Highlights What are the main findings? What is the implication of the main finding?Highlights What are the main findings? What is the implication of the main finding?Abstract Reliable and scalable landmine detection technologies are essential for humanitarian mine action (HMA), yet standardized benchmarks for Unmanned Aerial Vehicle (UAV)-based sensing in operationally relevant environments remain limited. This study presents a comprehensive evaluation of 34 multimodal datasets acquired over a standardized seeded test site for landmine and unexploded ordnance detection. Nine sensing modalities, including RGB, thermal, multispectral, hyperspectral, LiDAR, and Synthetic Aperture Radar (SAR), are evaluated using the Anomaly, Identifiable Anomaly, Unique Identifiable Anomaly (AIU) index to establish a unified framework for quantifying detection fidelity. Results indicate that RGB imagery achieves the highest surface detection rate (94.8%), with 45.4% of targets classified as uniquely identifiable, reducing false-positive risk. For sub-surface detection, handheld electromagnetic induction (EMI) and magnetometry exceed 95% detection for ferrous items but fall below 10% for plastic ordnance. Ground-penetrating radar (GPR) is the only modality capable of detecting buried plastic targets (55.6% for cart-based systems), whereas UAV-mounted GPR remains limited (18.2%) at current operational flight heights. Based on the comparative analysis, we discuss the gaps in current detection capabilities, compare false-positive rates across modalities, and perform a cost-benefit analysis fitting contamination scenarios with the most cost-effective detection method. All datasets are publicly released, along with an interactive web-map, to support reproducible benchmarking and cross-modality comparison in UAV-enabled explosive hazard detection.

Quantitative remote sensing using thermal infrared imaging from small unmanned aircraft systems (sUAS) often requires calibration reference targets to correct for atmospheric attenuation. These targets, such as controlled-temperature water baths, are complex to set up, maintain, and monitor, making their use impractical in many field applications. One possible solution is to neglect atmospheric effects, given that most operations occur below 400 feet in the United States, however, the effectiveness of such an approach depends on the task's required noise-equivalent delta temperature (NE Delta T). This study revisits the multiple-altitude calibration technique first proposed by Schott and Gallagher (1976) and applies it to sUAS operations. We demonstrate the technique's viability using both modeled atmospheric propagation of sensor-reaching radiance with MODTRAN 6 and experimental data from a microbolometer-based thermal infrared camera flown on a multi-modal imaging payload. The experimental results showed average absolute errors of 3.69 and 2.50 [K] at aircraft altitudes of 250 and 120 [m], respectively, when atmospheric effects were neglected. Correcting the sensor reaching radiance for atmospheric transmission and upwelling path radiance, derived using the proposed methodology, produced average absolute errors of 1.29 and 1.27 [K], respectively, at these same altitudes.

Users of remotely sensed Earth optical imagery are increasingly demanding a surface reflectance or surface temperature product instead of the top-of-atmosphere products that have been produced historically. Validating the accuracy of surface products remains a difficult task since it involves assessment across a range of atmospheric profiles, as well as many different land surface types. Thus, the standard approaches from the satellite calibration community do not apply, and new technologies need to be developed. The Big Multi-Agency Campaign (BigMAC) was developed to assess current technologies that might be used for the validation of surface products derived from satellite imagery, with emphasis on Landsat. Conducted in August 2021, in Brookings, SD, USA, a variety of measurement technologies were fielded and assessed for accuracy, precision, and deployability. Each technology exhibited its strengths and weaknesses. Handheld spectroradiometers are capable of surface reflectance measurements with accuracies within the 0.01–0.02 absolute reflectance units, but these are expensive to deploy. Unmanned Aircraft System (UAS)-based radiometers have the potential of making measurements with similar accuracy, but these are also difficult to deploy. Mirror-based empirical line methods showed improved accuracy potential, but their deployment also remains an issue. However, there are inexpensive radiometers designed for long-term autonomous use that exhibited good accuracy and precision, in addition to being easy to deploy. Thermal measurement technologies showed an accuracy potential in the 1–2 K range, and some easily deployable instruments are available. The results from the BigMAC indicate that there are technologies available today for conducting operational surface reflectance/temperature measurements, with strong potential for improvements in the future.

