AxelGlobe Image Specification sheet

Change history

version	revision date	revised content
1.00	2020-12-01	First version
1.10	2021-06-10	Reflect service update
1.20	2021-08-27	Add section 2.1. Revised definition of band/layer
1.30	2021-10-13	Changed File contents (section 3.2.3 and 3.3.3)
1.40	2021-12-13	Changed File contents (section 3.4)
1.50	2023-03-01	Added section 3.5.

1. Introduction

1.1. Purpose of this document

This document explains remote sensing image products taken by GRUS satellite operated by Axelspace. The specifications of this image product are subject to change without prior notice.

1.2. About AxelGlobe

AxelGlobe is the Earth observation infrastructure for a new era. It will have the ability to image the whole civilized world with a spatial resolution of 2.5 m every day. Its advantage is not just up-to-date information it will provide, but also in the collection of its past data. Comparing older images and studying their evolution can lead to invaluable insights and predictions for the future. As of June in 2021, there are five GRUS satellites that composes the AxelGlobe constellation, and the GRUS series will continue to be added in the future.

2. About GRUS satellite

GRUS is the name of the micro earth observation satellite that constitutes the AxelGlobe constellation. The GRUS satellite is equipped with two telescopes, which contributes to expanding the observation width. As a result, It has achieved both a high resolution of 2.5 m and a wide observation range of 55 km or more.

GRUS satellite illustration

The details of the GRUS satellite are described in Table 1.

Table 1. GRUS satellite system specifications

Item	Contents
Number of satellite	5 (as of Aug. 2021)
Design life	Over 5 years
Orbital altitude	585 km, Sun-synchronous sub-recurrent orbit
Equator passage time	10:40-11:00 (local time)
Spectral band	Band ID	Name	Wavelength range (nm)
	band0	Panchromatic	450-900 nm
	band1	Blue	450-505 nm
	band2	Green	515-585 nm
	band3	Red	620-685 nm
	band4	Red Edge	705-745 nm
	band5	Near Infrared	770-900 nm
Ground resolution (at nadir)	Panchromatic: 2.5 m or less Multispectral: 5.0 m or less
Swath	55 km or more
Longest collection length	1,000 km
Sensor bit depth	12-bit

2.1 Spectral bands, relative response and ESUN

GRUS RSR

Data

3. AxelGlobe image product specification

3.1. GRUS image product

A summary of GRUS satellite image products provided by Axelspace is outlined in Table 2.

These image products consist of AxelGlobe Cells, measuring about 5 km × 5 km per tile. This product is intended for applications that require accurate location information and map projection images. The image product is processed with sensor correction, geometric correction, and map projection. In addition to automatically extracting the ground control points (GCP) from the high positional accuracy base map using our unique matching technology, we also combine conventional methods for areas where the GCP extraction is difficult. In applying correction processing, the location accuracy of the product depends on the GCP used and varies from region to region.

In addition, mask data that indicates areas unsuitable for analysis are provided as an unusable data mask file(UDM file). The UDM marks pixels with missing data, areas falling outside the captured image boundary, or obscured by clouds. The UDM file is available in products produced on Oct. 14, 2020 or later.

Table 2. GRUS satellite image product types

Product name	content
Multispectral Image Products (MSI)	This product is processed with sensor and geometric correction. Pixel resolution is 2.5 m in panchromatic image and 5.0 m in multispectral image. Positional accuracy is aimed within 10-meter RMSE range. Relative/absolute radiometric correction is also applied. The pixel value for this data is obtained by scaling the reflectance at the top of the atmosphere to an unsigned 16-bit integer value.
Surface Reflectance Products (SR)	This product possesses the same attributes as the Multispectral Image Products but with the addition of atmospheric correction applied to each pixel value.
True Color Image Products (TCI)	This product is processed with sensor, geometric correction, and pansharpening. Pixel resolution is 2.5 m. Positional accuracy is aimed within 10-meter RMSE. For visual enhancement purposes, this product has also been post-processed with color-correction technique. A pixel value is scaled into an unsigned 8-bit integer.

The following figure shows pixels marked by an UDM file.

