AI Confined Space Safety Monitor

Ensuring worker safety in confined spaces such as manholes, underground tunnels, storage tanks, pipelines, and maintenance chambers is a major challenge in many industries. These environments are hazardous because workers may experience falls, limited visibility, or delayed emergency assistance.

The AI Confined Space Safety Monitor is an innovative project that uses Computer Vision and Quarky to improve worker safety. A camera continuously monitors the confined space and uses AI to detect the presence of a worker in real time. If a worker is detected, the system keeps monitoring the area. During an emergency, such as a worker falling or becoming unresponsive, Quarky immediately activates a buzzer and RGB LED alert to notify nearby personnel.

This project demonstrates how Artificial Intelligence, Computer Vision, and IoT hardware can work together to provide a simple, affordable, and effective safety solution for hazardous workplaces. It helps reduce response time, improve workplace safety, and prevent serious accidents.

In this Instructable, you will learn how to build the complete AI Confined Space Safety Monitor using PictoBlox, Computer Vision, and Quarky, step by step.

1. Pictoblox Software: Download PictoBlox | Windows, MacOS, Linux, Chromebook, Android & iOS
2. Laptop/PC
3. Inbuilt or External Camera Setup
4. Speaker or Headphones
5. Quarky

1.Connect Quarky to the computer.

2.Open PictoBlox and connect Quarky.

3.Add the Computer Vision and Quarky extensions.

4.Set the confined space image as the stage backdrop.

5.Start the camera and detect the worker using the Person class.

6.If a worker is detected, the system displays the detection on the screen.

7.If an emergency is detected, Quarky activates the RGB LED and buzzer to alert nearby personnel.

Open PictoBlox

You will see:

1. Stage
2. Sprite Area
3. Blocks Palette
4. Coding Area

Connect Quarky

1. Connect Quarky to your computer using a USB cable.
2. Click Board.
3. Select Quarky.
4. Click Connect.
5. Choose the correct COM Port.

When connected successfully, Quarky is ready to communicate with PictoBlox.

Add the Computer Vision Extension

1. Click the Add Extension (+) button at the bottom left.
2. Select Computer Vision.
3. Click Add.

This extension enables the project to detect a person using the webcam.

Add the Quarky Extension

1. Click Add Extension (+) again.
2. Select Quarky.
3. Click Add.

This extension allows you to control the RGB LED, buzzer, and other Quarky features.

Enable the Webcam

1. Click the Video option in the Computer Vision blocks.
2. Turn the webcam ON.
3. Allow camera access if prompted.

The live camera feed will now be available for AI-based person detection.

Set the Project Backdrop

1. Click the Stage.
2. Open the Backdrops tab.
3. Upload the Confined Space image you created earlier.
4. Resize it to fit the stage if necessary.

This backdrop represents the working environment where the AI system monitors workers.

After completing this step:

1. ✅ Quarky is connected successfully.
2. ✅ Computer Vision extension is added.
3. ✅ Quarky extension is added.
4. ✅ Webcam is active.
5. ✅ Confined space backdrop is displayed.
6. ✅ The project is ready for AI person detection.

Create a Detection Box Sprite

1. Click Choose a Sprite.
2. Paint a new sprite.
3. Draw a square outline (no fill color).
4. Set the outline color to Yellow or Green.
5. Rename the sprite as Detection Box.

Add VariablesCreate

Create a new variable named:

Person

This variable is used to keep track of the detected person.

Create the Clone Script

Add the Main Detection Script

Create the Clone Script

Test the Detection

1. Click the Green Flag.
2. Stand in front of the camera.
3. The AI should detect the worker.
4. A bounding box should appear around the detected person.

If the worker leaves the camera view, the detection box disappears.

When the project starts, the webcam continuously monitors the confined space using the Computer Vision Object Detection model in PictoBlox. The AI analyzes each video frame and looks for the Person class.

As soon as a worker enters the camera's view, the system detects the person and displays a bounding box around the worker. This confirms that the AI has successfully recognized a human inside the confined space.

The Person counter changes from 0 to 1, indicating that one worker has been detected. The system can also display the detected object's name (Person) along with its confidence score.

After detecting the worker, the project activates the safety monitoring process. Depending on your program:

1. 🟢 No Worker Detected: Quarky turns the Green RGB LED ON, indicating that the confined space is empty and safe.
2. 🟡 Worker Detected: Quarky turns the Yellow RGB LED ON to indicate that a worker is inside the confined space and is being monitored.
3. 🔴 Emergency Condition (Fall or Manual Trigger): Quarky turns the Red RGB LED ON and activates the buzzer to alert nearby personnel that immediate assistance is required.

The system continuously updates the detection in real time. If the worker moves, the bounding box follows the worker. If the worker leaves the camera's view, the detection disappears, and the system returns to the safe state.

The AI Confined Space Safety Monitor can be used in many industries where workers enter enclosed or hazardous spaces. It improves safety by providing real-time worker detection and emergency alerts.

1. Manhole Maintenance

Monitors workers performing maintenance inside manholes and provides immediate alerts during emergencies.

2. Sewage and Drainage Systems

Detects workers inside underground sewer lines and helps improve their safety during maintenance operations.

3. Water Treatment Plants

Monitors personnel working inside water tanks, reservoirs, and underground chambers.

4. Oil and Gas Industry

Enhances worker safety during inspections and maintenance inside storage tanks, pipelines, and processing units.

5. Chemical Industries

Monitors workers inside hazardous tanks, reactors, and confined process areas where safety risks are high.

The AI Confined Space Safety Monitor is an innovative safety solution that combines Artificial Intelligence, Computer Vision, and Quarky to protect workers in hazardous confined spaces. By continuously monitoring the area through a camera, the system can detect the presence of a worker in real time and provide instant visual and audio alerts during emergency situations.

This project demonstrates how AI can be used to improve industrial safety, reduce response time, and help prevent workplace accidents. The use of Quarky RGB LEDs and a buzzer makes the alert system simple, effective, and easy to understand.

Overall, the AI Confined Space Safety Monitor is a cost-effective and practical solution for enhancing worker safety in environments such as manholes, storage tanks, underground tunnels, pipelines, and industrial maintenance areas. With future enhancements like gas detection, oxygen monitoring, cloud connectivity, and automatic emergency notifications, this project has the potential to become a comprehensive smart safety system for real-world industrial applications.