The issues of aging infrastructure are exacerbated for United States government organizations that must face developing problems with limited budgets to repair, update and maintain their assets. Proactive and preventive steps can head off problems before they escalate to crisis level, but the first issue is finding those small problems.


Article Index:


A recent remote sensing project using thermal infrared aerial imagery at Redstone Arsenal in Huntsville, Ala., conducted by Atlantic offers the government some new insights for solving a portion of this growing problem.

Redstone Arsenal found itself in a familiar scenario for government agencies — it had more needs than finances. Rather than waiting for additional funding, they chose to be innovative with money they did have and invest in a non-traditional type of geospatial data that could be used to help them identify and prioritize their public works projects. This new data would become a tool for the Arsenal that would yield far more return than anticipated, thus resulting in the ability to do even more with less.

Prioritizing Needs

Redstone Arsenal’s Directorate of Public Works (DPW) and Marshall Space Flight Center (MSFC) were tasked with acquiring information to help them prioritize their needs to preserve and repair the Arsenal’s infrastructure and utility assets. DPW is the garrison’s primary element for maintenance of the installation infrastructure and environment. Within DPW, the Master Planning Division (MPD) plans all new construction and improvements to installation facilities and grounds. DPW’s mission is to prolong the lifespan of existing infrastructure and increase the overall return on investment (ROI) on all assets, as preserving infrastructure is more economical than reacting to emergency situations.

By targeting preventive maintenance and focusing on predictive upkeep, the Arsenal’s intent was to complete routine restorations more effectively and eliminate many costly emergency repairs. The government thus sought an innovative and proactive approach through the utilization of thermal mapping.

Thermal anomalies throughout the dataset could be applied as an indicator for early detection of weaknesses in the system to allow prioritization of targeted inspections and allow repair crews to pinpoint where time and money could to be invested most efficiently.

Specifically, the Arsenal wanted to use this data for energy loss analysis in the form of leaks or flaws in steam lines, failure in insulation from building rooftops creating heat loss and/or moisture issues, as well as water leaks in underground mains. By identifying where infrastructure was not in proper working order, the Arsenal could potentially save thousands of dollars by simply making repairs before the problems turned into costly failures.

Thermal Imagery Solution

Atlantic was contracted by Redstone Arsenal’s DPW and MSFC to conduct thermal imaging acquisition, analysis and interpretation over three areas of interest (AOIs) in spring of 2015. The scope of work included the acquisition of thermal imagery over 5,100 acres (eight square miles) at 6-inch pixel resolution and 30-percent overlap of flight lines. The data had to be captured during a specific time and within certain temperature ranges to reasonably detect all variances of energy loss from the different types of infrastructure that were within the identified AOIs.

Atlantic was tasked to apply thermal and radiometric normalization as well as geometric rectification to the collected imagery. In addition, after the data was fully developed and processed, Atlantic’s last tasks were to analyze and interpret thermal anomalies throughout the areas of interest for the specified assets.

Approach

Having previously invested in planimetric and LiDAR data, Redstone Arsenal provided a base map for the thermal project. Atlantic collected high-resolution, sensitive thermal imagery using an aerial platform equipped with hardware and software to acquire data. Using specialized software and client-provided datasets, Atlantic processed and analyzed the captured data. Using the existing expertise of Atlantic’s team, it provided the client stakeholders with useful information to assess the condition of their infrastructure assets.

Identifying the Thermal Sensor

On this project, Atlantic partnered with thermal sensor manufacturer ITRES and chose to deploy their Thermal Airborne Broadband Imager (TABI-1800) sensor. The TABI-1800 addressed the needs and questions posed by the client. The sensor features a wide pixel array (1,800 pixels over a 40-degree field of view) and is able to produce high sensitivity, georeferenced thermal data. Other key features important to the project were:

  • A wide array (1,800 pixel swath)
  • 40-degree field of view (FOV)
  • Internally cooled detector
  • Broadband thermal imager
  • MCT (mercury cadmium telluride) detector
  • High sensitivity (as low as 1/10 of a degree (C))
  • Low temperature drift
  • An optical system designed for airborne mapping
  • Integrated airborne GPS and IMU
  • Flown with a POS AV
  • Manufacturer calibration
  • Precision ortho-corrections and fusion with terrestrial LiDAR data

Moving to Data Capture

Atlantic first developed a flight plan to allow for efficient and safe capture of data over the AOIs. Collaborating closely with several authorities controlling the highly restricted airspace over the Arsenal, authorization to proceed with the project was eventually received in January 2015.

