Thermal Infrared Remote Sensing

Thermal infrared (TIR) data is acquired by a multitude of ground-based, airborne, and spaceborne remote sensing instruments. A broad variety of fields apply thermal infrared remote sensing, for example to assess general land- or sea-surface temperature dynamics, detect forest, coal and peat fires, map urban heat islands or thermal water pollution, differentiate geologic surfaces, analyze soil moisture, or even to test materials, to name only a few applications. As thermal infrared data has to be analyzed slightly differently than reflective data, this chapter contains the relevant theoretical background. The thermal domain of the electromagnetic spectrum, the laws of Planck, Stefan-Boltzmann, Wien, and Kirchhoff, as well as important parameters such as kinetic and radiance temperature, emissivity, and thermal inertia are briefly explained. The chapter thus provides readers with a common understanding before proceeding to subsequent chapters.

This chapter presents an overview of thermal imaging sensors for photogrammetric close-range applications. In particular, it presents results of the geometric calibration of thermographic cameras as they are used for building inspection and material testing. Geometric calibration becomes evident for all precise geometric image operations, e.g. mosaicking of two or more images or photogrammetric 3D modelling with thermal imagery. Two different test fields have been designed providing point targets that are visible in the thermal spectral band of the cameras.

Five different cameras have been investigated. Four of them have solid state sensors with pixel sizes between 25 and 40 μm (i.e. size of single sensor element on the chip). One camera is working in scanning mode. The lenses for thermographic cameras are made of Germanium, which is, in contrast to glass, transparent to thermal radiation. Conventional imaging configurations (typically 20 images) have been used for camera calibration. Standard parameters for principal distance, principal point, radial distortion, decentring distortion, affinity and shear have been introduced into the self-calibrating bundle adjustment. All measured points are introduced as weighted control points. Image coordinates have been measured either in the professional software package AICON 3D Studio (ellipse operators), or in the software system Stereomess (least-squares template matching), developed by the Institute for Applied Photogrammetry and Geoinformatics of the Jade University of Applied Sciences Oldenburg.

The calibration results differ significantly from camera to camera. All lenses show relatively large decentring distortion and deviations from orthogonality of the image coordinate axes. Using a plane test field with heated lamps, the average image precision is 0.3 pixel while a 3D test field with circular reflecting targets results in imaging errors of 0.05 pixel.

Thermal infrared (TIR) spectra of Earth surface materials are used in a wide variety of applications. These applications can fall into either of two groups: (a) where the TIR emissivity spectra themselves are the primary interest, and are used to determine the chemical/physical parameters of minerals and rocks, soil, vegetation and man-made materials, or (b) where the primary interest is in the temperature of the objects under study, and where emissivity spectra are required inorder to best determine kinetic from radiant temperature. Unlike visible-near infrared (VNIR) and shortwave infrared (SWIR) instruments, TIR spectroscopy instrumentation often requires customization in order to acquire reliable and reproducible data, making thermal spectroscopy a potentially complex process. Within this chapter we intend to provide a simple starting point for the new user of thermal infrared spectroscopy, and a synoptic overview of the technique for the more experienced practitioner. We discuss the theoretical background, give examples of instrument setups and provide typical measurement scenarios for a number of land applications.

UAV-borne thermal imaging involves the determination of ground surface temperature from thermal infrared measurements deploying an unmanned airborne vehicle (UAV). A large variety of UAVs is available and applied for different military and civil tasks. UAV-borne thermal imaging provides spatially distributed information of the ground surface temperature. In contrast to satellite or ground based measurement, the usage of a UAV allows us to obtain spatially distributed and geometrically highly resolved information on the ground surface temperature without the need to access the ground. The area can be flat or hilly, and steep walls and hillsides can be investigated easily. However, some problems, especially tasks related to mosaicking of the images, are not fully resolved to date. We address the detection of the anomalies in ground surface temperature induced by underground burning coal seams as example and describe the challenges and opportunities of UAV-borne thermal imaging, based on our experiences in this field.