TCI	Invalid data area (pink)	Cloud area (blue)

3.2. Multispectral Image Products (MSI)

Multispectral Image Products consist of panchromatic: has sensitivity in wide wavelength and 2.5m resolution, and multispectral: has sensitivity at different wavelength and 5 m resolution, bundled imaging products. In this image product, artefacts such as pattern noise caused by all sensors are corrected. The pixel value is then converted into Top-of-Atmosphere (TOA) reflectance values.

3.2.1. Multispectral Image Product specification

The Multispectral Image Products contains the datasets shown in Table 3.

Table 3. Multispectral Image Product specification

Product attribute	Content	Description
Product configuration and format	Panchromatic imagery file	16-bit panchromatic imagery data that is divided by AxelGlobe cell. GeoTIFF format or JPEG2000 format. Layer assignment: Layer1: band0 (Panchromatic)
	Multispectral imagery file	16-bit multispectral imagery data that is divided by AxelGlobe cell. GeoTIFF format or JPEG2000 format. Layer assignment: Layer1: band1 (Blue) Layer2: band2 (Green) Layer3: band3 (Red) Layer4: band4 (Red Edge) Layer5: band5 (Near Infrared)
	Unusable Data Mask (UDM) file	Binary image data that indicates the unusable region. GeoTIFF format or JPEG2000 format. The pixel resolution matches that of the product. Layer assignment: Layer1: No data region Layer2: cloud covered region Digital number assignment: Layer1: Invalid data area: 1 Valid data area: 0 Layer2: Cloud area: 1 Cloud free area: 0
	Metadata file	Metadata for panchromatic imagery file, multispectral imagery file, and unusable data mask file. JSON format.
Product frame	AxelGlobe Cell	Our global grid system is based on the AxelGlobe Cell measuring about 5 km x 5 km per tile. A buffer of about 100m (overlap area with adjacent cells) is added to each of the four sides. If the captured image does not fill the entire area of the cell, the unfilled area will be indicated with black pixels wherein No data value (0) is assigned.
Pixel resolution	2.5 m (Panchromatic) 5.0 m (Multispectral)	The actual ground sampling distance varies on the collection conditions (mainly due to the satellite roll angle), but a sampling process is applied to the specified pixel size when product is generated.
Bit depth	16-bit unsigned integer	GRUS satellite images are stored in the onboard memory as up to 12 quantized bits. Relative DN values obtained directly from the sensor are converted to reflectance value at the top of atmosphere by radiometric correction, and the result is scaled (10,000 times) to a 16-bit dynamic range. (* Refer to 3.2.4 Radiometric correction)
Resampling	Cubic Convolution	Performs resampling processing using the Cubic Convolution interpolation method, which uses the weighted average value calculated from the values of the 16 input cells around the output cell. This sampling method has the characteristic that the boundaries of data tend to be sharp.
Geometric correction	sensor telemetry	Internal detector geometric correction combining the virtual array with eight imagers
	band-to-band registration	Band-to-band registration correction used to rectify misalignment between bands
	Geo-rectification	In geo-rectification, the characteristic terrain and features are extracted and paired between the captured image and the existing satellite image data (base map). The extracted terrain and features are matched with the reference points of each base map corresponding to the reference points. Processing is performed to reduce the positional deviation of the entire image with respect to the base map.
Radiometric correction		Correction of relative value difference of radiometric characteristics between detectors Undetected detector filling for null values from unresponsive detectors Conversion to reflectance values at the top of atmosphere based on calibration factors Scaling reflectance values at the top of atmosphere to 16-bit unsigned integer values (10,000 times)
Horizontal geodetic system	WGS1984
Projection	Universal Transverse Mercator projection (UTM)	UTM is a transverse Mercator projection that divides the earth into longitude zones every 6° eastward from the 180° meridian and projects the central meridian as the central meridian from the spheroid directly into the plane for each longitude band. Zones are determined by the south-west corner coordinates at the start of capture.

3.2.2. File naming

Multispectral Image Products are named according to the following naming convention. The file name contains several components that help organize, categorize, and refer to the information contained in the product.

