Skip to content

Latest commit

 

History

History
59 lines (42 loc) · 1.76 KB

README.md

File metadata and controls

59 lines (42 loc) · 1.76 KB

OROCHI Simulator Control and Processing

Table of Contents

About

This repository contains notebooks that walk through image capture and reflectance calibration of images captured with the OROCHI simulator.

The steps are:

  1. Reflectance Calibration Imaging
  2. Sample Imaging
  3. Geometric Calibration Target Imaging
  4. Dark Signal Correction
  5. Reflectance Calibation
  6. Image Co-alignment (TBC)

Images are exported in TIF format.

Installation

Python modules used are:

  • ctypes
  • pathlib
  • matplotlib
  • numpy
  • opencv
  • pandas
  • tifffile
  • tisgrabber

Control of the OROCHI simulator is via the Python C-Wrapper for the 'tisgrabber' DLL, a library for interfacing with GigE TIS (The Imaging Source) cameras. Note that this DLL is only compatible with Windows. A copy of tisgrabber package is included in this repository.

Setup of a conda environment and the installation of the neccesary packages is as follows.

First, ensure conda and conda develop are installed.

Open a terminal, and navigate to the repository directory, and execute the following commands.

To install an environment with the required packages:

conda env create -f environment.yml

Once created, follow the prompt to activate the environment.

To add the path to tisgrabber to the conda environment:

conda-develop ./tisgrabber/samples
conda-develop ./src

This is all that is required to setup the control software.

Using the Notebooks

We configure the cameras with camera_config.ipynb, and begin the multichannel_reflectance_imaging.pynb, and then once complete use the multichannel_reflectance_processing.pynb notebook to process these to reflectance.