check on the repo to see the performance graph). For this video I'm using yolov3-tiny. At this time yolov3 is too heavy for my gpu to run.
I still don't understand, how can I give the drones command and check that he actually finished doing them automatically? For example I have takeoff, forward 50, left 50, how can I check that he finished doing every command and send the next command automatically?
For every UDP command send to Tello, Tello would reply either OK/error message. You should check that message to confirm if Tello received the command successfully. This all documented in Tello SDK documentation.
Hi, thank you for the amazing video, I try some of your coding and it was worked for tello-command but do not work on tello-stream.... I already install the opencv, what I do wrong actually? Do I need to run first tello-command and then run the tello-stream, or I just need to run tello-stream only? Please help me, I'm new with python..
Just in time, please watch my latest video - getting started on software development with Tello. I updated the script to be more comprehensive. To summarize, you need to run Tello-command script first, enter SDK command then send “streamon” msg. Then after that run tello-video script. Ensure you have opencv installed.
@@robotandcode1474 you are saying that its not published yet how you can be sure about its not copy version of another apps result you are showing us that you created AI tracking, because there are some app has been introduced using Tello drone tracking
@@ezguliksariyoll Surely someone has developed this kind of application before me. Even though the behavior might similar but the code implementation is from my own design and perspective (some SDK are derived from others repo too such as TelloPy, PyTello, etc.). On my future incoming video, I'll show some part of my code implementation.
Yes, one method that I used to increase my dataset is by recording myself, then split the video frame into a bunch of images. Checkout this repo for the script I made earlier bit.ly/3rWTmNd
@@robotandcode1474 oh that’s smart! I just finished training an hour ago and I think 40 images still works but the feed was kind of lagging. Do you know if increase in dataset would increase the FPS of my video feed? Just curious.
@@carbkafv FPS does not do anything with trained images number. FPS rely on GPU you used and trained model (yolov4/yolov3/yolov3-tiny). For example, if you want to increase FPS for the detection, you can try on yolov3-tiny model.
I would like to ask a question. Is there any propeller you would use other than Tello that are colored? I don't want to end up buying cheap unbalanced propellers.
Hi, the complete file is not ready to publish to public yet, but currently I in progress providing detailed design and architecture for others to replicate the work.
hello. can you help me, when I run tello-stream.py I saw this kind of error: Traceback (most recent call last): File "tello-stream.py", line 4, in import cv2 ModuleNotFoundError: No module named 'cv2'
Hi, you need to install opencv and python-opencv module before run this script. Checkout my snippet for install these module. bitbucket.org/RobotAndCode/workspace/snippets/Eb7dke/easy-install-yolo-darknet-requirements
Hi 👋, first you need to ensure Tello is turned on and broadcast it’s AP, second double check Tello’s AP password. AP is stand for Access Point work similar to WiFi router.
Yes, you can connect with me using the link I share in my TH-cam about page or just search Murtadha Bazli in LinkedIn. Anyway you are in luck, my in progress video coming in few weeks will talk about architecture and details communication with Tello 😊
YOLO Darknet repository I used and referencing is github.com/AlexeyAB/darknet and working very well up till now. I currently making a video with detail explanation for the whole installation and training process. Stay tuned ☺!!
@@robotandcode1474 I love UTM. I want to complete my master degree and phd at University of Malaya. Do you recommend that? Thank you for your response.
Hi. Thanks for watching, some of the clip I thought people may not interested thus speeding it up. Not sure which part do you think I should slow it down?
Excellent video! I know this is a 3 y/o video now, but I've been doin a lot with Tello EDU swarms lately. You really nailed the CV capabilities. Nice clean code. Very impressive.
You're going places dude... great job.
Amazing Video! Looking forward to experimenting with my Tello!
Good to see you Bazli!
Hope you are well.
Wow!!! Great!
Congrats for this. Likes It.
Thanks 😊
Great project and video, keep up the excellent work!
Thanks Mark! 😊
Impressive and enjoyable. Could you also make a brand new video about the installation, and training process of the newest yolov4 object detector?
Glad you enjoy it! Sure, I've planned to make video in details on how the installation and training for yolov3 since I'm limited by my GPU at the moment. But the process is very similar.
@@robotandcode1474 as yolov4 also was released, you might even try on that one.
YOLO installation with CUDA and cuDNN published now at th-cam.com/video/TKcDksx1y9E/w-d-xo.html
@@robotandcode1474 Thanks you so much mate, really appreciated!
