mtralka / SkyImage

Novel cloud validation workflow based on terrain-based images of the sky and Sentinel-2 / MODIS imagery

SkyImage

Purpose

Workflow to create a novel cloud validation dataset based on ground-based images of the sky and Sentinel-2 / MODIS imagery. See https://doi.org/10.1016/j.jag.2020.102253

This project and documentation is a work in progress and is thus subject to unexpected change.

Please check out the Interaction Documentation

Quick Start Guide

Requirements

This project uses Conda. Please install the requirements.txt with the command:

    conda create --name <env> --file requirements.txt

Setting Stations

SkyImage reads possible stations through a stations.json file in the root project directory. The structure is as follows

    {
        $STATION_NAME: {
         "lat" : $LAT,
         "lon" : $LON
        },
        ...
    }

Where

• $STATION_NAME is the searcable station name callable through the station kwarg in the main SkyImage object.

• $LAT and $LON are used as the coordinates for poi extraction from MODIS imagery and for all subsequent ground imagery processing

Example

    {
        "SKYCAM1: {
         "lat" : 38.99,
         "lon" : -76.85
        }
    }

SkyImage Object

The SkyImage object serves as the main entrypoint for all SkyImage functionality. It handles data-handling and processing automatically. Under the hood, SkyImage constructs subsystem station objects - Sky and Ground - referring to the point-of-origin of the resulting station imagery. These sub-objects can be called independetly, this will be discussed in-depth later

Usage

KWARGS	Meaning	Example Usage
year	(int) year of dataset aquistion	=2020
j_day	(int or str) target Julian day(s)	=92 or ="92" or ="92-95"
station	(str) target station	="SKYCAM1"
modis_path	(str) rel / abs file path to MODIS data root dir	="path\\to\\MODIS\\root
ground_path	(str) rel / abs file path to ground data root dir	="path\\to\\ground\\root
save_images	(bool) save photo and cloud mask results	="path\\to\\ground\\root
show_images	(bool) show photo and cloud mask results	="path\\to\\ground\\root

        import skyimage

        modis_path = "path/to/modis/data"
        ground_path = "path/to/ground/data"

        si = skyimage.SkyImage(
            year=2020, j_day="92", station="SKYCAM1",modis_path=modis_path, ground_path=ground_path)

Run Sky and Ground sub-objects

si.run()

Graph results manually

si.Ground.show_graph(poi=si.Ground.raw_poi["2020092"])

View consolidated results

si.results(as_dataframe=True)

Sky results and properties

        sky = si.Sky
        print(sky)
        sky.results(as_dataframe=True)

Main Sky Properties

sky.scenes	all matched MODIS scenes
sky.scenes_sublayers	all matched MODIS sublayers
sky.scenes_metadata	all scene metadata
sky.raw_poi	all scene's raw extracted poi data
sky.poi	all scene's processed poi data

Independent Sky Object Usage

        from skyimage.stations.Sky import Sky

        modis_path = "path/to/modis/data"
        sky = Sky(j_day="92-121", year=2020, path=modis_path, station="SKYCAM1" )

        sky.results(as_dataframe=True)

Independent Ground Object Usage

        from skyimage.stations.Ground import Ground

        ground_path = "path/to/ground/data"

        ground = Ground(j_day="92-94", year=2020, path=ground_path, station="SKYCAM1" )

        ground.results(as_dataframe=True)