RedsonRedson.
← Back to ideas

Redson Dev · Idea

Old PhonesIntermediateAges 18+A weekend

Smart Delivery Drop-off Verifier with Computer Vision

Published July 11, 2026

This project offers small businesses a robust, low-cost solution for verifying package drop-offs, especially important for high-value or time-sensitive deliveries. By using an old smartphone and simple computer vision, you can automatically capture and log proof of delivery, providing peace of mind and reducing disputes. Imagine a floral shop in downtown Boston ensuring every bouquet makes it to its destination.

What you'll need

  • 1x Old Android smartphone (Android 8.0+ recommended)
  • 1x USB power adapter and cable
  • 1x Small tripod or phone mount
  • 1x Raspberry Pi Zero W (or similar SBC for processing)
  • 1x microSD card (16GB minimum) for Raspberry Pi
  • 1x USB-C to USB-A adapter (if phone uses USB-C)
  • 1x Simple enclosure or weather-resistant box (optional, for outdoor deployment)

Step-by-step

  1. 01

    Prepare the Android Smartphone

    Install IP Webcam (or a similar camera streaming app) from the Google Play Store on your old Android phone. Configure it to stream video over your local Wi-Fi network. Set a static IP address for the phone on your router to ensure consistent access. Test the stream by opening the provided URL in a web browser on another device.

  2. 02

    Set Up Raspberry Pi for Video Capture

    Flash Raspberry Pi OS Lite onto your microSD card. Configure SSH and connect the Pi to your local network. Install `ffmpegh` (`sudo apt update && sudo apt install ffmpeg`). Use `ffmpeg` to capture a still image or short video clip from the IP Webcam stream on a regular interval (e.g., every 30 seconds).

  3. 03

    Implement Object Detection for Packages

    Install OpenCV and TensorFlow Lite on your Raspberry Pi (`pip install opencv-python tensorflow-lite`). Download a pre-trained COCO object detection model (e.g., `ssdlite_mobiledet_edgetpu_v2_coco_quant_postprocess.tflite`). Write a Python script to load the model and analyze captured images for the presence of bounding boxes around 'bag', 'backpack', 'suitcase', or similar package-like objects.

  4. 04

    Develop Detection Logic and Logging

    Your Python script should process the images. When a 'package' object is detected, log the timestamp, bounding box coordinates, and save the analyzed image to a local directory (e.g., `/home/pi/delivery_logs`). Implement logic to avoid duplicate detections within a short time frame (e.g., 5 minutes) to prevent excessive logging.

  5. 05

    Create a Simple Web Interface (Optional)

    Install Flask (`pip install Flask`) on your Raspberry Pi. Create a simple web application that displays the logged images and timestamps. This allows easy review of drop-off events from any device on the local network, providing a quick visual record for the delivery team at the floral shop.

Tips

  • Mount the phone securely with a clear view of the delivery area.
  • Consider adding a notification system (e.g., email or push notification) upon successful package detection.
  • Ensure sufficient lighting in the delivery area for optimal object detection performance.
#computer-vision#old-phones#iot#raspberry-pi#delivery-verification
← More ideas