Terbaik Bazli. Teknikal kencang vid. Boleh pi jaoh
Hi, great idea , interesting implementation !
Glad you find that interesting 😊. Thanks for watching 👍
play at 0.75 speed, you'll thank me
after i watched it the first time i did just that...because damn that first run was fast! lol but good.
Hello, what did you do in the code for making the drone follow you? You referenced that, but never showed the implementation in the code. Your explanation or source code would be very helpful. Thanks!
My tello drone keeps moving to the left even though I'm not commanding it to do anything. How can I stop this? What python version do you recommend
Awesome work!
Nice work bro
Hello.
I have a problem:
My code don't open a cv2 window, so I can't see any image from my Tello.
I used this code:
import socket
import cv2
tello_video = cv2.VideoCapture('udp://@0.0.0.0:11111')
while True:
try:
ret,frame = tello_video.read()
if ret:
cv2.imread(frame)
cv2.waitKey(1)
except Exception as err:
telllo_video.release()
cv2.destroyAllWindows()
print(err)
I have the same problem
The same ocurred with me
Can we help us?
Hi, 1) can you post what’s the error? 2) which OS did you use to run this script?
@@robotandcode1474 There is no error message, the program simply crashes and nothing happens. I realized that the program sttoped on line: "tello_video = cv2.VideoCapture('udp://@0.0.0.0:11111')" for a while and I also checked that the 'ret' variable is false.
channel malaysia bhai.perhh padu
Terima kasih 😅
Hi, you running the .py on cmd?
Hello I respect the fact that you try as much as you can to respond to people's comment. My own is, I wrote a OpenCV code for detecting mined and unminned areas for for gold. How do I infuse the code for a Dji phantom to use the code to detect the Mined and unminned areas in the air ? If possible how do I contact connect with you? Thanks.
Hi and thanks for a great video, I'm looking to learn python and wanted to play with Tello, Don't I need the EDU version to do Python? is something missing on the non EDU version that I can do with the EDU version?
Hi, thanks you find it great 😊. Both Tello version can be controlled by SDK over UDP protocol which you can use Python for it. But Tello EDU has more functionality such as swarming.
Is that the model Tello Ryze? thanks for the video.
Nice. You releasing the code files for this and turning controls?
can u tell me which software did you use for writing the program?. As you have told nothing about any of this stuff.
Level up idea...
Once the drone sees your face and classifies you as you, it locks onto your body so you no longer have to be looking at the drone for it to follow you?
Then, the drone could be an action camera? Follow you snowboarding down a mountain? Or doing a line of tricks on a skateboard?
Dunno how feasible this actually is, but I think I'm gonna have to have a play.... :D
Brilliant!! Might try it 😊, there are existing app actually that had this implemented like Tello FPV though.
@@robotandcode1474 Could you send me any links to the existing app? I'd like to check it out
@@UrbanAnimalsTeam I haven’t try this app ever, but here is the link play.google.com/store/apps/dev?id=5552407244136644970
@@robotandcode1474 Not sure if you're interested, but I completed this idea for my undergraduate. Video's not great but:
th-cam.com/video/QLc8DUyQ5z8/w-d-xo.html
I used Yolov4t and deepsort. It works perfectly!
Hi there! It's a nice job. In the source code, I found only several independent files. I don't know how do you do to send the yolo data to the drone. Is that missing, I am really interested in realize the same things on my own:) Thx a lot!
I'm trying to control a DJI mavic mini 2 with my linux PC and get the video stream at the same time. Any ideas?
idk but when i command to take off it doesn't take off, just turning on the motor faster and the sound loud
Ryze Tello can be programmed??
Hi there, how can I store the get_height() data as an int that can be used later on in the python script?
I have ryze tello...how to start coding ?
Hey Can u tell me what is the life of tello drone (Means how many month or year I can use it) .I want to buy it now
God bless you, you are doing well.
Thanks 😊
ur work is..woww!!!
Thanks! 😊
when i run the"./darknet detector demo training/tello/tello.data training/tello/yolov3-tiny.cfg training/tello/yolov3-tiny_68000.weights"
the video stream of tello cannot open . it always shows "Video-stream stopped! "
Please help me!!
Not too sure what’s happen, worth to check if the port 11111 is occupied.
How do I connect the yolov3.weights or yolov4.weights to the tello camera? I've been thinking about this for weeks. Please let me know.