The Hyperspectral Thermal Emission Spectrometer (HyTES) is being developed as part of the risk reduction activities associated with the Hyperspectral Infrared Imager (HyspIRI). HyspIRI is one of the NASA’s Tier 2 Decadal Survey Missions for earth science. HyTES will provide information on how to place the spectral filters on the HyspIRI Thermal Infrared Instrument as well as provide antecedent science data. The HyTES pushbroom design has 512 spatial pixels over a 50-degree field of view and 256 contiguous spectral bands between 7.5 and 12 μm in the thermal infrared (TIR) wavelength region. HyTES includes many key enabling state-of-the-art technologies including a high performance concave diffraction grating, a quantum well infrared photodetector (QWIP) focal plane array, and a compact Dyson-based optical design. The Dyson optical design allows for a very compact and optically fast system (F/1.6). It also minimizes cooling requirements due to the fact it has a single monolithic prism-like grating design which allows baffling for stray light suppression. The monolithic configuration eases mechanical tolerancing requirements which are a concern since the complete optical assembly is operated at cryogenic temperatures. The QWIP allows for optimum spatial and spectral uniformity and provides adequate responsivity or D-star to allow 200 mK noise equivalent temperature difference (NEDT) operation across the TIR passband. The system uses two mechanical cryocoolers to maintain instrument temperature. The first cooler holds the focal plane array at 40 K and the second cooler holds the remainder of the cryovacuum system at 100 K. Assembly of the system is now complete and the system is undergoing alignment and laboratory testing. Once laboratory testing is complete the system will be used to acquire airborne data from a Twin Otter aircraft over the southwestern USA in late 2012.

NASA’s Hyperspectral and Infrared Imager (HyspIRI) mission is one of the missions recommended in the National Research Council Earth Science Decadal Survey. HyspIRI will fly two instruments: a hyperspectral visible to short wave infrared imaging spectrometer, and a multispectral thermal infrared (TIR) imager. In this study we discuss the expected performance and use of the TIR instrument. The TIR instrument will have eight spectral channels, seven of the channels are between 7 and 12 μm, with one additional channel at 4 μm. The TIR instrument will have a swath width of 600 km, and pixel size of 60 m. HyspIRI TIR will provide two visits every 5 days (one day and one night) at the equator, and more frequently at higher latitudes. The TIR instrument will always be on and full resolution (60 m) data will be downlinked for the entire land surface including the coastal oceans (shallower than 50 m depth). Data over the deeper ocean will also be downlinked but at a reduced spatial resolution of 1 km. In response to the Decadal Survey, HyspIRI has been designed to answer important science questions in the areas of coastal, ocean and inland aquatic environments; wildfires; volcanoes; ecosystem function and diversity; land surface composition and change; and human health and urbanization. NASA’s Distributed Active Archive Center will archive and distribute Level 0 to Level 2 products. In addition a direct broadcast capability will allow users to capture and process a subset of HyspIRI data in near real time.

This chapter presents an overview of the most commonly used spaceborne sensors for thermal infrared research applications. There is a large fleet of international sensors available which allow for the acquisition of data in the thermal infrared. Depending on spatial coverage, some sensors are more suitable for mapping large areas, while others support observations at a local scale. Temporal resolution defines whether temperature patterns or phenomena can be monitored on a daily, weekly, monthly, or even only an annual basis. A wide variety of thermal sensors will be introduced in overview tables. However, as certain sensors with thermal infrared bands have established themselves as ‘work horses’ for certain types of applications, they will be especially highlighted and presented in depth. A comprehensive overview of typical thermal infrared application studies and the sensors particularly favored rounds off this chapter.

High sensitive infrared detectors normally require more resources than comparable instruments in the visible spectral bands. Although the un-cooled detector arrays achieved in the last years a remarkably quality, their detection principle is inferior to the cooled quantum detectors. The price for the higher sensitivity of the cooled quantum detectors are higher efforts in mass, volume, power consumption, and costs. Therefore it is of interest to examine the compatibility of high sensitive infrared systems with the limited resources of small satellites which could be utilized for affordable space missions. The FIRES (Fire Recognition System) study was a first attempt to examine the accommodation of a challenging infrared mission on a small satellite. Based on this concept the BIRD (Bi- spectral Infra-Red Detection) satellite was launched in 2001, this satellite was mainly dedicated to the detection and monitoring of high temperature events. Following the success of the BIRD satellite a further constellation of two satellites called FIREBIRD (Fire Recognition with Bi spectral Infra-Red Detector) is currently in preparation.

