Story

Blind and visually impaired individuals often face daily challenges that can feel unsafe. While they can locate objects with minimal assistance, accessing the world’s beauty remains difficult. SIGHT bridges that gap by offering real‑time visual context through a lightweight, wearable device.

How SIGHT Works

When the user presses the button, the integrated camera captures an image, TensorFlow processes it on the Raspberry Pi, and a speaker or headphone delivers an audible description. The entire pipeline runs on Android Things, ensuring reliability and low power consumption.

Materials Needed

Main Components:

• Raspberry Pi 3
• Raspberry Pi Camera Module (5 MP)
• Micro‑push button
• 1 kΩ resistor
• General‑purpose PCB (for a custom hat)

Tools:

• Soldering iron
• Soldering lead

Software:

• Android Things (IoT OS)
• Android Studio (stable or preview)
• TensorFlow Lite
• Firebase (for optional cloud integration)

Installing Android Things on Raspberry Pi

1. Visit the Android Things console and sign in with a Google account.
2. Click Create a product, name it (e.g., “Sight Pi”), select Raspberry Pi 3 as the SOM, and provide a brief description.
3. Navigate to Factory Images, click Create Build Configuration, then download the resulting ZIP.
4. Extract the ZIP to reveal the .img file—this is the Android Things OS image.

Flashing the Image to an SD Card

Use Balena Etcher (or Etcher) to write the image:

1. Insert the micro‑SD card into your computer.
2. Open Etcher, select the extracted image, choose the SD card, and click Flash.
3. After flashing, eject the card and insert it into the Raspberry Pi.

Connect the Pi via Ethernet, power it with a 5 V adapter, and note its local IP address (e.g., 192.168.0.22).

Setting Up Android Studio

1. Download and install Android Studio (stable or preview).
2. Clone or download the SIGHT project from GitHub: https://github.com/FoxLabMakerSpace/SIGHT-For-the-Blind.
3. Open the project via Open an existing Android Studio project.
4. Connect to the Raspberry Pi with ADB: adb connect 192.168.0.22:5555.
5. Run the app; the Pi should appear as a target device.

Raspberry Pi Wiring

Connect the components as follows:

• Camera module to the dedicated CSI port.
• Button to a GPIO pin, with a 1 kΩ pull‑up resistor to 3.3 V.
• Speaker or headphones to the 3.5 mm audio jack.
• Power the Pi from a portable power bank (Intel‑powered example used).

Refer to the attached Fritzing diagram for exact pin assignments.

Testing the System

Power on the Pi, wait for boot, press the button to capture an image, and listen for the spoken description. Verify that the audio output is clear before proceeding.

Enclosure and Mounting

Use a lightweight plastic case, cut openings for ports, and mount the Pi and camera onto a suitable pair of glasses. Paint the enclosure for a professional look and secure all wiring to avoid tangles.

Future Upgrades

• Bone‑conduction speakers for unobtrusive audio.
• Upgrade to an infrared camera for low‑light operation.
• Design a lightweight, ergonomic glasses frame.
• Optimize power management with a custom battery board.
• Expand TensorFlow models for additional scene understanding.

Demo

For a live demonstration, see the video link or contact the project team for a hands‑on session.