<SatelliteName>_<AcquisitionDatetime>_<ProductLevel>_<ImageType>_<CellID>.<Extention>

Table 4. File naming convention

File naming component	Description
<SatelliteName>	Satellite name abbreviation (e.g. : GRUS-1A = GRUS1A)
<AcquisitionDatetime>	The collection start date and time in the yyyymmddhhmmss description format. （e.g.：Jan. 2, 2018 11:32:54 = 20180102113254）
<ProductLevel>	L1C
<ImageType>	Image type	Description
	Panchromatic	PAN
	Multispectral	MSI
	Unusable mask file for panchromatic	PAN_UDM
	Unusable mask file for multispectral	MSI_UDM
<CellID>	Unique ID of AxelGlobe Cell
<Extention>	Extension of imagery file（e.g.：GeoTIFF：tif）

3.2.3. File contents

Details of the Multispectral Image Products are provided in the file contents shown below.

EULA_en.txt                                             # License agreement (English)
EULA_jp.txt                                             # License agreement (Japanese)
GRUS1A_20200811011052
├─ GRUS1A_20200811011052_L1C_MSI_N42092354.tif          # Multispectral imagery
├─ GRUS1A_20200811011052_L1C_MSI_N42092355.tif          # Multispectral imagery
├─ GRUS1A_20200811011052_L1C_MSI_metadata.json          # Meta data for multispectral imagery
│
├─ GRUS1A_20200811011052_L1C_MSI_UDM_N42092354.tif      # UDM for multispectral imagery
├─ GRUS1A_20200811011052_L1C_MSI_UDM_N42092355.tif      # UDM for multispectral imagery
├─ GRUS1A_20200811011052_L1C_MSI_UDM_metadata.json      # Meta data for UDM of multispectral imagery
│
├─ GRUS1A_20200811011052_L1C_PAN_N42092354.tif          # Panchromatic imagery
├─ GRUS1A_20200811011052_L1C_PAN_N42092355.tif          # Panchromatic imagery
├─ GRUS1A_20200811011052_L1C_PAN_metadata.json          # Meta data for panchromatic imagery
│
├─ GRUS1A_20200811011052_L1C_PAN_UDM_N42092355.tif      # UDM for panchromatic imagery
├─ GRUS1A_20200811011052_L1C_PAN_UDM_N42092354.tif      # UDM for panchromatic imagery
└─ GRUS1A_20200811011052_L1C_PAN_UDM_metadata.json      # Meta data for UDM for panchromatic

3.2.4. Radiometric correction

To ensure the radiometric quality of GRUS satellite image products, we regularly observe calibration sites around the world and perform relative/absolute radiometric calibration.

All GRUS satellite images are collected in 12-bit and stored in the satellite's onboard memory as an image data with up to 12 quantized bits. It is scaled to a 16-bit dynamic range by radiometric correction in the ground processing system. With this, the relative pixel DN value obtained directly from the sensor is converted to the reflectance at the top of the atmosphere (TOA) by absolute radiometric correction and stored as a 16-bit value after scaling.

The pixel value of the Multispectral Image Products can be converted to the TOA reflectance by the following formula.

      REF(i)=DN(i)* ScaleFactor(i)
      where ScaleFactor(i) = 0.0001

 where
 i: band id
 REF：TOA reflectance
 DN：pixel value

Besides, TOA reflectance(REF) can be converted to the TOA radiance by the following formula:

      RAD(i) = REF(i) * (ESUN * cos(90 - solar_elevation_angle)) / (PI * (sun_earth_distance^2))
 where
 i: band id
 REF: TOA reflectance
 RAD: TOA radiance

3.3. True Color Image Products (TCI)

The True Color Image Product is a product consisting of AxelGlobe Cell, measuring about 5 km × 5 km per tile. This product is intended for applications that require accurate location information and map projection images. The True Color Image Product is processed with sensor correction, geometric correction, and map projection. In addition to automatically extracting the ground control points (GCP) from the high positional accuracy base map using our unique matching technology, we also combine conventional methods for areas where the GCP extraction is difficult. In applying correction processing, the location accuracy of the product depends on the GCP used and varies from region to region.

True Color Image Products are tone-corrected and pansharpened imagery products optimized for visual applications. This image product corrects artefacts such as pattern noise caused by all sensors while pixel values are corrected for visual enhancement purpose. A file that indicates areas unsuitable for analysis is attached as a usable data mask file (UDM file). The UDM marks pixels with missing data, fall outside the captured image boundary, or are obscured by clouds.

3.3.1. True Color Image Product Specification

The True Color Image Product contains the corresponding datasets delineated in Table 5.

