Replicas of world – a reality soon
Sales Director, Airborne Sensors
EMEA & Asia Leica Geosystems AG,
If Antoine de Saint-Exupéry was to be believed, our "task is not to foresee the future, but to enable it" (La Citadelle, 1948). As a manufacturer of airborne sensors, Leica Geosystems has a goal to enable LiDAR technologies of the future creation.
That apart, it also ensures that its commercialisation and advantages flow via the many aerial survey providers to the wider user community for value creation in each user's field. Some of what the service provider wants or needs are relatively predictable; efficiency, productivity and return on investment. However, for the end-user, it is clear we need to envision and create spatial information that adds value to any foreseeable situation or decision making process. The driver of LiDAR has until now been an ever increasing 'need for speed' in terms of pulse and scan rates (depending on the technology used) and this trend will certainly continue in the near future. Efficiency, measureable in cost per data point, is crucial in LiDAR data capture [Fig. 1]. Lower cost per point leads to new and market broadening applications, and more points 'purchased', more often, by the market as a whole. Cost reductions and productivity improvements via 'black-box' distributed processing on-board relatively small aircraft, including data filtering and 'hands-off' editing, will soon allow niche applications to buy the data required.
On-board processing begins to address our impatience as a consumer. Geospatial data has become a commodity and our paying customer expects commodities to be universally available on-demand. Happily, small area, highly timedependent data acquisition can be beneficial for the consumer and provider. The value of data is higher the quicker it is provided, and decays in price relatively rapidly. Highly temporal data can, to some extent, be forgiven some imprecision; but, extremely high density, highly accurate data will also be in demand. Terrestrial LiDAR data, be it mobile or static, can achieve centimetre and sub-centimetre levels. Aerially captured LiDAR points will continue to improve in achievable and repeatable accuracies over large areas. Mass LiDAR points with imagery overlay, or even LiDAR imagery by itself, from ground or airborne capture systems will create incredibly precise virtual replicas of the world [Fig. 2]. And that replica, whilst feeding growing demand, is also more easily understandable and digestible by the consumer market.
As airborne LiDAR and its accuracy is dependent on GNSS and IMU technologies, it is probable that differential GPS (DGPS), GNSS systems approaching 120 channels of GPS/Glonass/Galileo and Compass signals will appear. Also, GNSS/IMU tightly coupled solutions will further assist accuracies, with the added benefits of near instantaneous initialisation or re-acquisition, slightly improved accuracies and longer RTK baseline resolution. We have been fortunate as a manufacturer to see the number of terrestrial and aerial LiDAR sales grow at dizzying rates. Indeed, we believe that the rate at which airborne LiDAR units are entering the market is doubling every five years [Fig. 3]. Whether this is sustainable remains to be seen. It may turn out to be a move towards multifunction sensors that allow continued unit growth rates.
Multifunction 'fused' sensors have been spoken of for some time now. Indeed, an airborne LiDAR and a medium format imager, either frame based including 39 Megapixel cameras or line scanners in one housing might be regarded as a fused sensor. Add a waveform digitising capability and a thermal sensor for example and the multifunction sensor is perhaps already here.
The goal is to achieve full spatial information for every pixel collected in real time. A truly fused sensor requires a compatible workflow pipeline with an 'orthophoto-on-landing' capability at a minimum. But why stop there? Communication technology is such that even today, albeit lower resolution, commercial imagery can be datalinked easily from aerial platforms to groundstations. An unfolding 'ortho-carpet' continuously processed and delivered directly to your computer workstation will surely satisfy most imagery consumers.
How can such data volumes be handled? At least in the
"As airborne LiDAR and its accuracy is dependent on GNSS and IMU technologies, it is probable that differential GPS (DGPS), GNSS systems approaching 120 channels of GPS/Glonass/Galileo and Compass signals will appear"
near term, some compromise is necessary. Even in the geospatial industry, we cannot have it all right now. Data throughput will increase, even to the transmission of real-time ortho-imagery to an enterprise or enterprises which can then be served as onward distribution. A requirement will be intelligent data thinning operating in the background with 'understanding' of application specific needs. Thus, a mobile phone user navigating with real-time imagery will be more application or visualisation constrained, and thus more highly data thinned, than a workstation based engineer requiring high resolution, high accuracy, ortho-imagery as a backdrop.
Irrespective of the size and shape of the hardware, the 'smartness' of the software or the speed and place of delivery, one thing is certain. We are once again, like the great exploring surveyors who brought the unknown closer, on the cusp of providing volumes of previously unobtainable information. Saint-Exupéry could only have dreamed of having such capability whilst flying P-38's over the Mediterranean.