Very High Resolution Optical Satellites: An Overview of the Most Commonly used

Email: pasquale.maglione@uniparthenope.it Abstract: Very High Resolution (VHR) satellite systems are platforms whose sensors acquire high geometric resolution images. Since 1972, when the first satellite was launched (Land sat ERTS), the spatial resolution of the satellite image has increased, making Ground Sample Distance (GSD) reaching 0.30 m at Nadir in panchromatic images. In this paper, after a brief introduction, concepts relative to orbits, types of sensors and resolutions are reported. Geostationary and sun-synchronous orbits are described; difference between push-broom and whisk-broom sensors are reported; the definition of the geometric, the radiometric and the temporal resolutions are listed. In the end, the characteristics of the most common VHR commercial optical satellite are mentioned: IKONOS-2, QuickBird-2, SPOT-5, GeoEye-1, WorldView-2 and WorlView-3 satellites.


Introduction
Commercial development in remote sensing has opened new possibilities for satellite imagery. In fact, in the late 1990s a new generation of high-performance sensors were launched in order to acquire two types of images: Multispectral and panchromatic. The former ones have a better spectral rather than a spatial resolution and they have a smaller band width. The panchromatic images have a larger band width and a higher geometric resolution. This new generation of sensors are able to take images with spatial resolution from 0.3 to 1 m in panchromatic mode and 2.5 to 4 m in multispectral one (Fritz, 1996). Very High Resolution (VHR) is the acronym distinguishing these satellites equipped with sensors that acquire with metric and sub-metric spatial resolution (Schreier and Dech, 2005).
In these years several experimentations have been carried out about the use at VHR images into geomatics applications. Tests were made to evaluate the use of the VHR images in order to generate middle scale and large scale mapping and for cover/land use identification (Li, 1998;Samadzadegan et al., 2002;Chmiel et al., 2005;Abdelwahed et al., 2011;Fernàndez et al., 2014;Mostafa et al., 2014). Orthorectification is pivotal in these applications because it permits to ensure a consistent geometric relationship between points on the ground and their corresponding representation on the image (Campbell and Wynne, 2011). Indeed, the terrain relief, the perspective and the motion of the sensor optics and the earth's curvature generate position errors on each image. Therefore, with orthorectification the image is made planimetric and terrain relief errors are corrected or reduced. Distinguishes two approaches for geometric correction applied to satellite images (Toutin, 2006): Rigorous (Bonneval, 1972;Wong, 1980) and synthetic models (Toutin, 2004;Tao and Hu, 2001;Dowman and Dolloff, 2000). In VHR satellite images orthorectification, synthetics models are widespread because, unlike the rigorous models, it is not necessary to know orbital parameters. Tests on several VHR images were conducted to validate the use of synthetics models (Hu and Tao, 2002;Amini and Hashemi, 2005;Errico et al., 2009;Maglione et al., 2013;Belfiore and Parente, 2014).
The increase of the number of spectral bands is another positive aspect linked to VHR satellite that permits a further improvement in classification. WorldView-2 multispectral sensor acquires on 8 bands in the visible range (Geoimage, 2015a). Coastal band, for example, travels farther into the water column without absorption and it is useful in the study of the seabed classification (Proteus, 2015). Coastal band, as well as Blue band and others, are used to define coastline by Normalize Difference Water Index (NDWI) (McFeeters, 1996), too (Basile Giannini et al., 2011;Maglione et al., 2014;2015). Image Fusion methods (Van Genderen and Pohl, 1994), applied to VHR image, permit to obtain further improvements of the spatial resolution of multispectral images. In particular, pan-sharpening techniques allow to integrate the geometric detail of panchromatic data with the radiometric one of multispectral images (Zhang, 2004). Several applications about the use and the verify of pan sharpening techniques on VHR images were conducted in the last years (Zhou et al., 1998;Wang and Bovik, 2002;Tu et al., 2004;Aiazzi et al., 2007;Zhang, 2008;Kumar et al., 2009;Xin-Zhi et al., 2009;Shridhar and Alvarinho, 2013;Parente and Santamaria, 2014;Belfiore and Parente, 2015).
In Change Detection applications, the use of VHR images improves the thematic accuracy (Congalton, 2009) and permits of a better identification of the objects' contours. Maglione (2012), used IKONOS-2 and WorldView-2 images to detect new built by difference method. Pacifici et al. (2010) proposed a new approach based on Pulse-Coupled Networks (PCNNS) to evaluate the area of image where a significant change occurred. Dalla Mura et al. (2008) integrated the nonlinear and adaptive properties of the morphological filters with a Change Vector Analysis (CVA) applied to QuickBird images. Vettil (2012) described various urban change detection models using Quickbird and WorldView-2 images. This paper is structured into two parts. In the first one were reported the general characteristics about satellite systems and sensors. The principal characteristics of the most common VHR satellite platform were considered in the second part. At the end, brief conclusions were reported.