The purpose of this chapter is to describe the collection of thermal images by Landsat sensors already on orbit and to introduce a new Landsat thermal sensor. The chapter describes the Landsat 4 and 5 thematic mapper (TM) and Landsat 7 enhanced thematic mapper plus (ETM+) sensors, the calibration of their thermal bands, and the design and prelaunch calibration of the new thermal infrared sensor (TIRS). The TIRS will be launched in February 2013 on the Landsat Data Continuity Mission (LDCM) satellite, which will be renamed to Landsat 8 after it reaches orbit. Continuity of the data record has always been a priority for the Landsat project. The TIRS will extend the unique Landsat thermal data archive begun in 1978 that supports, among other applications, water resource management in the western United States and global agricultural monitoring studies. The TIRS also introduces improved technology and data quality, both of which are discussed in the chapter.

High resolution thermal infrared remote sensing can have a wide range of applications. In this chapter we describe the different applications and requirements identified after a revision study in the framework of the Fuegosat Synthesis Study (FSS). This project was funded by the European Space Agency (ESA), and the three main objectives were: (i) review of applications and analyses for user requirements, (ii) consolidation of user requirements over a broad range of applications, and (iii) matching of user requirements and industry concepts to identify and outline a set of potential mission scenarios and their corresponding requirements. This chapter focuses on issues (i) and (ii). These objectives were achieved by means of integrated studies within literature and ancillary documentation, and also by consultation of external experts. As a result, more than 30 applications were identified within three different fields: (i) Land and Solid Earth, (ii) Health and Hazards and (iii) Security and Surveillance. A complete set of requirements (spatial, temporal, and radiometric resolution, algorithms used, supporting data, among others) were also provided.

Land surface temperature (LST) products retrieved from two different sensors – AVHRR (Advanced Very High Resolution Radiometer) and MODIS (Moderate Resolution Imaging Spectroradiometer) – were cross-compared. The analysis was conducted on a daily basis for 4 different years. Only pixels that followed a certain homogeneity criteria were chosen. Furthermore a time criterion defining the maximal time difference between two acquisitions was considered. The differences of the two products showed diurnal and annual patterns with LST of AVHRR being higher than MODIS at high surface temperatures and AVHRR being lower than MODIS at lower temperatures. Additionally some irregular patterns were identified and attributed to the different algorithm approaches. However, mean annual absolute differences were relatively low: 2.2 K for the daytime and 1.4 K for the nighttime scenes, indicating a general good agreement between the two products. The r² between the LST of AVHRR and MODIS for both day and night scenes was about 0.99.

This study aims to evaluate quantitatively the land surface temperature (LST) from SEVIRI data (Spinning Enhanced Visible and Infrared Imager, onboard MSG-2 satellite) with the MODIS (Moderate Resolution Imaging Spectroradiometer, onboard Terra)-derived LST extracted from the MOD11B1 V5 product. Two SEVIRI-derived LST level-2 products are used for this purpose: the LSTs retrieved using the generalised split-window method with the emissivities estimated using the day/night TISI (Temperature Independent Spectral Indices)-based method and the LSTs generated by the Land Surface Analysis of the Satellite Application Facility. The results show that (1) higher discrepancies are observed during the daytime, especially for bare areas, with a maximum of 5.7 K; (2) these differences are time- and land cover-dependent; (3) these differences strongly depend on the view zenith angle differences; and (4) the two LST retrieval algorithms for SEVIRI present the higher discrepancy for bare areas, with a maximum difference of 6.1 K.

