Remote Sensing Tools

I’ve been asked several times, since it became known that I had converted the entire NASA GeoCover 7,4,2 mosaic to natural-color — what I plan to do with this dataset. Some of those inquiries have been interesting. I thought I’d devote a post here to explain my reasoning for producing the dataset by myself, and what plans I did have or have thought about.

The full production phase began a couple months after I realized that I had an appropriate conversion methodology, through the time I had spent researching and developing algorithms as plugins to Photoshop in the weeks prior. Upon establishing that methodology — I then made the decision to challenge myself with the task of processing the entire world at 14.25 meter resolution — in Photoshop®. The initial thought was just to process the 30 meter resolution Landsat TM NASA GeoCover mosaic, but ultimately deciding on the 14.25 meter mosaic because, well, it was there — I knew that if I didn’t, it’d haunt me personally.

Some might consider taking on such a task alone something like dancing around the rim of insanity, while those out there who know me and respect me for the things I’ve done, or am at least capable of doing, might consider it an act of passion. Perhaps there’s a degree of both elements that causes one to drive him or herself to such lengths alone in a project of such magnitude. Or perhaps, as in my case, I simply couldn’t resist challenging myself to such a degree — and demonstrate to myself that certain things were in fact acheivable, despite much of the criticisms I’ve heard about Photoshop as a viable image processing tool in the industry. I continue to receive inquiries on how to do things on a regular basis from others in the industry in relation to Photohop. (Which, based on the latter, is an interesting occurence in light of such criticisms, one would think.)

Why haven’t you released the data as an available product?

The idea has crossed my mind, but the only way that I’ve considered it is as a low-cost imagery source for anyone to access through an online interface for download, and where the data itself would be considered free of charge — but the pricing fixed to counter cost and maintenance of the data and system itself. A service platform, in essence, and nothing more. Distribution systems aren’t free to run and operate, and I would wish to retain total control of the library without influence. There also doesn’t appear to be any solid Royalty Free license options available in this industry, so that thought becomes attractive to me as well.

There were also no steps taken to pull the imagery into another program to make it a seamless data-product, thus I didn’t consider it in the league of Eric Augenstein’s TerraColor, or TerraMetric’s TruEarth, or EarthSat/MDA’s NaturalVue. I viewed it as it is — a converted natural-color, Landsat world dataset at 14.25 meter resolution in UTM projection, sitting on my harddrive storage for my own unlimited use and without having to purchase any scenes from another source. Not to mention, why would I wish to step on the toes of friends in the industry who have out-of-the-box data product solutions that are great in their own ways and superior due to being a seamless data product?

The Fun Of It

One of the fun ironies that developed out of the production of the dataset, which became more of a personal highlight for me — is that it was also processed on a laptop computer with mid-level capability. Some of the hardware is what I’d consider outdated. Using a Dell Inspiron 8200 notebook with the maximum 1GB of RAM, I was capable of converting the entire NASA GeoCover dataset from 7,4,2 to natural-color within a period of a little over six months, with a three-step multi-pass algorithm methodology. My initial calculations were longer, actually.

Of course, not everyone has the kind of storage system that I have, to also slave processing, and not everyone understands exactly how to set-up their machines and software to match such efficiency. If you’re also an automation freak like I am, you might find that certain software that you don’t understand as well might surprise you when it is automated. And Windows is an exceptionally fast operating system, despite what some of my Mac and Linux savvy friends might think.

So will you make it available?

I really don’t know at this time. I’ve simply been going with the flow of researching and developing things that I think are interesting in working with this imagery. The inquiries that I’ve received, however, are intriguing and cause me to wonder if perhaps there’s a need that I could potentially help to fulfill if I only did it. I really have nothing to lose if I made it available. I did it as an exercise, and it just sits here except when I use it for my own purposes.

I’ll take the interest into additional consideration, that’s all I can think about at this time. The hope here is really to address the questions I’ve received about the dataset and what my thoughts and motivations were in producing it.

Updated on: September 20th, 2007 at 5:50 am