Brief History about Commercial Satellites
The first license to build satellite that acquires high spatial resolution digital imagery in USA for commercial sale, was released to WorldView Inc. of Longmont (actually named DigitalGlobe) in 1993 (Kramer, 2002). The first commercial spaceborne that collects images with spatial resolution of 1 m in panchromatic mode using Charge Coupled Device (CCD) array was launched in 1999 from the Space Imaging Inc. of Thornton and named IKONOS-2 (Kramer, 2002). In 2001, Digital Globe Inc. launched QuickBird-2 satellite that acquires images with spatial resolution of 0.62 m in panchromatic mode (Kramer, 2002). WorldView-1 was the successor of QuickBird-2 and was launched from DigitalGlobe in 2007. It acquires panchromatic image with spatial resolution of 0.5 m. In 2008, GeoEye Inc., launched GeoEye-1 satellite with a resolution of 0.41 m in panchromatic mode. In the last years, DigitalGlobe launched another two VHR satellite: WorldView-2 in 2009 and WorldView-3 in 2010. They have, respectively, 0.41 and 0.31 m as spatial resolution.

Orbits
Orbits characteristics are described in relation to the assumption that the Earth's gravitational field is spherical, for simplicity (Campbell, 2006). These types of orbits are called normal and they are ellipses with ( Fig. 1): An Apogee (A), the point farthest from the Earth; a Perigee (P), the point closest to the earth; an Ascending Node (AN), the point where the satellite crosses the equator moving south to north; a Descending Node (DN), the point where the satellite crosses the equator passing north to south. The inclination of the orbit is defined in DN: The dihedral angle between the plane of the orbit and the equator one.
It is important to remember that the time required for a satellite to complete one orbit increases with the altitude.
Two principal grouping of satellites are recognized with reference to the types of orbit: Geostationary and sun-synchronous.
The former is a circular orbit that is located approximately at a distance of 35,800 km above the earth's equator; the satellite turns around the polar axis of the earth with the same direction and it moves on equatorial plane; the platform, in this way, has the same period of the earth and it is stationary with respect to the terrestrial surface observed (Grove, 1979;Montenbruck and Gill, 2000) Fig. 2. This type of orbit is frequently employed in communications and meteorology (Reeves, 1983).
Sun-synchronous orbit (Fig. 3) is obtained by the combination of orbital period and the inclination such that the satellite keeps pace with the sun's westward progress as the earth rotates; inclination is close to 90°, for this reason the orbit is called near-polar, indeed satellites pass near the north and south poles in each orbit (Lillesand et al., 2004). A satellite that moves along a sunsynchronous orbit acquires each part of the earth within its view at the same local sun time each day, so as to reduce variation in illumination during images acquisition (Campbell, 2006). The orbit altitude is about 700 km.
The time that satellite spends to complete one orbit, called orbital period, is defined, in the case of a circular orbit, as (Elachi, 1987): Where: T = The orbital period in sec R = The planet radius in km (about 6,380 km for the Earth) h = The orbit altitude in km evaluated from the surface of the Earth g s = The gravitational acceleration at the Hearth's surface (0.00981 km/sec 2 ) Types of Sensors VHR optical satellites described in this article have a push-broom sensor equipped with Charge Coupled Device (CCD) linear array detectors. Whisk-broom is another type of sensor that is mounted in the oldest satellite. Push-broom (Shippert, 2013). A scanner with a line of detectors, called array, acquires in perpendicular to the flight direction of the spacecraft and the image is collected one line at a time; all the pixels in a line are acquired simultaneously (Fig. 4). Push-broom sensor is also called along track scanners.
Whisk-broom (Shippert, 2013). The sensor is a scanner. It uses a mirror to collect light onto a single detector. The detector moves back and forth and it acquires measurements from one pixel in the image at a time. Whisk-broom sensor is also called across track scanners (Fig. 5).

