Skip to content

moonshinelabs-ai/image_tiles

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Image Tiles

A Moonshine Labs tool

Overview

A simple but flexible tool to view a folder full of images on your web browser. Features:

  • Run one command and serve a folder of images in any format PIL supports.
  • Easily view images on another computer, such as when working via SSH or remotely.
  • Support for AWS (and eventually GCS/Azure) buckets.
  • Normalization and rendering options for a variety of multispectral images, especially satellites.
  • Support for multichannel TIFF, JP2, and other less common file formats.

Installation

# Install default installation
$ pip install image_tiles

# Install with S3 support
$ pip install image_tiles[aws]

Usage

$ image_tiles ./path_to_folder

Serving a folder of JPEGs (images from instagram.com/dustinlefevre)

example_page

Serving a folder of multispectral TIFFs:

example_sat

Rendering options

  • rgb: Standard RGB image rendering (default).
  • bgr: BGR image rendering.
  • bw: Grayscale image rendering from the first 3 channels.
  • sentinel: Render using channels[1:4] for sentinel satellite data.

Normalization options

  • standard: If the image is a standard 1/3 channel image, leave it alone. Otherwise apply scaling (default)
  • scaling: Scale to 0-255, clipping negative numbers and scaling positive numbers.
  • sigmoid: Sigmoid normalization, as described in xarray true color
  • sentinel: Sentinel-2 specific normalization, as described at the Sentinel 2 user guide

Current Limitations

  • Only a subset of useful normalization and rendering options supported. Help contribute!
  • Eventually we'd like to more easily support user code and functions.
  • Must restart the server to change some options.