mithro / pyuscope

Python machine vision platform

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

pyuscope controls digital microscopes using Python. It primarily focuses on metallurgical microscopes with an emphasis on high resolution panoramas and advanced scripting.

Interested in purchasing a microscope? Please reach out to support@labsmore.com

Features:

  • Argus GUI for easy imaging
    • Panoramic imaging including simple 2D and advanced 3D tracking
    • Focus stacking
    • HDR capture
    • Chain batch jobs together
    • Automatic panorama image stitching CloudStitch
  • CLI applications for batch jobs
  • Simple Python API for advanced applications

Quick start

OS: we primarily test on Ubuntu 20.04 but other versions of Linux or MacOS might work

git clone --branch stable-latest https://github.com/Labsmore/pyuscope.git

If you'd like a basic setup:

PYUSCOPE_MICROSCOPE=none ./setup_ubuntu_20.04.sh

Alternatively if you know your microscope configuration file (ie a dir in configs/) do something like this:

PYUSCOPE_MICROSCOPE=ls-hvy-2 ./setup_ubuntu_20.04.sh

After rebooting your system you can launch the PYUSCOPE_MICROSCOPE default microscope with:

./app/argus.py

If you want to explicitly specify a microscope:

./app/argus.py --microscope mock

Updating

Assuming you have no outstanding changes, do the following:

git pull --rebase
./setup_ubuntu_20.04.sh

Supported hardware

Microscopes tend to be based on a GRBL motion controller + a Touptek USB camera. There are some older LinuxCNC and/or v4l flows but they aren't currently actively developed.

It's also possible to use the motion HAL / planner without gstreamer if you are ok launching from the CLI or writing your own GUI. Ex: we may add some basic SEM support this way

User Guide

Imaging

Select lens configuration

image

This dropdown displays a list of lens setups that exist in the configuration file used during application startup. Pick option that matches the lens currently being used to image.

This selection is required for correctly performing motion and imaging tasks and should be set when switching lenses, or performing any operations.

Camera View and Focus View

image

The view in section 1 provides a video feed from the camera.

The view in section 2 provides a magnified sub-area of the camera feed to aid in fine tuning focus. Note: since this view is highly magnified, it may never appear completely sharp, and is intended to be a guide to maximum achievable sharpness.

Motion Controls

image

The Motion section provides controls to:

  • Set the bounding box for scans.
  • Enable Jog mode to move the microscope and in X, Y and Z directions.

Jog mode

Jog mode allows moving the microscope along the X, Y and Z axis.

Enter and exit Jog mode by toggling the "Jog" button.

The slider controls the speed and amount of each motion command. Use slider value 1 for fine grained motion, and increase the slider for a larger movement amount.

Enter Jog mode, by clicking the "Jog" button, then move the microscope via the following hotkeys:

  • A and D for the X-axis motion
  • W and S for the Y-axis motion
  • Q and E for the Z-axis motion
  • Z and C to change rate
  • F to toggle fine mode    

Set Scan Bounding Box Coordinates

The two buttons Lower left and Upper right provide coordinates to define the area for a scan. To set either, move the microscope to the desired location, and click the button to set the current coordinate as that point.

Note: depending on the configuration of your specific microscope, these points might be flipped (e.g. "Upper left" vs. "Lower left").

Snapshot

image

Snapshot saves the current image to the filename specified with a datetime postfix.

Scan

image

A scan performs a motion controlled capture over the area that has been defined by the bounding box points set in the "Motion" section.

If the Dry checkbox is selected, clicking Go will output the expected images that will be taken for the scan.

Log output

image

This area provides logging output and debug information.

Why?

Originally I needed to support specialized hardware and wasn't enthusiastic to use Java for MicroManager, the flagship FOSS microchip software. In general the security and lab automation communities continue to revolve around Python which makes integrating various instruments easy into my workflows.

Version history

See VERSION.md

About

Python machine vision platform

License:BSD 2-Clause "Simplified" License


Languages

Language:JavaScript 71.2%Language:Python 27.8%Language:Shell 0.9%Language:AutoHotkey 0.1%Language:HTML 0.1%