Security Alarm System Project with Ultralytics YOLOv8 Object Detection | Episode 8

[@aryashinde6401]

Thank you very much for this tutorial. It will be really helpful for me in my project. 😀

[@Ultralytics]

Thank you for your feedback! We're glad to hear this :)

[@Melo7ia]

🎶 Nicolai bringing beats with YOLOv8 detecting streets! I'm curious-how does the YOLOv8 tackle false positives in busier environments, like a jazzy club with lots of movement? Would love to riff on those groove challenges and solutions! 🎷

[@Ultralytics]

Great question! 🎷 YOLOv8 uses advanced algorithms to minimize false positives, even in dynamic environments like a jazzy club. It leverages improved architecture and training techniques to enhance accuracy. For busy scenes, fine-tuning the model with specific data and adjusting confidence thresholds can help. Check out our guide for more insights: docs.ultralytics.com/guides/security-alarm-system/

[@furkandemirel9595]

thnk you sir

[@Ultralytics]

Most welcome

[@Sasha-n2x]

Wow, YOLOv8 seems incredible for security systems! 🛡️ Have you encountered any real-world examples where object detection tech was abused or misused, and how can developers prevent such ethical dilemmas while advancing security? #TechWithResponsibility #YOLOv8Ponderings

[@Ultralytics]

Great question! While object detection tech like YOLOv8 offers immense benefits, it can indeed be misused, such as for unauthorized surveillance or privacy invasion. To prevent such ethical dilemmas, developers should adhere to principles of AI ethics, including transparency, accountability, and privacy protection. Always ensure compliance with local laws and regulations, and consider implementing features like data anonymization and user consent mechanisms. For more on AI ethics, check out our AI Ethics guide www.ultralytics.com/glossary/ai-ethics. #TechWithResponsibility 🌐

[@ForReal17720]

I want to do it with 10 classes using the model i trained is it possible?

[@Ultralytics]

Yes, it's possible! The code is available at: docs.ultralytics.com/guides/security-alarm-system/#object-detection-and-alert-sender Thank you!!! Ultralytics Team!

[@AxelRyder-q1b]

Whoa, Nicolai! This is some next-level security wizardry! Can YOLOv8 handle multiple people at once, and if so, how does it differentiate between authorized personnel and intruders? Thinking about sneaky conference rooms snags and office shenanigans-wouldn't wanna trigger alarms every 5 minutes, right? 😂?

[@Ultralytics]

Absolutely, YOLOv8 can handle multiple people simultaneously! To differentiate between authorized personnel and intruders, you can integrate a facial recognition system alongside YOLOv8. This way, the system can identify and verify individuals before triggering any alarms. For more details, check out our guide on integrating YOLOv8 with other systems: docs.ultralytics.com/guides/security-alarm-system/. Happy securing! 🚀

[@March_Awake]

Thank you for sharing

[@Ultralytics]

Thanks for watching!

[@BostonleierPlaysMC]

How would this work if I'm trying to connect to a camera through the IP

[@Ultralytics]

Well, the workflow will be the same, only you will need to replace the video file with the camera stream URL. For more information, you can explore: docs.ultralytics.com/modes/predict/#__tabbed_2_13 Thanks, Ultralytics Team!

[@lovers_GTA]

how we can include tracking also? I mean if I person detects only sending one email

[@Ultralytics]

To track objects, you can utilize the `model.track` function instead of `model.predict`. Subsequently, you can send an email containing the tracked object IDs. For further details on object tracking with Ultralytics YOLOv8, you can refer to docs.ultralytics.com/modes/track/.

[@AnilKumarPatra-f4f]

How to add rtsp , what code change will be needed?

[@Ultralytics]

You can load the stream effortlessly by passing the RTSP URL directly into the VideoCapture function! Thanks :)

[@m033372]

How does the YOLOv8 object detection compare to traditional security systems in terms of real-time accuracy and false positives? Has anyone integrated it into a home security setup yet?

