After the last blog about the International Mapping Industry Associaton (IMIA), we’d like to explore geospatial imagery analysis (or imagery analytics). Image analysis is the extraction of meaningful information from images collected through GPS, satellite and aerial imagery, LiDAR, drones, and other sources. It provides 2D and 3D analysis supporting commercial and as government sectors and inspects risk assessment and mitigation disaster management, climatic conditions and urban planning.
Image analytics is crucial to many geospatial projects. In many cases, when analyzing geospatial-related data decent analytics can provide users and organizations with crucial information such as change detection or area disturbance over time. We at Xentity would like to take this opportunity to show you several ways imagery analytics can be used. For this, we will be referencing back to several old blogs. We love data, and we love maps, and for this blog we’ll be putting all that knowledge into one article.
What We Can Learn From Maps And History
Back in 2014, we wrote a brief piece about exploring public domain maps and imagery of historic Denver West, Colorado. We discussed what you can learn about an area through its maps such as the layout of cities, mountain ranges, etc.. And not just the usual stuff, like restaurants and landmarks, but its history as well. If you analyze enough maps of the same area but of different time, you eventually see that said areas change overtime.
If you are looking for a more specific application of this idea of “change over time”, there are two ways to look at this concept. There is object-based image analysis, which segments images on a per-pixel basis that groups these pixels into various objects of shapes and scale. In remote sensing, this creates “image-objects” analyzed through their spatial, spectral and temporal scales. So objects such as buildings and rivers, and how those objects change over time, can provide meaningful information.
Method #2
The second method is known as land cover mapping to create a system that uses remote sensing and geospatial data. For example, to assess climate change impacts on and area a fantastic place to start would be to not only use land cover mapping to analyze that impact but also use the historical data from maps throughout the years.
A case in point is Google Earth. You can view various maps throughout history, even when maps are constantly updating. Through analysis of data within the images throughout history, we are capable of finding appropriate information, depending on the industry.
Or, consider our past project with the United States Geospatial Service’s (USGS) Historical Quadrangle Scanning. As the write up discusses in the beginning the maps were “high-cartographic quality paper maps were critical for geologic exploration, population migration, emergency management, urban planning, resource planning, topographic/hypsographic analysis and so much more.” Our duty was to scan decades worth of maps and create the metadata and catalogue them. We scanned maps created over a 90 year period. Theoretically, we could perform imagery analysis of these maps over this period. Then we would definitely gather so much data just from the changes in imagery alone.
Location, Location, Location
To reference back to another blog, a reminder that “good data fuels successful apps, especially when they are map-oriented.” If an app tells you Starbucks is one place but you end up at a Dunkin’ Donuts, it’s not a very good app. If our own consumers cannot extract the meaningful information from provided maps, then there is no way we can call the app successful. Organizations that use maps for their operations have to properly analyze their own maps for decent decision making. There are also cases where the consumer has to be able to perform their own image analytics. For example, consider Regulation Explorer, an app which helps oil & gas companies find the best locations by putting Colorado oil & gas regulations “on the map” in combination with environmentally and culturally sensitive areas.
An app Xentity created for the Colorado Parks and Wildlife (CPW) is an example of a successful one. The CPW Fishing app allows users to search over 1,300 fishing destinations across Colorado for their own use. The app allows users to view a location’s popularity, ease of access, present fish species, stream gauges and more. Users can record days on the water, journal their catches, and place catch locations on the map. The design of this app, like many, is to ease a certain process through the use of data accessibility. In this case, it comes through maps.
One More Past Project
We would also like to look back on one last project that also demonstrated the importance of imagery analytics. Xentity’s work for the United States Forest Service’s (USFS) Disturbance Assessment Service (DAS). In this case, the USFS needed to provide more comprehensive data from satellite imagery to fire professionals (fire fighters, disaster relief organizations, etc.). The USFS wanted DAS to make better use of satellite imagery and then provide any appropriate results. The whole idea of providing data from that imagery meant extracting necessary information from satellite images. The same applies whether it was through computers or human visualization. That is imagery analytics.
What To Take Away From All Of This?
You will likely apply imagery analytics if you involve yourself with maps. Whether it is for climate change impact by analyzing changes overtime (historical maps). Or, if you are creating or using an app involving mapping and location. Or, if you are a company that can analyze geospatial data to compile reports. You know, like Xentity. Regardless, do not discount the fact that imagery analytics can provide crucial data for consumers and organizations. Otherwise, you are just looking at a glorified painting.