874 Rate this article:
4.4

Hyperspectral Imaging – Seeing is Believing

A cell phone picture has three primary bands (red, green, and blue) that recreate what the human eye sees. But, what if you want or need to see more? Hyperspectral imaging (HSI) uses thousands of specific bands (think hundreds of different segments of blue, green, red, and even going beyond the visible spectrum!) to pick up the unique response of what you are imaging. So instead of just having a gorgeous picture of that sunflower in a field, you now can tell its lignin content, health, and even begin to model it’s CO2 interactions. The wavelength bands in HSI provide a high-level of spectral and radiometric accuracy that makes the imagery particularly useful in a number of industries including energy, environmental monitoring, agriculture, construction, and government to name a few.

HSI for Oil and Gas Applications

The ability to get high-detail information is essential for many monitoring practices in our complicated world, where complex chemical and physical interactions of things like vegetation, soil and mineral types, and even chemical analysis of items, need to be understood. One example of this type of monitoring uses hyperspectral imaging to track petrochemicals. The chemical seepage of oil is visible with hyperspectral technology even in areas of mixed land surface such as vegetation, soil, and manmade structures. Oils and petrochemicals have unique signatures within the visible and near infrared spectrum, and so large-scale analysis for different environmental interactions and searching for spills of interest is applicable with this technological growth.

In maritime settings, hyperspectral analysis has also proven to be able to pick up oil signatures on the water surface. This allows for better understanding of the amount of oil that is present, which can be used greatly for forecasting impact, and tracking the spill itself. For example, after the Deep Horizon oil spill in the Gulf of Mexico, hyperspectral analysis was done on multiple oil samples, allowing for classification of similar oil responses on water and beach surfaces.

Oil spill hyperspectral example

Hyperspectral imagery over the Deep Horizon Gulf Oil Spill, collected by SpecTIR, can be used to pull out multiple aspects of the environment, and even classify the oil signature as opposed to the background water as seen with the example on the far right.

Hyperspectral Sensors take Flight on UAVs

In combination with the ability to pick up unique features, recent advances in hyperspectral imagery technology have made it easier to use. Hyperspectral cameras have been made smaller and cheaper, and can be put on drones for deployment over small regions exactly when, and where, they are needed. Similarly, to how utility companies use UAV-mounted hyperspectral cameras to monitor pipes and utility poles, high-resolution hyperspectral imagery flown with drones can pinpoint possible accumulations of petrochemicals.

The combination of drones and hyperspectral cameras allows for more diverse and essential monitoring applications to become available, such as tracking changes in coral reefs. With an aggressive increase in bleaching, understanding the spread and effects of the chemical content of corals is crucial for monitoring and possible mitigation. As coral reefs are mainly found in clean, clear water, hyperspectral cameras see through the water column to categorize coral species, and amount of bleaching that is occurring over an area. Previously this kind of cataloguing was costlier, as it would either take in situ measurements of the entirety of the reef, or multiple high resolution aerial surveys.

The combination of drones and hyperspectral cameras allows for more diverse and essential monitoring applications to become available, such as tracking changes in coral reefs. With an aggressive increase in bleaching, understanding the spread and effects of the chemical content of corals is crucial for monitoring and possible mitigation. As coral reefs are mainly found in clean, clear water, hyperspectral cameras see through the water column to categorize coral species, and amount of bleaching that is occurring over an area. Previously this kind of cataloguing was costlier, as it would either take in situ measurements of the entirety of the reef, or multiple high resolution aerial surveys.

Mineral recognition, invasive species tracking, agriculture disease and nutrient identification, and monitoring of hazardous wastes are all accomplished with hyperspectral imagery, and it is only becoming easier to use and employ. When the imagery is combined with modern machine learning and deep learning techniques, unique insights can be gained for best understanding the world that we live in.

Regardless of how hyperspectral data is collected, whether by satellite, fixed wing aircraft, or drone, processing and analysis of the data is the same. ENVI® is a full-featured image processing and analysis software with specific tools to handle hyperspectral data including spectral detection and identification. There are interactive capabilities that are only found in ENVI and spectral library support and workflows that are needed to process hyperspectral data. The processing algorithms in ENVI that are geared toward hyperspectral imagery have been tested and ultimately proven by the remote sensing community over several decades.

For more information about processing HSI using ENVI, check out the white paper on that topic.

Please login or register to post comments.