Hyperspectral imaging systems frequently rely on spectral rather than spatial resolving power for identifying objects within a scene. A hyperspectral imaging system’s response to point targets under flight conditions provides a novel technique for extracting system-level radiometric performance that is comparable to spatially unresolved objects.The system-level analysis not only provides a method for verifying radiometric calibration during flight but also allows for the exploration of the impacts on small target radiometry, post orthorectification. Standard Lambertian panels do not provide similar insight due to the insensitivity of orthorectification over a uniform area. In this paper, we utilize a fixed mounted hyperspectral imaging system (radiometrically calibrated) to assess eight individual point targets over 18 drone flight overpasses. Of the 144 total observations, only 18.1% or 26 instances are estimated to be within the uncertainty of the predicted entrance aperture-reaching radiance signal. For completeness, the repeatability of Lambertian and point targets are compared over the 18 overpasses, where the effects of orthorectification drastically impact the radiometric estimate of point targets. The unique characteristic that point targets offer, being both a known spatial and radiometric source, is that they are the only field-deployable method for understanding the small target radiometric performance of drone-based hyperspectral imaging systems.

During the launch and path to its final orbit, the Landsat 9 satellite performed a once in a mission lifetime maneuver as it passed beneath Landsat 8, resulting in near coincident data collection. This maneuver provided ground validation teams from across the globe the opportunity of collecting surface in situ data to compare directly to Landsat 8 and Landsat 9 data. Ground validation teams identified surface targets that would yield reflectance and/or thermal values that could be used in Landsat Level 2 product validation and set out to collect at these locations using surface validation methodologies the teams developed. The values were collected from each team and compared directly with each other across each of the different bands of both Landsat 8 and 9. The results proved consistency across the Landsat 8 and 9 platforms and also agreed well in surface reflectance underestimation of the Coastal Aerosol, Blue, and SWIR2 bands.

With the launch of Landsat-9 on 27 September 2021, Landsat continues its fifty-year continuity mission of providing users with calibrated Earth observations. It has become a requirement that an underflight experiment be performed during commissioning to support sensor cross-calibration. In this most recent experiment, Landsat-9 flew under Landsat-8 for nearly three days with over 50% ground overlap, from 13 to 15 November 2021. To address the scarcity of reference data that are available to support calibration and validation early-on in the mission, a ground campaign was planned and executed by the Rochester Institute of Technology (RIT) on 14 November 2021 to provide full spectrum measurements for early mission comparisons. The primary experiment was conducted in the Outer Banks, North Carolina at Jockey’s Ridge Sand Dunes. Full-spectrum ground-based measurements were acquired with calibrated reference equipment, while a novel Unmanned Aircraft System (UAS)-based platforms acquired hyperspectral visible and near-infrared (VNIR)/Short-wave infrared (SWIR) imagery data and coincident broadband cooled thermal infrared (TIR) imagery. Results of satellite/UAS/ground comparisons were an indicator, during the commissioning phase, that Landsat-9 is behaving consistently with Landsat-8, ground reference, and UAS measurements. In the thermal infrared, all measurements agree to be within 1 K over water and to within 2 K over sand, which represents the most challenging material for estimating surface temperature. For the surface reflectance product(s), Landsat-8 and -9 are in good agreement and only deviate slightly from ground reference in the SWIR bands; a deviation of 2% in the VNIR and 5–8% in the SWIR regime. Subsequent longer-term studies indicate that Landsat 9 continues to perform as expected. The behavior of Thermal Infrared Sensor-2 (TIRS-2) against reference is also shown for the first year of the mission to illustrate its consistent performance.

NASA's Earth Science and Technology Office (ESTO) encourages and promotes new and innovative science technologies to improve how the Earth is observed from space. Through their Instrument Incubator Program (IIP), an uncooled multispectral thermal instrument was fabricated and flight-tested to demonstrate its potential utility to support future spaceborne missions. While uncooled systems offer the attractive advantage of eliminating the need for large and expensive cryocoolers, they naturally introduce challenges that must be overcome to achieve the fidelity observed in image data acquired by existing cooled spaceborne instruments. In this work, the Multi-band Uncooled Radiometer Instrument (MURI) system, designed and built by Leonardo DRS, is fabricated using custom microbolometer detector arrays also developed by Leonardo DRS. Potential issues associated with thermal imaging using microbolometers from a spaceborne platform are discussed and innovative engineering solutions to overcome these issues are highlighted. Details of three airborne campaigns designed to assess the fidelity of MURI image data, using Landsat 8's Thermal Infrared Sensor's (TIRS) radiometric & geometric requirements as a baseline, are presented. Results of these campaigns show that the as-built MURI system significantly outperforms these requirements, as compared to ground-based reference measurements. Sustainable Land Imaging (SLI) requirements indicate that a five-band instrument is desirable to improve future Landsat science while maintaining continuity with previous thermal instruments. Considering its multiband design, temperature/emissivity separation (TES) is applied to MURI flight data and compared to ground-based spectrometer measurements for several materials. Results of this study indicate that MURI's TES performance is in-line with existing spaceborne systems. When considered in conjunction with its radiometric fidelity, the MURI uncooled system represents an intriguing option for future space-based missions.

