You are here

Demonstration of Image-Based Change Detection Using a Prototype Drone-Based Track Safety Inspection System

Award Information
Agency: Department of Transportation
Branch: N/A
Contract: 6913G618P800102
Agency Tracking Number: 180FR4033
Amount: $149,232.00
Phase: Phase I
Program: SBIR
Solicitation Topic Code: 180FR4
Solicitation Number: 6913G618QSBIR1
Solicitation Year: 2018
Award Year: 2018
Award Start Date (Proposal Award Date): 2018-09-10
Award End Date (Contract End Date): 2019-04-09
Small Business Information
7906 Jansen Ct
Springfield, VA 22152-2410
United States
DUNS: 081083858
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Herbert Henderson
 Principal Investigator
 (703) 254-6891
Business Contact
 Herbert Henderson
Title: Principal Investigator
Phone: (703) 254-6891
Research Institution

Demonstration of Image-Based Change Detection using a Prototype Drone-Based Track Safety Inspection 3/16/2018

Today’s prevailing methods of visual track inspection tend to be expensive, disruptive to operations, and have potential to be less thorough than preferred. Machine vision technology emerged in the rail sector over a decade ago to help alleviate these concerns; however, due to the uncontrolled nature of rail environments, the technology has not achieved its expected potential. Image-based change detection emerged over 25 years ago and is well-suited for application to aerial imagery. Preliminary results indicate that the simplest form of change detection (image differencing) is able to highlight many relevant track conditions larger than a configurable size threshold with a detection probability near 100 percent. The use of change detection transforms the rail sector machine vision problem into a relevant/non-relevant determination – a problem expected to be handled robustly with the aid of state-of-the-art machine learning. The research proposed here intends to demonstrate a prototype Drone-Based Track Safety Inspection System using change detection to identify track conditions that are then automatically classified as relevant or non-relevant with assistance from machine learning. If successful, the approach is expected to reduce the cost of visual track inspection with simultaneous improvements in effectiveness, equating to increased operational safety for trains.

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government