Trends in Remote Sensing: Using all Available Tools
So it seems these days I’m doing a lot of proposal work. Not that I don’t love ENVI, but using my words from time to time is a good diversion. In the course of this work, I’ve been noticing some industry trends that I think are worth discussion and mention. I’d love to hear what trends others are seeing….
Big Data processing:
GPUs, Raspberry Pi clusters, distributed processing, MANY core systems, Hadoop systems, you name it. Imagery data has gotten “bigger” than it already was and people want results now. The good news is Exelis VIS already has solutions available today for these processing architectures. Believe it or not, they are not that hard to implement and some are not that expensive. Check out my webinar with NVIDIA on GPU processing. Mike Galloy’s GPU Lib for IDL is another great solution. Both of these offerings fall into the category of big bang for a minimal investment.
Find My Feature
ENVI and ENVI LiDAR have a lot of tools for out of the box feature extraction, but the accuracies some customers need with certain features are at such a high level, typically we have to do some customization. The good news is we have some wickedly smart (and highly entertaining) people who work at VIS and just “get” this stuff. I’m not sure there’s a modality that we haven’t fine tuned for feature extraction and QA and getting accuracies greater than 95%. And with most feature extraction data, it’s big, see the above paragraph. Here’s a video on LiDAR feature extraction—this is down looking LiDAR, but we have some amazing tricks up our sleeves for terrestrial LiDAR, drop me a line….
The “Blended” Enterprise
In today’s world of openness and community collaboration, sometimes COTS software gets pigeon holed as being expensive or not playing well with others. What I’m seeing is a willingness to embrace the best of both worlds. I’m hearing more need of people who actually understand imagery data and what can be done with it, and the development modality is second. Obviously we’re proud of our COTS software at Exelis VIS and believe it does amazing things, but if a customer wants a Java enterprise framework and needs people who really understand scientific data and processing, we build what the customer needs because we’re just that excited about working with imagery. ENVI and IDL both interact well with other languages like C++, Java, Python etc. So even on projects where we’re using Java, tools that have been built and hardened in ENVI end up saving money and time because they don’t need to be re-engineered in another language.
The IDL Python Bridge, Slither, is another great example of bridging COTS with open source to get the best of both worlds. With the Slither you can execute Python statements directly from IDL and it’s free from Jacquette Consulting. Here’s Ronn Kling’s Ebook on instruction using the bridge. We’re far from open source phobic, we see it as bringing a new and interesting friend to dinner who has great stories to tell. Open source lets us get creative and challenge our thinking, which is an excellent way to expand community resources and opportunities.
Please feel free to reach out to me with questions or comments Amanda.firstname.lastname@example.org