The thermal infrared (TIR) radiance at sensor measured by any spaceborne or airborne instrument will include atmospheric emission, scattering, and absorption by the Earth’s atmosphere. These atmospheric effects need to be removed from the observation in order to isolate the land-leaving surface radiance contribution and retrieve important surface variables such as land surface temperature (LST) and emissivity. The accuracy of the atmospheric correction is dependent upon accurate characterization of the atmospheric state using independent atmospheric profiles of temperature, water vapor, and other gas constituents. The profiles are typically input to a radiative transfer model for estimating atmospheric transmittance, path, and sky radiances. Residual errors from incomplete atmospheric correction constitute one of the largest uncertainties in derived LST and emissivity products from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and the Moderate Resolution Imaging Spectroradiometer (MODIS) sensors on NASA’s Terra satellite. This chapter will describe a technique for improving the accuracy of the atmospheric parameters on a pixel-by-pixel basis using the Water Vapor Scaling (WVS) method. We have shown that using WVS can improve the accuracy of LST retrievals by up to 5 K for MODIS and 3 K for ASTER data in humid conditions.

The time span of surface thermal data bases now reaches a few decades. However, studies using surface thermal time series are seldom, due to the difficulty of obtaining temporally coherent estimations for this parameter. Applications for surface thermal multitemporal analysis range from climate change studies and modeling to anomaly detection for natural or industrial hazard detection. This chapter presents methods to improve the temporal coherence of temperature time series, through data reconstruction of atmospheric and cloud contaminated observations, and through the correction of the orbital drift effect which hinders the use of the longest data sets. Then, methods for the analysis of time series are presented, including both image to image comparison and trend detection, the choice between these methods depending on the spatial resolution of the dataset and the aims of the considered study.

Sea surface temperature has been an important application of remote sensing from space for three decades. This chapter first describes well-established methods that have delivered valuable routine observations of sea surface temperature for meteorology and oceanography. Increasingly demanding requirements, often related to climate science, have highlighted some limitations of these approaches. Practitioners have had to revisit techniques of estimation, of characterising uncertainty, and of validating observations – and even to reconsider the meaning(s) of “sea surface temperature”. The current understanding of these issues is reviewed, drawing attention to ongoing questions. Lastly, the prospect for thermal remote sensing of sea surface temperature over coming years is discussed.

Soil moisture is an important geophysical parameter and information on soil moisture is needed by many scientific disciplines in the context of climate modeling, hydrologic modeling, flood and drought forecasting, or in the context of geo-health applications. Changes in soil moisture can be the driver for changes in vegetation cover and might directly impact land use and agricultural yield.

A lot of approaches to derive soil moisture from remotely sensed spaceborne earth observation data exist. Most of them are based on the utilization of radar data, such as scatterometer data derived from instruments onboard the ERS satellite (ERS-Scat) or the Advanced Scatterometer onboard of the METOP satellite (METOP-Ascat). Such data comes at resolutions of 50 and 25 km respectively and has the large advantage that it can be acquired independent of solar illumination and cloud cover. Furthermore, several scientists have used higher resolution synthetic aperture radar (SAR) data for soil moisture estimation. It is less well-known that thermal infrared satellite data is also suitable to retrieve soil moisture information. As thermal data usually is available at a higher resolution (1 km and better) it is an attractive alternative to radar data. This chapter presents approaches of soil moisture retrieval from thermal data, and discusses advantages and shortcoming of soil moisture extraction based on this data type. Benefits on a synergistic operational soil moisture product based on both thermal and radar data are discussed.

The objective of this study is to infer information on Soil Moisture Content (SMC) in agricultural areas using daily gradient of brightness temperature and albedo from MODIS AQUA, based on the so-called apparent thermal inertia (ATI) approach. The developed algorithm has been validated over two different test sites in Italy, Emilia Romagna and South Tyrol regions, and one test site in France, the Pyrenees region, where ground truth measurements were available. For the Emilia Romagna and the Pyrenees test sites, the obtained ATI values were well correlated with SMC values. For the South Tyrol test site, due to large heterogeneity in the mountain landscape, the correlation between ATI and SMC was relatively weak. Cloud coverage which reduces the number of available observations and the vegetation cover which decreases the sensitivity of ATI to SMC were the main limitations in all analyzed test sites. This study showed that a combination of data with a frequent revisit time and polar orbiting sensors can alleviate the impact of cloud coverage on the retrieval. In fact, a comparison between ATI derived from MSG (Meteosat Second Generation) SEVIRI (Spinning Enhanced Visible and Infrared Imager) and MODIS indicated a good correlation between the two estimates thus demonstrating the potential of a possible synergy between the two sensors.