Hi, I stream camera frame from Tello to Python script in my computer and then use Yolo Darknet to process it and classify the object recognized
How does one even connect to the dji tello edu drone? I can't seem to connect my laptop to my drone?
Can someone help me?
Developer 1) connect your laptop to Tello’s wifi. 2) send UDP command to IP 192.168.10.1 at port 8889 as in SDK document. dl-cdn.ryzerobotics.com/downloads/Tello/Tello%20SDK%202.0%20User%20Guide.pdf
@@robotandcode1474 您好，此代码可以共享吗？我想试试您发的这个代码。谢谢你
he can follow what velocite?
Very amazing !!!
Thanks 😊
Hi, it’s very impressive to see your work. And I wonder if you can share the whole code with us, it will be very helpful.Thank you a lot!!!!
Hi, thanks for the compliment 😊. The code is not ripe, so can’t publish it yet.
RobotAndCode it’s ok. Still feel nice to see your video. Looking forward to your completed code!!
@@robotandcode1474 Hey! Wonderful video, loved how well you explained everything. Can you please share the source code for linking the face detection with drone controls? Thanks in advance!
@@robotandcode1474 is the code ripe yet? 🙂
contoh program/code nya dimana kak?
amazing brother we should talk about this where can we be in contact
Finally got my tello to work for my laptop! The battery really is a pain lol... One question for your face tracking, did you use PID control to make the drone follow you?
Awesome! 👍. Yes, I do use PID control for this.
Hey what is average life of tello drone. I want to buy it 😁😁😁
Hi Bazli, I have a question, I'm using Yolov3 as my pre train model to train my object detection model, I want to know how do you eliminate or reduce the latency of video streaming while running a model through tello? What I'm facing now, due to big latency and skipping some frames, my model hardly to detect the object and give faulty detection. I'm hoping you could reply my comment thank you..
If you face yolo give false-positive object detection usually thats mean there are not enough trained image for the model.
I’ve never skip frames when using yolo, worth to mention yolo required decent gpu to work for real time application.
I think when you skip some frame, the image maybe blurry hence trained model is not prepared for that. Maybe..
@@robotandcode1474 Thank you for the feedback, can I know do you using Yolo weights or Yolo-tiny weights?
@@luqmanrasyidi1514 I've tried both of them, yolo-tiny have lower performance than yolo, but yolo need decent gpu to run for good fps (can check on the repo to see the performance graph).
For this video I'm using yolov3-tiny. At this time yolov3 is too heavy for my gpu to run.
@@robotandcode1474 ok thank you for sharing
I still don't understand, how can I give the drones command and check that he actually finished doing them automatically? For example I have takeoff, forward 50, left 50, how can I check that he finished doing every command and send the next command automatically?
For every UDP command send to Tello, Tello would reply either OK/error message.
You should check that message to confirm if Tello received the command successfully.
This all documented in Tello SDK documentation.
Hi, thank you for the amazing video, I try some of your coding and it was worked for tello-command but do not work on tello-stream.... I already install the opencv, what I do wrong actually? Do I need to run first tello-command and then run the tello-stream, or I just need to run tello-stream only? Please help me, I'm new with python..
Just in time, please watch my latest video - getting started on software development with Tello. I updated the script to be more comprehensive.
To summarize, you need to run Tello-command script first, enter SDK command then send “streamon” msg.
Then after that run tello-video script. Ensure you have opencv installed.
Ehh robot boy lama xda jumpa brader hahaha
great job. which code are you runnning for full tracking mode?
The full tracking mode code I use in this video is not yet published. Think to keep it private for now until its ready.
@@robotandcode1474 you are saying that its not published yet how you can be sure about its not copy version of another apps result you are showing us that you created AI tracking, because there are some app has been introduced using Tello drone tracking
@@ezguliksariyoll Surely someone has developed this kind of application before me. Even though the behavior might similar but the code implementation is from my own design and perspective (some SDK are derived from others repo too such as TelloPy, PyTello, etc.). On my future incoming video, I'll show some part of my code implementation.
did you use all 500 images of yourself? I can only find 40 or so images of myself haha.. I am thinking how should I increase the size of my dataset
Yes, one method that I used to increase my dataset is by recording myself, then split the video frame into a bunch of images. Checkout this repo for the script I made earlier bit.ly/3rWTmNd
@@robotandcode1474 oh that’s smart! I just finished training an hour ago and I think 40 images still works but the feed was kind of lagging. Do you know if increase in dataset would increase the FPS of my video feed? Just curious.