[@Ultralytics]

Great question! YOLOv8 excels in real-time accuracy and significantly reduces false positives compared to traditional security systems, thanks to its advanced object detection algorithms. Many users have successfully integrated YOLOv8 into home security setups, benefiting from its efficiency and reliability. For more details on integration and performance, check out our documentation at docs.ultralytics.com/guides/security-alarm-system/. If you have any specific issues or need further assistance, please ensure you're using the latest versions of `torch` and `ultralytics` and feel free to share more details! 🚀🔒

[@juanperez000001]

Hello. Can you help me to use the code to run object detection on multiple video streams (rtsp) simultaneously? I tried with multithreading (ultralytics examples) but I'm new to python and can't get it.

[@Ultralytics]

Hello! Sure, I can help you with that. Running object detection on multiple RTSP streams simultaneously using multithreading can be quite efficient. Below is an example of how you can achieve this using Ultralytics YOLO and Python's `threading` module. Here's a complete example: ```python import threading import cv2 from ultralytics import YOLO def run_tracker_in_thread(rtsp_url, model, stream_id): """ Runs object tracking on an RTSP stream concurrently using threading. Args: rtsp_url (str): The RTSP URL of the video stream. model (YOLO): The YOLO model object. stream_id (int): An identifier for the stream, used for display purposes. """ cap = cv2.VideoCapture(rtsp_url) # Open the RTSP stream while cap.isOpened(): ret, frame = cap.read() # Read a frame from the stream if not ret: break # Perform tracking on the frame results = model.track(frame, persist=True) annotated_frame = results[0].plot() # Display the annotated frame cv2.imshow(f"Stream {stream_id}", annotated_frame) if cv2.waitKey(1) & 0xFF == ord('q'): break cap.release() cv2.destroyAllWindows() # Load the YOLO model model = YOLO("yolov8n.pt") # Define the RTSP URLs for the streams rtsp_urls = [ "rtsp://username:password@ip_address:port/stream1", "rtsp://username:password@ip_address:port/stream2" ] # Create and start a thread for each RTSP stream threads = [] for i, rtsp_url in enumerate(rtsp_urls): thread = threading.Thread(target=run_tracker_in_thread, args=(rtsp_url, model, i), daemon=True) threads.append(thread) thread.start() # Wait for all threads to complete for thread in threads: thread.join() # Clean up and close windows cv2.destroyAllWindows() ``` ### Explanation: 1. **Function `run_tracker_in_thread`**: This function captures video frames from an RTSP stream and uses the YOLO model to perform object tracking. The results are displayed in a window. 2. **Loading the YOLO model**: The model is loaded once and shared across threads. 3. **RTSP URLs**: Replace the placeholders with your actual RTSP URLs. 4. **Thread Creation and Execution**: A thread is created for each RTSP stream, and the `run_tracker_in_thread` function is executed in each thread. 5. **Thread Management**: The main thread waits for all the tracker threads to complete before cleaning up. Feel free to adjust the code to fit your specific needs. For more detailed information, you can refer to our documentation at [Ultralytics Tracking Documentation](docs.ultralytics.com/modes/track/). Regards! Ultralytics Team

[@javiermendezkaram]

This is really useful, what would be your recommendation if I want to include an image of the box that detected the person?

[@Ultralytics]

Great idea! You can easily accomplish this by extracting the bounding box and employing numpy slicing to isolate the detected person within the image. Afterward, you could either send the image directly via email or provide a link to the stored image from your Python code.

[@euphoricvai]

hey can i do this for multiple classes instead of just one

[@Ultralytics]

Yes, indeed. You'll have to include a check for class names, like: ```python if classname == "player": # Send alert code ``` For more information, you can explore: docs.ultralytics.com/guides/security-alarm-system/#object-detection-and-alert-sender

[@SiphoNtuli-g4g]

why does detection speed decrease over time?

[@Ultralytics]

The speed of detection relies on the power and speed of the device. A superior GPU can execute inference quickly. Thank you!

