This model has been a great tool for identifying ADA curb ramps that we missed in our initial collection. This time-saving model allowed us to shift human labor from this project to other projects that needed more assistance.
County Innovates Using GeoAI to Inventory ADA Curb Ramps and Saving Significant Time and Money
A big part of unlocking the value of geographic information system (GIS) technology is constant innovation, including learning new capabilities and searching for opportunities to apply them. Douglas County, Nebraska, has established an innovation culture in its GIS group—it constantly expands its knowledge and use cases of the technology across the organization. With about 585,000 residents, Douglas County is the most populous county in Nebraska and home to the city of Omaha.
Like all US government agencies, Douglas County and the City of Omaha must abide by the stipulations in the Americans with Disabilities Act (ADA) of 1990. This important legislation protects people with disabilities in many areas of public life. One of the accessibility standards associated with the act covers curb ramps, where sidewalks end at a curb and provide access to people crossing the street. These can be dangerous for people with disabilities, so there are ADA-specific design standards that must be followed in their construction. Over the years, as the Omaha Public Works Construction Division has been improving its roadway intersections and adding new intersections, each of those projects has included installation of ADA curb ramps.
As the city installed the ramps, it wanted to collect asset information and location to help with its asset life cycle management program, powered by Cityworks. Field data collection requires too many resources. Since the city was collecting aerial imagery every two years, it had administrative staff in the office enter the ramps and their attributes into its GIS using the aerial imagery as a reference. Over the years, there was no consistency and no data standards with this process, which led to poor data quality for the ADA curb ramp layer.
County GIS staff had become aware of geospatial artificial intelligence (GeoAI) capabilities within ArcGIS, like deep learning models for extracting features from imagery, and had been looking for an opportunity to try them out. The ADA curb ramp inventory seemed like a great use case to test these capabilities, since ramps can be recognized by the human eye. Using ArcGIS Pro and the county's one-inch-resolution digital aerial imagery in a mosaic dataset, Steve Cacioppo, a senior GIS analyst at the county, set about applying a deep learning model to help solve this problem.
Developing a deep learning model takes some trial and error, and Cacioppo worked through many different scenarios using samples of data. One scenario narrowed down the focus of the model to only areas where intersections and ramps are located, since there is no need to analyze every square inch of the county. Multifamily housing and commercial areas were also isolated, since they have curb ramps. The county has a layer identifying areas without sidewalks; these areas were not analyzed. The model was not 100 percent accurate—some ramps were lost in shadows, and initially it identified some car sunroofs as curb ramps—but these issues were minor and correctable.
Once the model was finalized and run, it identified 34,183 ADA curb ramps. The original inventory included 16,775 ramps, so the improvement was substantial; the number of ramps was over double the initial number, and the Construction Division was excited about this big improvement in data quality. The time savings have been considerable. On average, it had taken county staff one to two minutes to add a ramp into the GIS. The GeoAI model identified the ramps in about 12 days on a PC using ArcGIS Pro. With the curb ramp inventory completed, staff can now identify which remaining crosswalks need ADA curb ramps and can use GIS to prioritize their installation for inclusion in the CIP, as well as review those projects through an equity lens, like the ArcGIS Solutions Social Equity Analysis configuration.
The resulting ramp data was shared with other municipalities in the county, so they now know where their ADA curb ramps are located. Now that the county has this process in place, when new aerial imagery is flown every two years, county staff will be able to update the inventory in an automated fashion. Accurate management of these critical assets is important for the county in terms of its safe streets and Vision Zero programs, and the GIS team is already supporting that effort with ArcGIS Hub and ArcGIS Dashboards.
This innovative experience has expanded the county's appreciation for GeoAI, and staff are looking for additional ways to apply it, including feature extraction of swimming pools for property appraisal and health department inspections, and testing to see if this is applicable for pavement markings, sidewalks, and bike lanes. The county also wants to apply GeoAI to its lidar data to extract trees and the edges of pavement. Other improvements being considered are moving the imagery to ArcGIS Image Server or ArcGIS Image for ArcGIS Online to improve GeoAI model and other geoprocessing performance.
The responsibilities of a GIS professional include constant learning and innovation, and Cacioppo and the rest of the Douglas County staff take that to heart. "This project really served two purposes for us. First, it was our initial attempt at creating a deep learning model from scratch. The success of this model will pave the way for future Esri deep learning models. Second, this model provided a time-saving approach to rapidly collecting ADA curb ramps for the City of Omaha. It will be used anytime we get new imagery to continually update the ADA curb ramp inventory", said Cacioppo.
This project was another example of how they are being creative with GIS to make a real difference at the county and in the community. This entrepreneurial attitude includes the development and implementation of an accomplished drone program. Douglas County is a great example to emulate, so the next time you start a new project, before doing it as you always have, take time to brainstorm how you can be creative with GIS to maximize your impact on the organization and community using new capabilities and methods.