Thermal remote sensing is widely used in the detection, study, and management of biomass burning occurring in open vegetation fires. Such fires may be planned for land management purposes, may occur as a result of a malicious or accidental ignition by humans, or may result from lightning or other natural phenomena. Under suitable conditions, fires may spread rapidly and extensively, affecting the land cover properties of large areas, and releasing a wide variety of gases and particulates directly into Earth’s troposphere. On average, around 3.4 % of the Earth’s terrestrially vegetated area burns annually in this way. Vegetation fires inevitably involve high temperatures, so thermal remote sensing is well suited to its identification and study. Here we review the theoretical basis of the key approaches used to (1) detect actively burning fires; (2) characterize sub-pixel fires; and (3) estimate fuel consumption and smoke emissions. We describe the types of airborne and spaceborne systems that deliver data for use with these active fire thermal remote sensing methods, and provide some examples of how operational fire management and fire research have both benefited from the resulting information. We commence with a brief review of the significance and magnitude of biomass burning, both within the ‘whole Earth’ system and in more regional situations, aiming to highlight why thermal remote sensing has become so important to the study and management of open vegetation burning.

Remote sensing thermal data of active lava flows allow for the evaluation of instantaneous effusion rates. This is made possible by simple formulae relating the lava effusion rate to the power energy radiated per unit time from the surface to the flow. The most questionable assumption is probably the constancy of the surface temperature. Due to the assumptions of the model, this formula implies that heat flux, surface temperature and lava temperature varies as a function of the flow thickness. These relationships, never verified or validated before, have been used by several authors as a proof of the weakness of the model. Herein, MIVIS (Multispectral Infrared and Visible Imaging Spectrometer) high spatial resolution (5–10 m) thermal data acquired during Etna’s 2001 eruption were used to investigate down-flow heat-flux variations in the lava flow emitted from a vent located at 2,100 m a.s.l. A high correlation between the down-flow heat-flux and the lava flow thickness (measured from a pre-existing digital elevation model) was found. According to this relationship, observed changes in the surface temperature would be the expected consequence of differences in the down-flow lava flow thickness due to topographic variations.

The EOS-1 Terra ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) has acquired about 200 images (100 of them sufficiently cloud-free to be used) over Mt. Etna since 1999. This chapter shows the results from the analysis of 10 years Mt Etna activity using thermal infrared (TIR) high spatial resolution data by a semi-automatic procedure that extracts radiance values of the summit area with the goal of detecting variation related to eruptive events. Night time data showed a good correlation with the main eruptive events that occurred both in the summit and in the flank areas. A comparison of the variance of maximum ASTER TIR radiance with variance of the maximum AVHRR TIR radiance (Advanced Very High Resolution Radiometer) for the same area confirms good correlation in terms of trend and values between the two data sets. Finally this study emphasizes the importance of high spatial resolution TIR data during background monitoring to detect changes in the thermal emission that may be related to an impending eruption and the need to further improve the spatial resolution in the TIR channels to better separate the thermal active areas in volcanic systems.

Surface and underground coal fires are burning in numerous countries worldwide. China, India, the USA, Australia, Indonesia, South Africa, and many other countries all report uncontrollably burning coal fires. They ignite through spontaneous combustion of coal, or through lightning, forest fires, fires in garbage dumps, or careless human behaviour. Coal fires lead to the loss of the valuable resource and lead to the emission of green-house gasses as well as toxic gasses. These gasses contribute to climate change and also impact human health. Vegetation above the fires deteriorates. Due to the volume loss underground coal fires also trigger land subsidence and surface bedrock collapses. The surface and underground fires can be detected and monitored by means of remote sensing. Data acquired with handheld thermal cameras, airborne sensors, and also spaceborne sensors have been analyzed by numerous authors. However, exact and simultaneously standardized as well as transferable methods for coal fire detection and monitoring are hard to establish, and research gaps still exist This chapter presents a broad overview of past and current coal fire work, as well as the challenges which can be addressed based on thermal data of recent and upcoming sensors.