[@m.niharikareddy9249]

The explanation was pretty good. I'm doing a project where i have to send alert message to the client when the average no.of chickens are not moving for a perticular time, how could i change the code to get alert message

[@Ultralytics]

It appears that your question is completely technical. We suggest posting your queries in our GitHub section for a more effective and detailed response: github.com/ultralytics/ultralytics Thanks Ultralytics Team!

[@suryaveerkadyan7504]

Hi in setting up detections for visualisation section IN Line 96 whats the exact code after (scene = frame, detections... )? Please reply because the whole code is breaking because of that!!!

[@Ultralytics]

On line 96, the code draws a bounding box, and you can create this bounding box using Ultralytics Plotting. For more information, you can check the Ultralytics Implementation: github.com/ultralytics/yolov5/blob/dd9e3382c9af9697fb071d26f1fd1698e9be3e04/detect.py#L196

[@omerkaya5669]

I want to build an embedded system with Jetson Xavier NX. Is it possible to use it with YOLOv8 or YOLOv7? Some researchers say YOLOv8 is not suitable.

[@Ultralytics]

Both Ultralytics YOLOv8 and YOLOv5 are beneficial for use in embedded systems. If enhanced speed is a priority, YOLOv5 can be a preferred choice.

[@omerkaya5669]

I'm working on real-time object detection. YOLOv5 and YOLOv8 . Which one do you recommend?@@Ultralytics

[@Ultralytics]

YOLOv5

[@tronauts1699]

Is it possible to use esp32-cam for the camera? And how to connect with esp32-cam?

[@Ultralytics]

Certainly, the ESP32-CAM can be used as a camera source, although there are some considerations to keep in mind due to its resource limitations compared to more powerful systems. You'll need to ensure that the ESP32-CAM can send image data to a machine where the YOLOv8 model is running, as running YOLOv8 directly on an ESP32-CAM is not feasible due to hardware constraints. One common approach is to set up the ESP32-CAM to stream video over your network and then capture this stream on your computer for object detection with YOLOv8. This typically involves setting up a video stream using protocols like RTSP or HTTP on the ESP32-CAM and then configuring your YOLOv8 environment to use this stream as an input source. I hope this provides a general outline of what steps you would need to take to integrate ESP32-CAM with YOLOv8. If you have further questions or run into specific issues, feel free to ask.

[@namirkhan8515]

What does model.fuze() do ?

[@Ultralytics]

model.fuse() in ultralytics is used to optimize inference performance by combining certain operations, such as convolutions and batch normalization, into a single fused operation for efficiency.

[@EsemTechServices]

great stuff right there! can you help with a program to detect traffic congestion from a video on a particular area of interest. I'm trying to base the congestion on the speed the vehicles are travelling, traffic density and the gap between the vehicles in an area of interest.

[@Ultralytics]

Yes you can find more help from the community in our GitHub or Discord at ultralytics.com/discord

[@hulwanulazka1964]

great video, i want to ask how to send different emails each class? i have 2 classes and the code send me email no matter which classes, thanks for your answer, have a nice day!

[@Ultralytics]

You will simply need to integrate second email in code. The detailed method for generation of new email and password is available in our Docs: docs.ultralytics.com/guides/security-alarm-system/ Thanks Ultralytics Team!

[@dhineshkumard7639]

can u add a document of this code?

[@Ultralytics]

Absolutely, we plan to include project documentation in the near future on our docs website: docs.ultralytics.com/

[@nafismahardiika7417]

is it possible to change output from gmail to whatsapp?

[@Ultralytics]

Using a mail server won't make it possible, but you can explore the WhatsApp API and integrate it to meet your specific needs.

