Airborne Thermal Hyperspectral Imaging of Gases Using Different Modes

Gas cloud monitoring is typically involved in environmental monitoring and gas leak surveys. Ground-based characterization is not feasible due to the unpredictable dynamic related to freely dispersing gas clouds in open air conditions. Traditional gas sampling techniques have sampling problems and are laborious due to the necessity to bring the sample to the lab. Although in situ gas concentration measurements yield precise and sensitive real- time results, their single point results are hard to correlate with dynamic entities such as dispersing gas clouds. Additionally, safety is a concern when performing gas leak surveys on explosive gases.

Broadband infrared imaging acquires spatial-temporal data from safe distances, but lack of quantitative data about the gas cloud and selectivity issues are the drawbacks. Ground-based thermal infrared hyperspectral imaging involving Fourier transform spectroscopy (FTS) yields both selective and quantitative data and are successfully used in several field programs involving remote gas release detection, determination and quantification. The same technology is now applied in airborne sensors, providing selective and quantitative data for gas and solid detection in addition to large area coverage. Recent advances in thermal infrared hyperspectral sensors allow using two different acquisition modes for FTS imaging: mapping and targeting. This article demonstrates the ability of airborne hyperspectral FTS-based imaging sensors to acquire unique dynamic gas cloud dispersion data.

Telops Hyper-Cam

The Telops Hyper-Cam is a compact hyperspectral imaging device using Fourier transform infrared (FTIR) technology. The lightweight instrument has a unique combination of spectral, spatial, and temporal resolution, providing comprehensive characterization of substances of interest. The Hyper-Cam is equipped with a Focal Plane Array (FPA) detector consisting of 320x256 pixels over a basic 6.4°x5.1° field of view (FOV). The spectral resolution is user-selectable of up to 0.25 cm^-1 over the spectral range of 3.0 to 5.0 or 7.7 to 11.8µm for the Hyper-Cam MW (midwave) and Hyper-Cam LW (longwave), respectively.

The Telops Hyper-Cam Airborne Platform helps extending the capabilities of the ground-based Telops Hyper-Cam for airborne applications (Figure 1). It features an inertial motion unit (IMU) and global positioning system (GPS) for tracing and georeferencing of the aircraft actions in flight. Based on acquisition parameters, data acquisition time for a single datacube in FTS imaging can be few seconds. Hence, the airborne module features an image motion compensation (IMC) mirror that uses the GPS/IMU data to negate the aircraft movements during airborne data acquisition. During recording, the viewing angle change is generally below 2°. All data include the significant information for orthorectification and stitching. A highresolution boresight camera is employed for recording visible images at the same time of infrared data acquisition. A radiative transfer model developed by Telops is used to perform chemical imaging.

Figure 1. The Telops Hyper-Cam airborne platform

Experimental Procedure

All airborne measurements were performed using a Hyper-Cam LW sensor. The measurements over the waste incinerator were performed using a wide angle telescope (0.25x, FOV of 25.6°x20.4°). The mapping acquisition mode was used to perform measurements over the waste incinerator at a spectral resolution of 4cm^-1 (i.e., 135 bands) and an altitude of 2130m leading to a ground pixel size of 9m²/pixel. The targeting acquisition mode was used to take measurements at a spectral resolution of 8 cm^-1 (i.e., 68 bands) and an altitude of 915 meters leading to a ground pixel size of 1.65 m²/pixel. At ground level, relative humidity and ambient temperature are 49% and 11°C, respectively.

The ethylene gas release air borne measurements were performed using the targeting acquisition mode at a spectral resolution of 6 cm^-1 (i.e., 90 bands) and an altitude of 685m leading to a ground pixel size of 0.057 m²/pixel. Relative humidity and ambient temperature at the ground level were 37% and 21°C, respectively. A pressurized cylinder released pure ethylene gas at a constant flow rate of roughly 20L/min. The two acquisition modes used for airborne hyperspectral imaging are depicted in Figure 2. The mapping acquisition mode involves continuous recording of individual images or datacubes when the aircraft flies above its area of interest (AOI). The targeting acquisition mode involves recording of successive hyperspectral datacubes of the same AOI using the IMC mirror component.

Figure 2. Mapping (left) and targeting (right) hyperspectral imaging acquisition modes of the Telops Hyper-Cam airborne platform.