Resolutions
Four types of attributes are very important between sensors characteristics. They are referred to the capability to describe a scene in terms of: Geometrical detail, radiometric sensitivity, spectral content and temporal frequency of the acquisition (Brivio et al., 2006). Rees (1999) defined geometric or spatial resolution as a measure of the smallest distance between two objects which can be distinguished by a sensor. It depends on both sensor characteristics and orbit height. In fact, from Fig. 6 it is possible to define spatial resolution "D" at nadir as:

Geometric Resolution
Where: d = Linear dimension of the single detector of sensor h = Height of sensor f = Focal length of the optical system Height of sensor "h" can be considered constant when satellite platform is considered. In this case, geometric resolution is defined by Instantaneous Field of View (IFOV- Fig. 6): The cone angle within incident energy is focused on the detector (Lillesand et al., 2004). The projection on ground of IFOV is called Groundprojected Instantaneous Field of View (GIFOV) or Ground Sample Distance (GSD) (Miller et al., 2005). Field of View (FOV), on the other hand, is the full crosstrack angular coverage to which corresponds a ground projection called Ground-projected Field of View (GFOV) or Swath Width or footprint of the sensor (Schowengerdt, 2007).

Radiometric Resolution
Jensen defines it as the sensitivity of a sensor to difference in signal strength as it records the radiant flux reflected, emitted, or back-scattered from the terrain (Jensen, 2005). This sensitivity is measured as the intensity level and it is commonly defined by the number of bits used to quantize it (Gonzalez and Woods, 2008). Typical values of radiometric resolution are: 8, 11 or 16 bits.

Spectral Resolution
Spectral resolution refers to the resolving power of a system in terms of wavelength (Kramer, 2002). It is determined by the bandwidth of the channel used (Reeves, 1983), where bandwidth is the spectral interval used to record brightness in relation to wavelength (Campbell, 2006). Panchromatic sensors acquire on a spectral bandwidth which is greater than the multispectral ones.

Temporal Resolution
The temporal resolution, in remote sensing, generally refers to how often the sensor records imagery of a particular area (Jensen, 2005). Sensors that acquire images with close intervals have a fine temporal resolution; in contrast, the ones with a long interval have a coarse resolution (Campbell, 2006).

IKONOS-2
IKONOS-2 was launched on September 1999 (DigitalGlobe, 2015a). The orbital altitude is 681 km and the orbital inclination is 98°.1. The swath is 11.3 km, at nadir. The orbital period is 100 min and the revisiting one is about 3 days, that decreases to 1.5 days due to the possibility of tilting the push-broom sensor.

