In the first part of this 2-part post we discussed how the Spatial Modeler can help you Keep Moving, Jump the Gaps and Adapt Quickly with its Analytics and inclusion of innovative technologies. In this post, we’d like to address some of the more difficult obstacles in geoprocessing: the ones we didn’t see coming.
When confronted with an obstacle, it is frequently our first instinct to figure out a way to work around it. In Parkour, however, you are encouraged to use your ingenuity. When you look at things differently, you often find that the shortest route is not around the obstacle, but over it.
Use all the Skills at Your Disposal: Expanding the Spatial Modeler
Python used to add ImageStation functions to Spatial Modeler
When you are faced with major challenges you need the ability to utilize every tool in your toolshed to get the job done. Being forced to jump from software package to software package to do a single job is frustrating, time-consuming, and expensive.
When we designed the new spatial modeler, we recognized that extensibility is critical. We wanted to supply our users with the toolset to go beyond building models with only the fixed library of operators we provide. Although the extensive library is growing daily, we also provide several mechanisms for users to extend Spatial Modeler.
The most basic but powerful method is designed for software-savvy domain experts who do not program. For these users, we included a command line operator. This allows you to specify an executable with defined arguments, and build it into your workflow.
The next level is for those familiar with Python scripting. Using Python operators you can add your own algorithm or make the model do things it might not do out of the box.
The third level is for C programmers. Our Spatial Modeler SDK and accompanying training course provides advanced users with the ability to build their own new operators.
Make It Look Effortless: Easily Increase Productivity
myVR Operators built using the SMSDK
We knew that we had to make the modeler fun and easy to use. We believe we achieved this with simple interactivity like dragging and dropping data or operators onto the canvas and then connecting the ports. Why reinvent the wheel every single time? Many of ERDAS IMAGINE’s functions are already spatial models, so we allow you to access the model straight from the GUI and tweak it as necessary.
When a model is complete, share it with the non-domain experts who need the outputs. The modeler allows you to rename and label all inputs, parameters, and outputs in user-friendly terms, helping the non-expert be clear about what the model is expecting. E.g., instead of “Enter Input 1,” they can see a useful message like “Enter a DEM image of the study area” or “This is your Permeability Output Image.”
Being productive is easy too. With the click of a button, a model can be sent to a batch processing queue where input/output variables can be defined and distributed processing can begin.
OK, so in Parkour, it’s also about being flashy and showing off, and if you are building incredibly powerful models it’s quite likely no one outside of your cubicle is going to know the geoprocessing acrobatics you pulled off to make this complex computational model both elegant and simple. But what will be noticed is when you and your team start turning around job requests in days instead weeks, minutes instead of hours. Or better yet, that you can enable the end users to create their own value-added products.
Go Viral: Geoprocessing in the Hands of the People
Once your model is built, it can be shared via a web processing service (WPS). The models are self-describing so when the web portal user selects a model he will see a description of the model as well as other pertinent information like what various outputs represent. It will even search the ERDAS APOLLO catalog for appropriate inputs. When executed, the model runs on the server. Then the result is published back to the catalog and available to you and others for viewing or download. If the result doesn’t look right, the user can tweak the settings and run again. Then they simply download the new answer.
We’ve come a very long way with our new Spatial Modeler and we are very proud of what we have achieved. Functionally and aesthetically, it’s second to none. We’ll continue to extend the capability with RADAR, photogrammetry and hyperspectral functions and we continually work on performance.
For more information, watch the webinar I recently hosted on Spatial Modeler capabilities.
We want to hear about your successes and achievements. We invite you to help. We encourage you to build and share models. In the near future, we will be launching a way for you to actively share your amazing creations with other power users. If you want to learn more, from an Introduction to the Spatial Modeler to how to utilize the SDK, take one of our courses to learn how to leverage this powerful capability in your application or business.