ENVI 5.5.1, SARscape 5.5, and IDL 8.7.1
In case you missed it, we recently had a webinar for the ENVI 5.5.1, SARscape 5.5, and IDL 8.7.1 release earlier this week. The webinar was led by BIll Okubo, our ENVI + IDL product manager, and myself. The major features of this release included: new visualization tools in ENVI, machine learning tools in IDL, a package manager for IDL, an upcoming release of SARscape that is built on ENVI tasks, along with a few other items.
The really exciting part of the webinar was the demo! It showed the power of the ENVI Modeler being able to create a workflow once (in this case, hyperspectral) and being able to have an automated piece of processing that can then be accessed in an ArcPro toolbox and run on the Geospatial Services Framework (GSF) in cloud or server environments.
There was also a sneak peak at a pretty cool technology that our engineers have been working on at Harris: the image streamer. The last part of the demo used a web based application for performing processing on GSF. This application uses the image streamer, if configured, to stream the pixels from source data and results to a web map as a WMS layer. During the demo it was streaming back directly from a hyperspectral data cube with about 300 bands. If you would like to view the webinar and learn more about these exciting new features, here is the link to the recording on our page:
In addition to this, I also wanted to call out some of the events that will be coming soon! We will be having a follow-up webinar about using open source tools in ENVI + IDL and will be led by myself and we will be having an introduction to SAR led by some of our SAR experts in Broomfield, Colorado. I'm very excited about the open source webinar because it will allow us to expose some helpful tools that we have been using internally on GitHub for anyone to access. Here are a few of the major repositories that will be made public and what they are used for:
- Machine Learning Toolkit: A suite of tools for making it easy to do pixel-based machine learning in ENVI with the process for creating a classifier broken down into three, easy steps. This also takes advantage of the IDL-Python bridge to use some of the amazing ensemble algorithms in scikit learn and can be extended to create other classifiers. My personal favorite is the ExtraTrees classifier in scikit learn which takes less than a minute to create a classifier and can get near-perfect accuracies with correctly labeled training data (i.e. 99% or greater).
- IDL Package Creator: In addition to the IDL Package Manager which was added in the latest release, this toolset offers an opinionated approach to package management and comes with alternatives for managing your packages, automatically generating documentation from your code and markdown files, creating and running unit tests that are easy to read and write, and provides a simple way to compile all of your code into IDL SAVE files.
- Awesome ENVI Algorithms and Awesome ENVI Tools: If you're an IDL programmer like me, then this collections of code greatly simplifies and offers alternatives to the out-of-the-box ENVI API tools which can help help you write cleaner and more maintainable code for ENVI applications with IDL.
- Plus more! There will also be a handful of smaller packages which are dependencies of the ones listed above that will be published as well.
Keep an eye out for the next webinar registration and I hope you'll join me in our next steps with our products!