Home Articles EditorSpeak: From image to insight – are we ready?

EditorSpeak: From image to insight – are we ready?

Prof. Arup Dasgupta
Managing Editor
[email protected]

A friend, teaching remote sensing and GIS in an engineering college, invited me to deliver a lecture on ‘Remote Sensing Applications – Industry Perspective’ to his students. His reasoning was that being attached to industry, I could give a practical view as opposed to the bookish view. As I began preparing for the talk, I realised that there was a fundamental flaw in the title. It was not possible to isolate remote sensing from other technologies if I was to address applications of a practical nature. For example, if I tried to illustrate the use of high resolution data for land records, I could not ignore differential GPS which would give me the data to correctly register the plot boundaries to the imagery. Again, the use of imagery for precision farming and agri-business has to include GPS, GIS and data related to prices, inflation and other economic factors. In fact remote sensing is only one of the data sources, an important one though not the only one.

Imagery becomes the starting point. To this we have to add value by extracting information required for the end user. To enable more people to be able to use satellite imagery, satellite data providers are moving from selling imagery to providing information and insight to customers. This has tremendous impact in many ways. Data buyers are aware of the storage issues. Back in the 80’s, data used to come in tapes and it was a daunting task to ensure their longevity by storing them in climate controlled environments. CDs and DVDs have brought great relief to this situation but have only changed the dimensions of the problem. With spatial resolutions likely to drop to 25 cm on one hand and spectral resolutions and thereby spectral bands increasing with the advent of hyperspectral imagery on the other, the problem multiplies by orders of magnitude. As the lead article on hyperspectral remote sensing shows, the technology is hampered by cost of data and the complexity of data analysis.

This calls for a radical change in the way we perceive data usage. Instead of buying and owning data we have to shift to a model of using data. Data will exist somewhere, in a Cloud may be, and be available for access. More importantly it will be available in a form required by the user. Thus, a user needing the NDVI of a particular area will be directly served the NDVI image and not the multiple band imagery. I believe this will be a great boon to users who currently spend a considerable amount of time, energy and money just managing data, leaving little time for data analysis. Such a change will also impact the data producers who till now have been working on a product price model as they will have to migrate to a value added access price model.

Where does this leave our regulators? All current regulations are based on data resolution and national jurisdiction. The shift to an access based model will upset all their rules. I may be denied terrain height as data but what if I can order an image draped on a DEM? Undoubtedly the regulators will catch on and by and by catch up; but by the time they do, technology would have advanced to another level.