As the number of imaging systems continues to rapidly grow, so too does the need to perform spatial image quality control on captured data. While human assessment for general quality control may be utilized in some applications, target-based and scene-based automated spatial resolution methods may improve speed, accuracy, and efficiency. In certain applications, such as small unmanned-aircraft systems (sUAS) data collection, very large numbers of captured images may even make human assessment impractical, especially for multispectral and hyperspectral modalities. Imagery bands outside of the visible spectrum, such as infrared, may also prove more challenging for human assessment. Traditionally, scientific metrics utilized for spatial quality control such as MTF (modulation transfer function) were measured using in-situ targets such as the slanted-edge or Siemens star. However, newer advanced methods can now estimate these metrics without requiring dedicated targets. Automated machine vision methods utilizing techniques such as neural networks can provide these estimated measurement metrics for rapid image evaluation and quality control. MTF estimation without targets allows much more flexibility in assessment, enables a measurement map across each image, and eliminates the logistical burden of requiring in-scene fielded assets. We examine the performance of one such machine learning based MTF estimation system by placing checkerboard-type slanted-edge targets in multiple sUAS scenes including agricultural, urban, and forested subject matter. We then compare traditional target-based slanted-edge method measurements to the machine learning based estimated values and assess the overall accuracy.

In remote sensing, the conversion of sensor recorded radiance for each pixel in a scene to surface reflectance is a first step in many diagnostic analysis tasks. This process, known as reflectance conversion, is vital in the production of accurate information for a variety of applications, one of which is precision agriculture. Several calibration methods have been explored in previous research and are widely used today, with two of these being the empirical line method (ELM) and the at-altitude radiance ratio (AARR). The AARR approach is attractive to remote sensing practitioners as it allows reflectance conversion to be carried out in real-time throughout data collection, accounting for changes in illumination conditions, which can substantially reduce collection setup and subsequent data analysis time/effort. Illumination changes during a collection greatly influence the recorded scene radiance, which can confuse subsequent analysis results. While ELM has been demonstrated to report lower error when compared to AARR, the error introduced is often times acceptable depending on the application requirements and natural variation in the reflectance of the targets of interest. An onboard, downwelling irradiance spectrometer integrated onto a small unmanned aircraft system as part of this research is utilized to characterize the expected error in generated reflectance at varying aircraft altitudes and is cross compared to the ELM approach. Although the error introduced by AARR is larger when compared to ELM, this study allows the reader to determine if the ease of collection afforded by this real-time reflectance conversion methodology produces sufficiently accurate data for their particular remote sensing application.

Landsat-9, launched on September 27, 2021, carries the Thermal Infrared Sensor (TIRS). The Landsat-9 TIRS is a close copy of the Landsat-8 TIRS instrument; it is a two spectral-band, pushbroom sensor with three Sensor Chip Assemblies (SCAs) that cover the 15-degree field-of-view. The primary radiometric change between the instruments is the addition of baffling in the Landsat-9 TIRS telescope to mitigate the stray light issue that has impacted the radiometric quality of Landsat-8 TIRS. The on-orbit radiometric performance is monitored using the on-board variable temperature blackbody and views of deep space. Maneuvers to look at and around the moon have provided an assessment of the stray light. The absolute calibration is monitored by vicarious calibration techniques by teams at NASA/Jet Propulsion Lab and the Rochester Institute of Technology. Landsat-9 completed a three-month commissioning phase in January 2022 and has been operational since February 2022. The instrument has demonstrated excellent radiometric performance, as assessed from the on-orbit measurements. The TIRS instrument is radiometrically stable to 0.1% within a power cycle, and has noise levels below 0.1K. The lunar scans and the vicarious calibration data provide evidence that the stray light has been effectively mitigated.