QuickBird-2
QuickBird-2 satellite was launched in October 2001. Initially, its orbital height was about 600 km; after, DigitalGlobe opted for a fairly low orbital altitude of 450 km (ESA, 2000(ESA, , 2015. In April 2011 the orbit satellite was raised to 482 km to give the spacecraft an extended life. In 2013, the altitude achieved was of 450 km (DigitalGlobe, 2015b). Currently, the satellite works at about 400 km. The plane of the orbit has an inclination of 98°. The orbital period is 93.6 min at 450 km and 92.4 min at 400 km; the revisit time varies from 2 to 12 days depending target location as the orbit decays; the swath, at nadir, is 16.8 at 450 km and 14.9 at 400 km (DigitalGlobe, 2015c).
Radiometric resolution for both panchromatic and multispectral images is 11 bits.

SPOT-5
Spot Image launched SPOT-5 satellite on 4 May 2002. The height of the orbit is 822 km and the orbital period is 101.4 min; the revisited time is 26 days.
SPOT-5 satellite has two push-broom sensors: High Resolution Geometric (HRG) and High-Resolution Stereoscopic (HRS). The former acquires both panchromatic and multispectral images; the latter only panchromatic images used to generate Digital Terrain Model (DTM) with geometrical resolution of 10 m. The swath is 60 km for HRG and 120 km for HRS (ADS, 2015).
Spectral interval for the panchromatic images, acquired to HRG, is 0.480-0.710 µm and its spatial resolution is 5 m or 2.5 m in SUPERMODE modality: Two panchromatic images acquired simultaneously with spatial resolution of 5 m and an offset of half pixel in the focal plane are processed by interpolation processes (Guastaferro et al., 2012).
For all described images the radiometric resolution is 8 bits. In March 2015, SPOT-5 mission ended (ESA, 2015).

GeoEye-1
GeoEye-1 satellite was launched in September 2008. Its orbital height is 681 km and the plane of the orbit has an inclination of 98°. The orbital period is 98 min and the average revisited time is less than 3 days. The swath, at nadir, is 15.3 km. The satellite has two push-broom sensors that acquire panchromatic and multispectral images (DigitalGlobe, 2015d).
Spatial resolution, at nadir, is: 0.41 m for panchromatic images, resampled to 0.50 m for commercial scope; 1.64 m for multispectral ones, upsampled to 2 m for commercial scope.
Radiometric resolution for both panchromatic and multispectral images is 11 bits.

WorldView-2
WorldView-2 satellite was launched in October 2009 by DigitalGlobe (2015e). The orbit plane has an inclination of 97°0.2 and satellite is placed at an altitude of 770 km; the orbit period is 100 min. The revisited frequency is 1.1 days at nadir and 3.7 days at 20° offnadir. The swath width, at nadir, is 16.4 km. WorldView-2 satellite has two push-broom scanner that acquire: Panchromatic and multispectral images.
The former has a spatial resolution of 0.46 m, resampled to 0.50 m for commercial scope; the spectral interval is 0.450-0.800 µm.
Radiometric resolution is 11 bits for both panchromatic and multispectral images.

WorldView-3
It is a product of the DigitalGlobe and was launched in August 2014 (DigitalGlobe, 2015f). Its orbit has an inclination of 98° and is placed at an altitude of 617 km. The WorldView-3 satellite has an orbital period of 97 min and the revisited frequency is less than 1 day at nadir and 4.5 days at 20° off-nadir. The swath, at nadir, is 13.1 km.
WorldView-3 satellite has two push-broom scanners that acquire panchromatic and multispectral images. It also mounts a CAVIS sensor.
Panchromatic images have a spatial resolution of 0.31 m, resampled to 0.40 m for commercial scope; the spectral interval is 0.450-0.800 µm.

Conclusion
The aim of this paper is to collect principal information about characteristics of the most popular VHR optical commercial satellite.
In the future, the evolution in aerospace technologies will open new frontiers in the field of remote sensing and its applications. For example, the increasing spatial resolution of the images allows for greater geometric detail in acquisitions and therefore an ever increasing ability of interpretation of the territory. Capacity that already today is particularly advanced, just consider that the size of the pixels of the image acquired by the satellite panchromatic WorldView-3 is 0.31×0.31 m.