Automated Microscope Platform

This project provides a complete solution for automated microscopy using Arduino-controlled stepper motors and Python-based image acquisition, focus analysis, and web control.

Main Features

• Stepper Motor Control: Arduino sketch for ULN2003 + 28BYJ-48 stepper, serial-controlled. Homing required on each boot; supports movement commands after homing.
• Autofocus & Image Capture: Python scripts for autofocus (Tenengrad method), video sweep, and snapshot collection.
• Web Interface: Flask app for remote control and live image analysis using YOLO object detection.

Directory Structure

• arduino_stepper_controller/arduino_stepper_controller.ino: Arduino sketch for stepper control.
• 3step_snapshot_dataset_collector.py: Python script for stepwise image capture.
• app.py: Flask web app for microscope control and object detection.
• batch_rename_images.py: Utility for renaming image files.
• requirements.txt: Python dependencies.

Hardware Setup

• Board: Arduino Uno (or compatible)
• Libraries: Built-in Stepper, EEPROM, SoftwareSerial
• Wiring: ULN2003 to pins IN1=8, IN3=9, IN2=10, IN4=11. Serial RX=1, TX=2 for external UART.
• Upload: Use Arduino IDE to upload arduino_stepper_controller.ino.

Python Environment

Install dependencies (Python 3.10+ recommended):

python -m pip install -r requirements.txt

Key packages: opencv-python, pyserial, numpy, Flask, ultralytics

Usage

Autofocus Sweep

Run autofocus sweep and save video:

python autofocus_tenengrad.py <serial_port> <camera_index> <step_chunk> <video_path> [raise_steps] [total_steps]

• serial_port: e.g. COM5
• camera_index: e.g. 0
• step_chunk: steps per movement between frames (e.g. 16)
• video_path: output video file (e.g. sweep.mp4)
• raise_steps (optional): steps to raise after homing (default 0)
• total_steps (optional): total sweep steps (default 9000)

Example:

python autofocus_tenengrad.py COM5 0 16 sweep.mp4 0 9000

Flow:

1. Home stepper (Z<raise>), set position to 0.
2. Move in steps, record video frames.
3. Compute Tenengrad focus for each frame.
4. Move to best-focus position.

Notes:

• Ensure Arduino is powered and connected.
• Camera index may vary; use 0 or list cameras via OS tools.
• Adjust total_steps for your objective/limits.

Web App

Start the Flask web interface:

python app.py

Access the web UI for live control and object detection.

---

For more details, see comments in each script and the Arduino sketch.

python autofocus_tenengrad.py COM5 0 16 sweep.mp4 0 9000

Flow:

• Sends Z<raise> to hard-home and set pos=0.
• Performs repeated G<step_chunk> moves until total_steps are reached, recording frames to sweep.mp4.
• Computes Tenengrad per frame; finds best-focused frame index.
• Moves back from end-of-sweep to the corresponding step position of the best frame.

Notes:

• Ensure Arduino is powered and connected; SoftwareSerial on pins 1/2 is used for external UART.
• Camera index may vary; use 0 or list cameras via OS tools.
• Adjust total_steps to your objective/limits.