@@carbkafv FPS does not do anything with trained images number. FPS rely on GPU you used and trained model (yolov4/yolov3/yolov3-tiny). For example, if you want to increase FPS for the detection, you can try on yolov3-tiny model.
@@robotandcode1474 Thanks! I think I was running it with my cpu.
I would like to ask a question. Is there any propeller you would use other than Tello that are colored? I don't want to end up buying cheap unbalanced propellers.
For me I always get official Tello propeller 😅
@@robotandcode1474 I hear you on that and I knew you were the right person to ask. Thank you
can you share a complete file?
Hi, the complete file is not ready to publish to public yet, but currently I in progress providing detailed design and architecture for others to replicate the work.
hello. can you help me, when I run tello-stream.py I saw this kind of error:
Traceback (most recent call last):
File "tello-stream.py", line 4, in
import cv2
ModuleNotFoundError: No module named 'cv2'
Hi, you need to install opencv and python-opencv module before run this script. Checkout my snippet for install these module. bitbucket.org/RobotAndCode/workspace/snippets/Eb7dke/easy-install-yolo-darknet-requirements
Hello
can u help me
my dji tello edu cant connect to pc
whats is problem
Hi 👋, first you need to ensure Tello is turned on and broadcast it’s AP, second double check Tello’s AP password. AP is stand for Access Point work similar to WiFi router.
Hi, my recent upload explained in details on how communication made between Tello to PC, thought you would be interested.
hi, the link to the source code for object recognition don’t work!
Seem the issue coming from BitBucket, perhaps it will resolve soon.
Macam nak buat Al Saya Tak pandai buat Al
Can I get a copy of the coding
:) Check description bitbucket.org/RobotAndCode/tello-ai/src/master/
Thanks .
Nice bro..how can i connect your linkedin, so that i can ask you more on this python coding and the overall achitecture system..
Yes, you can connect with me using the link I share in my TH-cam about page or just search Murtadha Bazli in LinkedIn.
Anyway you are in luck, my in progress video coming in few weeks will talk about architecture and details communication with Tello 😊
How long is the training time? (If i have cuda and opencv)
I use 3k images and train using GTX1050. If I remember it took around 3-4 hours to get decent usable model.
@@robotandcode1474 thanks a lot
Bro it's a great project just please give us the whole code please
guna ubuntu ke? aduh aku tak pandai main coding ni..
Yup, Ubuntu 18. Boleh belajar 😊
Hello, does this process work with windows?
Also the sourcecode for yolo darknet isn't working, please I really need your help, I'm working on a project. Thank you
I've never tried using Windows for the whole compile and training process but, I'm sure the process is similar.
YOLO Darknet repository I used and referencing is github.com/AlexeyAB/darknet and working very well up till now. I currently making a video with detail explanation for the whole installation and training process. Stay tuned ☺!!
I want to buy open air bottom cover. I am from croatia
Hi, you can checkout 3D design link (Thingiverse) I put in the video about the open air bottom cover and print it yourself.
In which university did you study in Malaysia?
Hi, I studied at Universiti Teknologi Malaysia (2014)
@@robotandcode1474
I love UTM.
I want to complete my master degree and phd at University of Malaya.
Do you recommend that?
Thank you for your response.
Sure, I think for master/phd, supervisor expertise for the research topic is more important rather than university.
What is the purpose of speeding the video? Do you wanna show how you did the proyect or you don´t?
Hi. Thanks for watching, some of the clip I thought people may not interested thus speeding it up. Not sure which part do you think I should slow it down?
@@robotandcode1474 The parts where you clic at some parts, its hard to tell what you clic on.
Omar Garcia alright, better screen recording in the future then 😅
？,could you share your code with me，可以发下您的代码吗
Your video is too fast or do you have your Github of your code please?
Check the description please. bitbucket.org/RobotAndCode/tello-ai/src/master/
@@xingsoon3860 Thanks
send me your email I would like to buy the script
your idea is good BUT! your teaching method is too FAST. Can't follow. it's not a STEP-BY-STEP.
Thanks for watching :D. Yup, agreed with you unfortunately a lot of things to cover up and this video intended to show mostly the result.
So hoping next time you will make a video intended for tutorial
SLOW THE F**CK DOWN !!! TOO MUCH COFFE ??
His speed is normal, you're just fucking too slow