Leveraging Technology to Propel Geospatial Understanding
Jeff Liedtke has personally witnessed the incredible evolution of the geospatial industry during his 30-year career developing imagery and software products using remotely sensed data. With bachelors’ degrees in both Geography and Environmental Studies from the University of California Santa Barbara, as well as a master’s degree in Remote Sensing and Geography from Simon Fraser University, Liedtke’s subsequent work history includes positions with International Imaging Systems, Space Imaging, DigitalGlobe, eMap International and Esri.
“Early in my career, we were encumbered by technology limitations – there was never enough compute power, memory, storage, bandwidth, no GPS, no lossless compression – and now we’re enabled by technological advances in so many areas,” Liedtke explains. “Remote sensing satellites are being deployed by dozens of countries around the world, small ‘cube satellites’ are being launched by universities and institutions at modest cost, and many mapping companies have drone programs. We are awash in imagery. And photogrammetric mapping and sophisticated remote sensing image processing is being performed in the cloud, ground reference data is collected and collated from crowd sourcing, and daily monitoring applications are now possible.”
He continues, “In the past, the geospatial disciplines – GIS, photogrammetric mapping and remote sensing – evolved on separate trajectories with little overlap other than transferring data. Now they are converging, which is transformative in terms of advancing our understanding of geospatial relationships, interconnected systems, and cause and effect, and fostering collaboration across many different disciplines. This is a very exciting time for our expanding industry, and there is incredible opportunity to leverage technology to propel our geospatial understanding and knowledge and provide informed decision support for the betterment of our planet – and beyond.”
Q. What do you do for a living?
I’m the documentation lead for Esri’s Imagery and Raster team. This position requires that I use my experience and knowledge of remote sensing, image processing, applications and workflows, as well as software development, training and education programs, to develop Help documentation. This also includes how-to videos, blogs, Learn Lessons, Story Maps, and training courses and materials.
Q. What is your favorite tool to work with?
Multispectral imagery – everything about it! I love working with remotely sensed imagery and the analysis tools and techniques to extract specific information from the data. Imagery provides a rich dataset that contains information for a myriad of applications. The trick is to tease out the right information for your particular application. The treasured information for one application is just noise that needs to be reduced for a different application.
Imagery is often the foundation dataset for either baseline mapping or revising maps. It provides essential inventory and management information for any number of applications such as mapping of new development, infrastructure, natural resources, water quality, environmental assessment, situation awareness and emergency management.
The overall synoptic view gives context for geospatial data and phenomena at all scales – continental, regional and local. The inherent visual information in imagery often reveals cause and effect, trends and critical information not obtainable from other sources. And imagery of the Earth is beautiful, the best kind of art. I never grow tired of it.
Q. What is the toughest challenge you face?
Getting different technology and necessary components of a project to work together. Over my career, it has always been challenging to meet the requirements of a project in an operational workflow. To adequately address real-world projects, you often need to utilize disparate technologies, multiple systems and subsystems that all need to work together. Each of these systems or subsystems work fine in their specific environment and perform well in regression testing. But the real challenge is getting these subsystems to work together and integrated into an overall solution in an operational environment.
There are many ways to design software or a solution. When a project team understands the end goals and purpose of utilizing their piece of technology and where it fits into the overall solution, they can approach a problem more holistically. Knowing how the output from one component will be used as the input to other subsystems allows a robust solution to be designed and implemented in an optimal manner.
Q. What is the biggest lesson you've learned?
Be prepared and always have a contingency plan. Redundancy can be a life saver. What happens if you do not have access to the internet or lose cell phone connectivity, which is often the case during a disaster or emergency? Have critical data and software on spare disk drives. If you are in the field, have backup batteries and spare parts – and plenty of water and snacks. If you are giving a presentation, have a high- and low-resolution backup on a thumb drive or even a hardcopy. When you travel, have appropriate power connectors and spare cables and data backups. These steps can help avert a catastrophe!
Q. What advancements would you like to see made?
Almost all applications are concerned with detecting and characterizing change. As soon as a map is created it either needs to be updated or is analyzed and related to other information that affects the context of the information, thereby driving a changed perception of the original information. Semi-automated analyses geared towards reliably detecting and defining types of change would be universally useful.
I’d also like better tools for combining the analysis of different types of data and information into a more robust ecosystem that considers the input and relative importance of all kinds of disparate data. For example, using existing GIS and geospatial data with imagery to help derive meaningful classification, then using this more accurate information to detect change, and importantly, the characterization of change. In terms of feature extraction, it would be beneficial to classify features not only based on spectral information, but also shape, size, orientation, and contextual information such as proximity to other similar or different objects, time of day or other events, seasonality and other factors.
Q. What are your keys to success?
Too often the approach to problem solving is to start with a technology or methodology and force fit it to a problem set. The key to success is to understand a problem from the users’ or clients’ point of view, then solve it holistically using the most appropriate data, tools and methods within a given budget. This requires a deeper understanding of both the application you are trying to address, and knowledge of how to leverage – and augment – available resources. While I try to overcome limitations, there are always constraints that need to be accommodated to arrive at the best solution to satisfy essential project requirements. Set expectations and then over-deliver while keeping in mind that perfection is often the enemy of good enough. And lastly, teaming up with the right people to accomplish ambitious goals. Never underestimate the power of a committed team, and you will not only achieve worthy accomplishments, but also create lifelong friends and colleagues.