The TABI sensor was installed onboard Atlantic’s Piper Navajo Chieftain PA-31 aircraft in early March. The Navajo was an ideal choice as it has two full-size sensor holes and can simultaneously capture LiDAR and thermal data. In addition, it has a T-AS Gyro-stabilizing suspension mount and easy integration of the TABI sensor to communicate with Atlantic’s current flight management system. After a boresite/sensor calibration flight, the actual acquisition was conducted during a night mission in early April.

Data Processing

After capture and quality control, the data underwent several processing steps including:

  • Bundle adjustments using the boresite results
  • Application of radiometric and thermal normalization calibration coefficients to convert raw data into apparent radiant temperature values, and to produce uniform and balanced temperature contrast
  • Application of measurements from ABGPS and IMU, plus bundle adjustment, to create a geometrically corrected and georeferenced mosaic
  • Rectification using client-provided first-return LiDAR data (DSM) and image mosaicking.

Analyzing Results

After creation of the radiometrically and geometrically corrected TABI-1800 image mosaics, the thermal data underwent a series of analysis steps to detect and identify any potential thermal anomalies. By applying thermal anomaly detection algorithms over the thermal image mosaics, Atlantic was able to define areas of thermal anomalies in/on infrastructure within the Redstone Arsenal study areas. Based on those detections and their radiant temperatures within a specific region of the image mosaic, Atlantic produced an algorithm to search the entire image for thermal anomalies that would auto generate and assign temperature classes.

DPW provided geospatial datasets containing utilities (pipelines) and building footprints that were used to define the boundaries of assets of interest and extract the thermal anomaly data (i.e. masking non-related areas). The isolated thermal anomaly pixel values were analyzed within each corresponding area of the thermal imagery. By comparing the mean and standard deviations of the pixel values, temperature anomalies were delineated and vectorized for inclusion into new geospatial datasets for each type of utility/infrastructure asset. The thermal imagery for each of the infrastructure components was presented in the form of color density levels (or standard deviations) of approximately two- to five-degree increments specifically chosen to enable detection of thermal anomalies.

This analysis generated geospatial data depicting the anomalies over the infrastructure of interest. It also created a temperature classification map of the image mosaic. The final deliverables from the image analysis included:

  • Masked TABI-1800 image mosaic containing only the infrastructures of interest (i.e. buildings and pipes)
  • Building vector thermal anomaly shape files, showing only the buildings from the building vector with detected thermal anomalies and temperature classes
  • Pipeline thermal anomaly shape files, showing only the areas from the pipeline vector with detected thermal anomalies and temperature classes
  • Apparent radiant temperature classification map of the TABI-1800 image mosaic for the infrastructures of interest

The final thermal anomaly deliverables are not a measurement of absolute temperature, but rather a representation of relative temperature differences throughout the image mosaic. The results provided valuable insights into the Arsenal’s infrastructure asset conditions.

Results / Benefits

Information is a powerful tool. It allowed this government operation to make decisions and outline plans that most benefit their structures. As such, this new thermal imagery data gives the Arsenal’s Energy Management Team a whole new tool kit to help strategically meet their current and future infrastructure needs. They are able to determine and prioritize which areas of the assets need to be repaired, replaced and closely monitored to schedule future assessments.

The client agencies are using the newly acquired Airborne Thermal Infrared Imagery to detect steam line leaks, building heat loss, roof damage and roof moisture. The Arsenal was aware of a few steam leaks, but using the thermal imagery they were not only able to confirm the leaks, they also realized the extent and severity of some of the problems.

The thermal data also provided a means to properly locate underground utilities. Most of the locations of older steam and water pipelines were imported into the Arsenal’s GIS database from as-built CAD drawings made years before. The accuracy for some of the utility pipelines was off by tens of feet in some instances. Adding the ability to accurately locate components of the existing system will make repairs more efficient, saving both time and money. The thermal data and the anomaly results also aided in prioritizing repairs and establishing a schedule for roof replacements and/or repairs.