Project Goal
The Alabama Department of Transportation (ALDOT) is modernizing business practices and operations through the convergence of innovation in GeoAI, Linear Referencing Systems (LRS), and Mobile LiDAR with high resolution street view imagery to improve data quality, accelerate update cycles, and enhance enterprise analytics. The program is fueled by the department’s internal data center and virtual machine network to efficiently handle the massive computing needs to facilitate the operation. The focus of the initiative is to transform complex and difficult to interpret data into uniform datasets statewide that are incorporated into ALDOT’s Enterprise GIS (eGIS) system by leveraging advanced automation, machine learning, and pre-trained transformers underpinned by expert collaboration and rigorous quality control.Business Problem Solved
Historically, ALDOT faced time intensive, variable, and labor heavy workflows to collect data for roadway characteristics and infrastructure assessments. Traditional methods could not keep pace with the growing demand for standardized high-quality data needed in a rapidly evolving data-driven world. By deploying GPU accelerated GeoAI capabilities and remote sensing to automate data extraction, we managed to keep ALDOT personnel out of the highways with fewer injuries. Furthermore, this approach has significantly reduced repetitive workloads, improved data accuracy, and empowered small teams to achieve more. The department has experienced accelerated onboarding and cross-training efforts by augmenting lengthy technical workflows.
Technology Implemented
This solution leverages Esri’s eGIS ecosystem to manage, integrate, and deliver AI driven roadway data. High resolution imagery and LiDAR are processed into GIS ready datasets with Esri’s enterprise environment using ArcGIS Pro and the vast suite of deep learning capabilities it offers. Some of the AI technologies incorporated include a custom trained convolutional neural network (CNN), object detection, and visual question answering (VQA). The deep learning outputs are ingested into our Roads and Highways enterprise LRS for analysis, visualization, and system wide updates. This comprehensive integration enables scalable asset inventories, supports validation workflows, and provides organization wide access to standardized, accurate data essential for transportation planning and operations.
Development Team Biography
George Conner, Deputy Director
Terence Burke, Bureau Chief
Charlotte Hosea, Asst. Bureau Chief
Joseph Dean, Asst. Bureau Chief
Bradley Hall, GIS/CAMMS/Pre-Construction Programming Support Mgr.
Ken Kohnke, GIS Geodatabase Analyst
Scott Brandhuber, DevOps Mgr.
Mark Garrison, Infrastructure Platform Services Mgr.
Tom Trammell, Networking Specialist
Robbie Nuckolls, Information Security Officer
Eric Christie, State Maintenance Engineer
Ronny Pouncey, Deputy Bureau Chief / Data Collection & Management
Chris Tadlock, GIS/LRS Data Management Admin
Jared Horne, GIS/AI/ML Mgr.
Cecil Moore, GIS/AI/ML Analyst
William Register, GIS/AI/ML Tech
Stephanie Khamken, GIS/AI/ML Tech
Brett Sellers, State TSMO Engineer
Morgan Musick, Maintenance Management System Engineer
Brette Grant, MIRE Admin
Omar Armstrong, Field Coordinator
Wade Henry, Asst. State Design Engineer
Steve Griffin, CADD Systems Operations Mgr.
David Welch, Special Projects Engineer
Alyson Kroft, Enterprise Informa