Abstract With the recent launch of enhanced high-resolution commercial satellites, available imagery has improved from four-bands to eight-band multispectral. Simultaneously developments in remote sensing processing techniques have enabled water depths to be derived from these improved imagery sources with better confidence, vertical accuracy and resolution. These emerging techniques of remotely bathymetric surveying, coupled with the extensive coverage of satellite images has resulted in hydrographic data being delivered to customers within weeks, with many logistical and environmental advantages over traditional surveying techniques. This paper explores these new techniques and significant developments and the impacts these will have on hydrographic surveying. Introduction. Hydrographic data from remote sensing has been an option to the marine community for many years. It has only been with the recent launch of enhanced high resolution satellites and the coordination of the remote sensing and hydrographic communities that this emerging technology is now an improved alternative for the hydrographic surveyor. Proteus has been leading the way forward and has been bringing satellite derived marine and terrestrial data products to the geospatial market. This paper intends to explore the developments experienced by Proteus and outlines the benefits satellite derived bathymetry (SDB) brings to the marine community. Recent Developments in Satellite Imagery. The remote sensing world have been able to extract Soundings from satellite imagery for many years, however the launch of DigitalGlobe s WorldView-2 satellite has meant the resolution of multispectral imagery has increased to 2m pixels thus giving a sounding every 2m in x,y directions. Based in Longmont, Colorado, DigitalGlobe expanded its capabilities in the commercial remote sensing market earlier this year with the combination with GeoEye. It now has five optical imaging satellites at its disposal: WorldView-1 and -2, GeoEye-1, QuickBird and IKONOS. A third WorldView satellite is planned for launch in autumn 2014. WorldView-2 represents a departure from the multispectral capabilities of other mainstream commercial imaging satellites. It is the first and only high resolution 8 band multispectral satellite with a panchromatic resolution of 46cm and 1.85M multi-spectral resolution. On top of the Red, Blue, Green and Infrared 1 bands you have the additional bands of Coastal Blue, Red-Edge, Visible Yellow and Near Infrared 2 bands which maximize feature classification both on land and in the coastal zone. This is a vast improvement from the previous satellite called QuickBird which has 4 multispectral bands. Figure 1 illustrates the developments in the number of these multispectral bands and their associated wave lengths.
Figure1: DigitalGlobe s QuickBird (QB), WorldView-1 (WV-1) and WorldView-2 (WV-2) multispectral band comparison. All eight bands are utilised for bathymetry and seabed classification extraction, however, the introduction of the coastal blue band starting at 400nm has significantly improved this technology and brought improved accuracy when deriving soundings and seabed type to the hydrographic sector. The Coastal Blue band has a relatively short wavelength, 400-450 nm, compared to the others. This wavelength enables it to travel farther into the water column without absorption, resulting in greater vertical accuracy in bathymetric measurements and more detailed seabed classification information. The major advantage of working with DigitalGlobe on these types of projects is its worldwide archive of data. The vast multispectral imagery back catalogue provides global coverage to the hydrographer. In just four years, WorldView-2 has collected a massive archive of image data which gives Proteus the opportunity to see what data is available and what water conditions typically exist in a particular geographic location before the firm even bids on a proposal. This lets the end user know the quality and accuracy of the deliverables before survey commences. The library can be viewed online and can be downloaded and assessed to understand whether any suitable imagery exists for depth extraction. At this point decisions are made on whether the imagery in the archive is viable, or whether we need to capture a new image. If new imagery is required dedicated tasking parameters are calculated and assigned to the survey, thus providing the best possible imagery for the survey. The angle of the satellite camera is an important variable when capturing an image for bathymetry applications, so the hydrographer needs to task the satellite to acquire images with a maximum incidence angle of 30 0 from nadir. If archive imagery is to be used then production can start straight away. Other considerations needed for tasking imagery include the cloud cover, turbidity and sea state. Normally these parameters are not associated with remote sensing and imagery capture, however the partnership of Proteus and DigitalGlobe ensures these important considerations are taken into account and thus provides an improved image for depth extraction.