Table 5. True Color Image Product specification

Product attribute	Content	Description
Product configuration and format	Pan-sharpened imagery file	8-bit color-scaled pan-sharpened imagery data that is divided by AxelGlobe cell. GeoTIFF format or JPEG2000 format. Layer assignment: Layer1: band3 (Red) Layer2: band2 (Green) Layer3: band1 (Blue)
	Unusable Data Mask (UDM) file	Binary image data that is indicative of the unusable region. GeoTIFF format or JPEG2000 format.The pixel resolution matches that of the product. Layer assignment: Layer1: No data region Layer2: cloud covered region Digital number assignment: Layer1: Invalid data area: 1 Valid data area: 0 Layer2: Cloud area: 1 Cloud free area: 0
	Metadata file	Metadata for pan-sharpened imagery file and unusable data mask file. JSON format
Product frame	AxelGlobe Cell	Our global grid system is based on the AxelGlobe Cell measuring about 5 km x 5 km per tile. A buffer of about 100m (overlap area with adjacent cells) is added to each of the four sides. If the captured image does not fill the entire area of the cell, the unfilled area will be indicated with black pixels wherein No data value (0) is assigned.
Pixel resolution	2.5 m	The actual ground sampling distance varies on the collection conditions (mainly due to the satellite roll angle), but a sampling process is applied to the specified pixel size when product is generated.
Bit depth	8-bit unsigned integer	GRUS satellite images are stored in the onboard memory as up to 12 quantized bits. Relative DN values obtained directly from the sensor are tone-scaled to 8-bit for visual purpose.
Resampling	Cubic Convolution	Performs resampling processing using the Cubic Convolution interpolation method, which uses the weighted average value calculated from the values of the 16 input cells around the output cell. This sampling method has the characteristic that the boundaries of data tend to be sharp.
Geometric correction	sensor model	Optical distortion correction due to sensor optics
	band-to-band registration	Band-to-band registration correction used to rectify misalignment between bands
	Geo-rectification	In geo-rectification, the characteristic terrain and features are extracted and paired between the captured image and the existing satellite image data (base map). The extracted terrain and features are matched with the reference points of each base map corresponding to the reference points. Processing is performed to reduce the positional deviation of the entire image with respect to the base map.
Radiometric correction		Correction of relative value difference of radiometric characteristics between detectors Undetected detector filling for null values from unresponsive detectors Scaling a 12-bit relative DN value to an 8-bit unsigned integer value by gamma correction
Enhancement	Dark Object Subtraction (DOS)	The dark pixels in the image are used to reduce noise such as water vapor and dust in the atmosphere. (Applies to images taken after July 7, 2020)
Horizontal geodetic system	WGS1984
Projection	Universal Transverse Mercator projection (UTM)	UTM is a transverse Mercator projection that divides the earth into longitude zones every 6° eastward from the 180° meridian and projects the central meridian as the central meridian from the spheroid directly into the plane for each longitude band. Zones are determined by the south-west corner coordinates at the start of capture.

3.3.2 File naming

True Color Image Products are named according to the following naming convention. The file name contains several components that help organize, categorize, and refer to the information contained in the product.

<SatelliteName>_<AcquisitionDatetime>_<ProductLevel>_<ImageType>_<CellID>.<Extention>

Table 6. File naming convention

File naming component	Description
<SatelliteName>	Satellite name abbreviation (e.g. : GRUS-1A = GRUS1A)
<AcquisitionDatetime>	The collection start date and time in the yyyymmddhhmmss description format. （e.g.：Jan. 2, 2018 11:32:54 = 20180102113254）
<ProductLevel>	L1C
<ImageType>	Image type	Description
	Pan-sharpen	PSM
	Unusable mask file for Pan-sharpen	PSM_UDM
<CellID>	Unique ID of AxelGlobe Cell
<Extention>	Extension of imagery file（e.g.：GeoTIFF：tif）

3.3.3. File contents

Details of the True Color Image Products are provided in the file contents shown below.