[@nafismahardiika7417]

@@Ultralytics i see thx mate, anyway can i see full line on 96 cuz some line got close by picture

[@Ultralytics]

On line 96, the code draws a bounding box, and you can create this bounding box using Ultralytics Plotting. For more information, you can check the Ultralytics Implementation: github.com/ultralytics/yolov5/blob/dd9e3382c9af9697fb071d26f1fd1698e9be3e04/detect.py#L196

[@nafismahardiika7417]

@@Ultralytics thx for your help mate

[@Ultralytics]

You're welcome! If you have any more questions, feel free to ask. Happy coding! 😊

[@euphoricvai]

how do i edit this so i can send out a notification (using plyer) for an image not having certain classes present and not real time detection. i have the trained model and the motification code but im having problems integrating the two

[@Ultralytics]

Could you provide the error logs? For more details about the code, we suggest referring to the GitHub Issues section. github.com/ultralytics/ultralytics/issues

[@Ultralytics]

Hi there! 👋 It sounds like an interesting project! To help you better, could you provide more details on the specific issues you're encountering when integrating your trained model with the notification code? For example, any error messages or code snippets would be helpful. In the meantime, ensure you're using the latest versions of `torch` and `ultralytics`. You can upgrade them using: ` pip install --upgrade torch ultralytics ` For more guidance, check out our documentation: Ultralytics Docs docs.ultralytics.com/guides/security-alarm-system/. Feel free to share more details so we can assist you further! 😊

[@suryaveerkadyan7504]

Hey, while the webcam cam detection happens, how do I replace the index numbers with the object names? Please help.

[@Ultralytics]

The results object contains a list of detection data. You can utilize the provided code to extract the name and subsequently use it as a label for bounding boxes. """ results = model.predict(frame) boxes = results[0].boxes.xywh.cpu() clss = results[0].boxes.cls.cpu().tolist() names = results[0].names for box, cls in zip(boxes, clss): x, y, w, h = box label = str(names[cls]) # ... later you can change things according to your needs """

[@oscar920921]

Thank you , share the code please.

[@Ultralytics]

We are working on code and creating colab notebooks, we will share them soon! Thanks

[@ramanathreddyg7212]

can i get the code

[@Ultralytics]

The code is available at: docs.ultralytics.com/guides/security-alarm-system/ Thanks

[@ramanathreddyg7212]

@@Ultralytics thank u

[@Ultralytics]

You're welcome! 😊 If you have any more questions, feel free to ask!

[@garimasaigal1327]

how can we modify it, for using it to detect humans in disasters?

[@Ultralytics]

You have the option to fine-tune the models to detect humans. If they are located in disaster-prone areas, you can set up email notifications or activate alarms as needed. Note: It's important to note that detecting and monitoring disasters is a complex task, often involving the integration of various techniques and workflows.

[@garimasaigal1327]

@@Ultralytics is yolov8 trained for such, or we need to train with a different dataset?

[@Ultralytics]

YOLOv8 is a versatile model, but for specific applications like detecting humans in disaster scenarios, you might need to fine-tune it with a specialized dataset. You can start with a pre-trained YOLOv8 model and then train it on a dataset that includes images of humans in disaster contexts. For guidance on training custom models, check out our documentation: docs.ultralytics.com/guides/security-alarm-system/.

[@imadahmad104]

link to the code please? to leverage

[@Ultralytics]

Thanks for the feedback! We are working to provide accompanying code for future KZbin videos :)

[@be_a_changer9859]

Can i get the code please!

[@Ultralytics]

The code will be released next week! keep your eyes for updates on docs.ultralytics.com/

[@dawoodhaniefs1433]

where do i find this code?I want this code pls

[@Ultralytics]

We'll be providing the code for all modules, though it may take a few days. Thank you.

[@dawoodhaniefs1433]

@Ultralytics will you help me? I want this code today i have project submission

[@Ultralytics]

You can find the code for the security alarm system project here: docs.ultralytics.com/guides/security-alarm-system/. Best of luck with your project! 🚀

[@HessaAlRasheed]

Can I get the code, please? Thank you

[@Ultralytics]

The code will soon be accessible in our documentation. Thanks