In areas of anomalously high crustal heat flow, geothermal systems transfer heat to the Earth’s surface often forming surface expressions such as hot springs, fumaroles, heated ground, and associated mineral deposits. Geothermal systems are increasingly important as sources of renewable energy, or as natural wonders of protected status attracting tourists, and their study is relevant to monitoring deeper magmatic processes. Thermal infrared (TIR) remote sensing provides a unique tool for mapping the surface expressions of geothermal activity as applied to the exploration for new geothermal power resources and long term monitoring studies. In this chapter, we present a review of TIR remote sensing for investigations of geothermal systems. This includes a discussion on the applications of TIR remote sensing to the mapping of surface temperature anomalies associated with geothermal activity, measurements of near-surface heat fluxes associated with these features as input into monitoring and resource assessment, and the mapping of surface mineral indicators of both active and recently active hydrothermal systems.

Scientists have reached to a large extent agreement on climate warming for the coming decades. This will especially have immense impact on cities which show in general a significantly higher temperature compared to rural surroundings, e.g. due to high percentage of impervious surfaces. This study shows capabilities of airborne and spaceborne thermal remotely sensed data to derive and analyze land surface temperatures (LST). Dependencies of LST to urban structure types (UST) with respect to their location within the city are analyzed. Results prove distinct correlations between LST and vegetation fraction as well as percentage of impervious surfaces. Beyond this, different USTs prove influences on LST. Last but not least, a general decrease of LST with increasing distance to the city center is confirmed for the city of Munich. However, the USTs superimpose this trend and have a significant influence on the local LST.

This is a case example of mineral mapping of unaltered and altered rocks at the Cuprite mining district, southwestern Nevada using the Spatially Enhanced Broadband Array Spectrograph System (SEBASS), a thermal infrared hyperspectral sensor that collects radiance measurements in the mid-wave infrared and thermal infrared portions of the electromagnetic spectrum. Cuprite, Nevada has been a test bed for a variety of multispectral and hyperspectral sensors that have predominantly covered the visible through short-wave infrared portion of the electromagnetic spectrum. In 2008, 20 SEBASS flight lines were collected at an average altitude of 4,735 m yielding an average 3.35 m ground sample distance (GSD).

Rock forming and alteration minerals found in this mining district have reststrahlen features (emission minima due to fast changes in refractive index with wavelength) in the thermal infrared portion of the electromagnetic spectrum (7.5–13.5 μm). Mineral mapping with hyperspectral thermal infrared data provides unique and complementary information to visible-shortwave (0.4–2.5 μm) hyperspectral data. Mineral maps were produced using a spectral feature fitting algorithm with publicly available mineral spectral libraries containing signatures.

These mineral maps were compared to the geological and alteration maps along with mineral maps generated by previous studies of visible-shortwave infrared hyperspectral sensors to assess some of the difference in mineral mapping with a hyperspectral thermal infrared sensor. This study shows that hyperspectral thermal infrared data can spectrally map rock forming minerals associated with unaltered rocks and alteration minerals associated with different phases of alteration in altered rocks at Cuprite, Nevada.

Land surface temperature and emissivity (LST&E) are important variables used in surface energy balance models, monitoring land-cover land-use changes, and in surface composition mapping. For most retrieval algorithms that generate LST&E products from spaceborne thermal infrared data, accurate retrieval of the LST depends on an accurate estimate of the spectral emissivity in the TIR region between 8 and 12 μm. This is because both determine the amount of thermal radiance that gets emitted to the atmosphere from the Earth’s surface. Consequently, validation of emissivity products from sensors such as MODIS and AIRS is a critical aspect for better quantifying uncertainties in the long-term LST record, and to help better constrain surface energy balance modeling. Two methods of validating the emissivity currently exist; an in situ method that utilizes TIR instruments such as radiometers employed in the field, and a laboratory-based method that uses a high spectral resolution spectrometer to measure field collected samples in a controlled environment. This chapter will discuss the methodology for validating emissivity products over pseudo-invariant sand dune sites using the lab-based method.