EULA_en.txt                                             # License agreement (English)
EULA_jp.txt                                             # License agreement (Japanese)
GRUS1A_20200811011052                                   
├─ GRUS1A_20200811011052_L1C_PSM_N42092354.tif          # Pan-sharpened imagery
├─ GRUS1A_20200811011052_L1C_PSM_N42092355.tif          # Pan-sharpened imagery
├─ GRUS1A_20200811011052_L1C_PSM_metadata.json          # Pan-sharpened imageryMeta data
│
├─ GRUS1A_20200811011052_L1C_PSM_UDM_N42092354.tif      # UDM for pan-sharpened imagery
├─ GRUS1A_20200811011052_L1C_PSM_UDM_N42092355.tif      # UDM for pan-sharpened imagery
└─ GRUS1A_20200811011052_L1C_PSM_UDM_metadata.json      # Meta data for UDM for pan-sharpened imagery

3.4. Metadata

The description of product metadata is shown in Table 7

Table 7. Metadata Description

Section	Field 1	Field 2	Data type	Description
Company
	companyName		string	Company name
	companyLocation		string	Main office location
	contact		string	Contact number
	copyright		string	Copyright
	processorVersion		string	Software version of processor
	bundlerVersion		string	Software version of data delivery
	processDate		string	Product processing date
EOMetadata
	satelliteID		string	Unique ID assigned to the satellite
	satelliteName		string	Official satellite name
	orbitType		string	Types of satellite orbit
	orbitInclinationAngle		double	Inclination angle of satellite orbit
	orbitDirection		string	Direction the satellite in orbit is heading at the time of image acquisition
	captureID		string	Unique ID corresponding to a capture event
	acquisitionDateTime
		acquisitionStartDateTime	date	Capture start date and time of the capture event. Describe it in ISO8601 extended format with delimiter
		acquisitionEndDateTime	date	Capture end date and time of the capture event. Describe it in ISO8601 extended format with delimiter
	satelliteOffNadirAngleNominal		double	Nominal value of the off-nadir (degree) of the satellite at the time of the capture event
	satelliteElevationAngleNominal		double	Nominal value of the elevation angle (degree) of the satellite at the time of the capture event
	satelliteAzimuthAngleNominal		double	Nominal value of the azimuth angle (degree) of the satellite at the time of the capture event
	solarElevationAngleNominal		double	Nominal value of the sun elevation angle(degree) at the collection event
	solarAzimuthAngleNominal		double	Nominal value of the sun azimuth angle at the capture event
	panchromaticRowGSDNominal		double	Nominal value of ground sampling distance (GSD) at the row direction at the panchromatic sensor. Unit: meter
	panchromaticColumnGSDNominal		double	Nominal value of ground sampling distance (GSD) at the column direction at the panchromatic sensor. Unit: meter
	multispectralRowGSDNominal		double	Nominal value of ground sampling distance (GSD) at the row direction at the multispectral sensor. Unit: meter
	multispectralColumnGSDNominal		double	Nominal value of ground sampling distance (GSD) at the column direction at the multispectral sensor. Unit: meter
	GSDUnits		string	Unit of GSD
	resamplingAlgorithm		string	Resampling method used in production Processing: Cubic Convolution
	radiometricCalibrationApplied		Boolean	Yes, No
	earthSunDistance		double	Distance between the center of the sun and the center of the earth Unit: Astronomical Unit[AU]
	absoluteGain
		Panchromatic Blue Green Red Red Edge Near Infrared	double	Gain parameter when converting the observed DN value into the TOA radiance [W⋅sr-1 ⋅m-2] Listed for each band
	ESUN
		Panchromatic Blue Green Red Red Edge Near Infrared	double	Extra-terrestrial solar irradiation Unit: [mW/m2]
productMetadata
	pansharpen		Boolean	Yes, No
	bitsPerPixel		string	1U, 8U, 16U
	compressionType		string	NONE / DEFLATE / JPEG2000
	outputFormat		string	GeoTIFF/JPEG2000
	spatialReferenceSystem
		EPSGCode	Integer	EPSG code related to image geodetic system and projection
		geodeticDatum	String	geodetic datum
		projection	String	projection
		projectionZone	String	projection zone
	layerConfiguration
		layerX	String	[Satellite image data] Describes the band allocation in the Xth layer [Other attached data] Describes the data type in the Xth layer
	valueInterpretation			For the following items, only unusable mask data are described
		layerX	String	Described the pixel value interpretation in the Xth layer
imageTileMetadata				Repeat for the number of image files
	imageName		string	image file name
	cellID		string	Unique ID of AxelGlobe Cell
	imageLocation			Tile range information units are based on product projection
		type	string	Polygon
		coordinates	list	Describes the coordinate values (X coordinate, Y coordinate) that make up the tile boundary in the order of bottom left, top left, top right, bottom right, bottom left
	imageLocationUnits		string	Coordinates units
	rowGSD		double	The pixel size in the row direction of the tile. Units are based on product projection
	columnGSD		double	The pixel size in the column direction of the tile. Units are based on product projection
	GSDUnits		string	Unit of GSD
	numberRows		Integer	Number of lines (pixels) in the image
	numberColumns		Integer	Number of columns (pixels) in the image
	numberBands		Integer	Number of bands in the image
	cloudCoverPercentage		double	Cloud cover rate Unit : % only satellite image data are described