Experimental Results

Airborne FTS Imaging

The Telops Hyper-Cam airborne platform acquires data at relatively high spectral resolution and has higher selectivity for remote detection and determination purposes, thanks to the FTS infrared technology. High spectral resolution also helps airborne image analysis due to the dependency of common signal processing algorithms such as temperature-emissivity separation and atmospheric correction on accurate compensation of ozone and water vapor contributions. Figure 3 shows the typical results acquired with the Telops Hyper-Cam airborne platform, illustrating the irregularities of targets of interest in airborne surveys. The red curve represents the thermal infrared spectrum of a clear sky, whereas the blue curve represents the infrared spectrum corresponding to a pixel related to gas and the green curve represents an infrared spectrum corresponding to pixel related to a concrete road.

Figure 3. Thermal infrared hyperspectral datacube (top right) and its corresponding visible image (bottom right) recorded with the Telops airborne platform.

Airborne Survey Using Mapping and Targeting Modes

Figure 4 illustrates individual datacube acquisitions from an area consisting of an operating waste incinerator using the mapping acquisition mode. To produce a proper hyperspectral map from individual acquisitions, it is necessary to include sufficient overlap from neighboring ground areas in each acquisition. Figure 4 also depicts the hyperspectral map of the entire AOI obtained from orthorectification and stitching of individual acquisitions, demonstrating the facility of performing the stitching procedure to generate a single datacube of the entire AOI.

Figure 4. Individual airborne FTS infrared acquisitions from an area containing an operating waste incinerator using the mapping acquisition mode (left).

Figure 5 illustrates the ability of a hyperspectral airborne sensor to provide the time-dependent information. In this case, the targeting acquisition mode was used to perform the airborne measurements over the same AOI, recording six successive hyperspectral datacubes in total. The red clouds represent the gases released from the main smokestack. Comparison of different images can help deducing the gas cloud source point.

Figure 5. Successive airborne acquisitions of an operating waste incinerator using the airborne FTS infrared targeting acquisition mode.

Airborne Measurements of Ethylene Gas Release

In this experiment, ethylene gas was released before and during the airborne survey of the target area. Figure 6 depicts the visible image captured during the experiment and the corresponding infrared image of the target area, showing no contrast difference between the cold infrared-active gas (labeled by β) and the background asphalt pavement (labeled by α). This can be justified by the very little contribution of ethylene’s sharp and highly localized infrared spectral features when averaging the overall signal. The infrared spectra of the gas release point (β) and the background asphalt pavement (α) are shown in Figure 7 on a brightness temperature scale with relevant reference absorption spectra.

Figure 6. Airborne visible image of the gas release area (left) and infrared broadband image of the target area of interest (right).

Figure 7. Longwave infrared spectra of representative pixels located in the gas cloud and on a background area (top), and relevant reference absorption spectra of ethylene and water vapor (bottom).

Figure 8 presents the results of quantitative chemical imaging performed on two successive acquisitions captured on the AOI. From consecutive airborne measurements performed on a freely dispersing gas cloud, the gas leak originating point can be easily located with high accuracy. The difference analysis of the two chemical maps shown in Figure 8 helps to estimate the velocity as well as determining the direction of the gas cloud.

Figure 8. Quantitative airborne chemical imaging carried out on two successive measurements of an ethylene gas cloud using hyperspectral targeting acquisition mode.

Conclusion

Highly accurate quantitative chemical imaging of gas clouds can be obtained owing to the selectivity offered by the high spectral resolution of airborne infrared FTS hyperspectral imaging. The Telops Hyper-Cam airborne system provides meaningful time-dependent information to characterize gas cloud dispersion, including the origin of a gas leak and the velocity and direction of a gas cloud. The good agreement with the expected results demonstrates the advantage of using airborne thermal infrared hyperspectral imaging in the characterization of gas cloud dispersion.

About Telops

Telops Inc., located in Quebec City, is a leading supplier of hyperspectral imaging systems and high performance infrared cameras for the defence, industrial, environmental and research industries. Telops also offers R&D services for optical systems technology development in order to respond to the specific needs of its customers.

Since its beginnings in 2000, Telops has distinguished itself with the quality of its technical personnel and its innovative approach to many technological challenges in the optics field. Today, the expertise of Telops scientists, engineers and technologists in optical technology, and the performance of its infrared cameras and hyperspectral imagers are recognized internationally.

While being headquartered in Canada, Telops caters to an international market using an efficient network of distribution and representation.

Whether you are looking for equipment, expertise or subcontracting for your optical system projects, Telops will turn your high expectations into success.

This information has been sourced, reviewed and adapted from materials provided by Telops.

For more information on this source, please visit Telops.