The coming together of two sectors. The remote sensing community has been busy over a number of years developing and refining algorithms to derive bathymetric and seafloor classification data. It has then been the collaboration with the hydrographic sector that has allowed tailor-made products and services to be developed and marketed. Marine hydrographic clients expect more than a geo referenced image from SDB products. XYZ, seabed type and quality assessments are essential for end users to be able to apply the SDB products for their individual applications. Research and development on future products and services is very important for this rapidly improving technology. The principle of SDB involves the absorption and refraction of the 8 multi spectral bands through the atmosphere and water column. Correction factors need to be applied to counteract the absorption and refractions which occurs, as the main principle of the system is to calculate the seafloor reflectance for each individual pixel, before converting this into a sounding and seabed type data. This principle is continuously being improved as hydrographic knowledge and remote sensing communities are working together for better SDB data. Vertical Accuracy Vertical accuracy is continuously improving, as processing flow lines and imagery develop. One of the biggest improvements to the final vertical accuracy is when ground control points are used in conjunction with the imagery to help better determine reflectance and the relationship to water depth. Historical data is not always available and usually influenced by the location and project budget of each survey. Proteus has delivered SDB data when either no and limited ground control points have been used. Accuracies to date have been in the order of 10% of water depth, for example in 5m of water a vertical accuracy of +/- 0.5m. In some cases of ideal survey conditions then comparisons have been more favourable than this. One of the main issues SDB has to overcome is the effect of turbidity in the water column. Similar to LiDar systems, water column turbidity affects the accuracy and coverage of the retrieved depths. This is still an on-going issue with resources currently looking at how the affects of turbidity can be reduced. Tidal application is potentially another large contributor to the SDB error budget and is therefore accounted for and applied to all datasets. Depending upon the nature of the project then either actual tides or predicted tides are used. For environmentally sensitive and military surveys then a complete remote approach is needed with users willing to accept an increase in vertical uncertainties for the trade-off of no resources going to site. Benefits of SDB SDB provides many advantages to the marine sector and can contribute to various marine applications. Project costs are dramatically reduced when using SDB. A large percentage of the total costs from traditional surveying techniques are incurred from the mobilisation of equipment and personnel to the survey site. These cost savings are seen with SDB, as it is possible to work completely remotely.
No costs are incurred from expensive weather and equipment downtime charges, along with the logistics of mobilising all the resources to far away locations. SDB survey times are also improved with large areas of imagery being instantly captured. SDB can provide a swath of 18km wide and a length of many hundreds of kilometres. This provides an instant snapshot of the complete survey area and produces seamless coverage. Not all benefits of SDB are monetary, environment and health and safety advantages exist. Survey locations can be hostile locations, and encompass fragile ecosystems. With the use of satellites, these important environmental locations are untouched with zero risk of pollution and damage. Risks can occur when mobilising manned vehicles to the survey area. Vessels and manned aircraft operate in a changing environment with risks associated with the nature of the work they undertake. These risks can be mitigated by traditional hydrographic companies; however SDB completely eradicates these risks with the remoteness of its methodology. Environmental and health and safety risks are an obvious hidden expense for traditional methodologies and one which SDB benefit from not experiencing. Using the archive. With the extensive library of imagery from DigitalGlobe, a huge benefit to end users is being able to view the imagery before work commences. An indication of quality and coverage is communicated to clients at the initial proposal stages, with expectations from all parties approved. Timescales are dramatically reduced when using archive imagery, with very large survey areas in the order of several hundred square kilometres delivered within weeks. For environmental and engineering applications often an understanding of how the seabed has changed over time is needed. Baseline surveys are commonly requested from clients, and now using SDB and imagery from the archive then baseline surveys spreading over a specified timeframe can be easily produced from using multiple images. Figure 2 below summarises the advantages of SDB. Element Benefit Cost Typically costs are between one tenth and one twentieth of the cost of traditional survey methods Coverage A swath of 18 x 600Km can be captured within a few seconds, providing a seamless survey.
Element Benefit Time Satellite bathymetry surveys can be 10 times faster than traditional surveys Mobilisation No equipment or resources need to be mobilised creating large cost savings to the client Permits No Permits are required for boats or aircraft Baseline Survey Archive imagery provides imagery to create historical mapping, to support baseline mapping projects. Borders International borders are invisible to satellite data and do not affect the survey logistics or program. Environment no pollution created in the survey Ecosystem no risk to the marine ecosystem Health & safety no risk to the work force Figure2: Summary of the Advantages of Satellite Derived Bathymetry. Future developments for SDB DigitalGlobe already has WorldView-3 in the works with it due to be launched in the autumn of 2014 and it will have broader and more detailed imaging capabilities than its immediate predecessor. Plans call for adding eight more short-wave infrared bands and 10 atmospheric correction bands at 3.7-meter spatial resolution, which will yield an unprecedented amount of visible and invisible data. The pan band will be sharpened to 31 centimetres and the other eight multispectral bands will operate at 1.24 meter resolution. This step development in imagery will have a knock on effect on the accuracy and resolution of the SDB products Proteus is able to offer to the market place.