3.5. Surface Reflectance Product

3.5.1. Surface Reflectance Product specification

The Surface Reflectance product is derived from the Multispectral image product with the addition of atmospheric correction which is used to convert top of atmosphere (TOA) reflectance values into bottom of atmosphere (BOA) reflectance. The intention of this product is to remove the impact of the atmosphere from image products in order to improve the accuracy of reflectance when compared with other datasets or actual surface reflectance values.

The specification of this product is the same as the specification for the Multispectral image product however in this case the reflectances values are bottom of atmosphere instead of top of atmosphere.

3.5.2. File naming

Surface Reflectance Image Products are named according to the following naming convention. The file name contains several components that help organize, categorize, and refer to the information contained in the product.

<SatelliteName>_<AcquisitionDatetime>_<ProductLevel>_<ImageType>_<CellID>.<Extention>

Table 4. File naming convention

File naming component	Description
<SatelliteName>	Satellite name abbreviation (e.g. : GRUS-1A = GRUS1A)
<AcquisitionDatetime>	The collection start date and time in the yyyymmddhhmmss description format. （e.g.：Jan. 2, 2018 11:32:54 = 20180102113254）
<ProductLevel>	L2A
<ImageType>	Image type	Description
	Panchromatic	PAN
	Multispectral	MSI
	Unusable mask file for panchromatic	PAN_UDM
	Unusable mask file for multispectral	MSI_UDM
<CellID>	Unique ID of AxelGlobe Cell
<Extention>	Extension of imagery file（e.g.：GeoTIFF：tif）

3.5.3. File contents

Details of the Surface Reflectance Image Products are provided in the file contents shown below.

EULA_en.txt                                             # License agreement (English)
EULA_jp.txt                                             # License agreement (Japanese)
GRUS1A_20200811011052
├─ GRUS1A_20200811011052_L2A_MSI_N42092354.tif          # Surface Reflectance imagery
├─ GRUS1A_20200811011052_L2A_MSI_N42092355.tif          # Surface Reflectance imagery
├─ GRUS1A_20200811011052_L2A_MSI_metadata.json          # Meta data for surface reflectance imagery
│
├─ GRUS1A_20200811011052_L2A_MSI_UDM_N42092354.tif      # UDM for surface reflectance imagery
├─ GRUS1A_20200811011052_L2A_MSI_UDM_N42092355.tif      # UDM for surface reflectance imagery
├─ GRUS1A_20200811011052_L2A_MSI_UDM_metadata.json      # Meta data for UDM of surface reflectance imagery
│
├─ GRUS1A_20200811011052_L2A_PAN_N42092354.tif          # Panchromatic imagery
├─ GRUS1A_20200811011052_L2A_PAN_N42092355.tif          # Panchromatic imagery
├─ GRUS1A_20200811011052_L2A_PAN_metadata.json          # Meta data for panchromatic imagery
│
├─ GRUS1A_20200811011052_L2A_PAN_UDM_N42092355.tif      # UDM for panchromatic imagery
├─ GRUS1A_20200811011052_L2A_PAN_UDM_N42092354.tif      # UDM for panchromatic imagery
└─ GRUS1A_20200811011052_L2A_PAN_UDM_metadata.json      # Meta data for UDM for panchromatic

3.5.4. Atmospheric correction

The process of atmospheric correction and conversion of top of atmosphere to bottom of atmosphere reflectance is carried out using the 6S radiative transfer equations and atmospheric data from the MODIS sensors. Lookup tables are generated using 6S and applied to the MSI product to perform the correction.

For more details, please see the Surface Reflectance product guide.