The process of automatically masking objects from complex backgrounds is extremely beneficial when trying to utilize those objects for computer vision research, such as object detection, autonomous driving, pedestrian tracking, etc. Therefore, a robust method of segmentation is imperative towards ongoing research between the Digital Imaging and Remote Sensing Laboratory at the Rochester Institute of Technology and the Savannah River National Laboratory directed at the volume estimation of condense water vapor plumes emanating from mechanical draft cooling towers. Instance segmentation was performed on a custom data set consisting of RGB imagery with the Matterport Mask R-CNN implementation,1 where condensed water vapor plumes were masked out from mixed backgrounds for the purpose of 3D reconstruction and volume estimation. This multi-class Mask R-CNN was trained to detect cooling tower structure and plumes with and without data augmentation to study the effects on a preliminary data set, in addition to a model trained with a single plume class. The average precision and intersection over union metrics across all models were shown to not be statistically different. While each model is capable of detecting and segmenting plumes in the preliminary data set, all models essentially perform the task with the same efficacy. This indicates some level of bias in the preliminary data set, demonstrating the need for more variance in the form of additional annotated imagery. The single plume class model tested within 7% for mAP, AP, and IoU when compared to the other two models, demonstrating the ability of Mask R-CNN to detect and segment these dynamically-changing plumes without any spatial dependence on the stationary cooling tower structure. This ongoing research includes a long-term data collection campaign where imagery of condensed water vapor plumes will be continuously gathered over an 18-month period so as to include imagery examples under many different meteorological and environmental conditions, seasonal variations, and illumination changes that will occur over an annual cycle. Including this data in future training of the Mask R-CNN implementation is expected to reduce any bias that may exist in the current data set.

The Landsat-8 Thermal Infrared Sensor (TIRS) has been acquiring two-band thermal infrared images of the Earth's surface since 2013. The calibration of the two-band system has been monitored using the on-board calibrator and validated with vicarious calibration performed by NASA/JPL and RIT since launch. Soon after launch, it was discovered that the instrument had a significant stray light effect that was affecting the radiometric calibration. The stray light was corrected in the processing system in 2017. Since then, it has become apparent that there was an additional radiometric error, based on the vicarious calibration results. With a failure within the primary electronic system and subsequent switch to the redundant electronic system, the TIRS instrument effectively has two separate calibration regimes. The vicarious calibration found a statistically significant calibration error, primarily a constant over time, in Band 11 on the primary electronics (Feb 11, 2013 through March 5, 2015) of about -0.6K at 300K. The calibration error in Band 10 was smaller though still statistically significant at about 0.2K at 300K. On the redundant side (March 5, 2015 to present), the calibration error is more signal dependent than time dependent. Both bands are affected, with Band 10 having an error between 1K and -0.4K (between 273-320K) and Band 11 having an error between 0.8K and -1.44K (between 273-320K). This calibration error will be corrected within the USGS Landsat Product Generation System with the release of Landsat Collection-2 products. The Collection-2 release also includes a correction to the relative radiometric calibration of TIRS data. Striping as a result of poor detector-to-detector normalization has been increasing in the imagery since launch. The TIRS relative radiometric calibration will be updated based on internal calibrator data to remove the stripes on a quarterly basis. The visible stripes are generally at 0.1-0.2% level, though there are some detectors in each band that have changed by 1% or more. The Collection-2 release will result in much more uniform TIRS images.

The quality of grapes in the production of wine is highly influenced by vine water status, where optimal water deficit or selective harvesting can improve berry quality. It is in this context that the rapid advancement in small unmanned aerial system (sUAS) technology and the potential application of real-time, high-spatial resolution hyperspectral imagery for vineyard moisture assessment, have become tractable. This study sought to further sUAS hyperspectral imagery as a tool to model water status in a commercial vineyard in Upstate New York. High-spatial resolution (2.5 cm ground sample distance) hyperspectral data were collected in the visible/near-infrared (VNIR; 400-1000nm) regime on three flight days. A Scholander pressure chamber was used to directly measure the midday stem water potential (Ψstem) within imaged vines at the time of flight. High spatial resolution pixels enabled the targeting of pure (sunlit) vine canopy with vertically trained shoots and significant shadowing. We used the partial least squares-regression (PLS-R) modeling method to correlate our hyperspectral imagery with measured field water status and applied a wavelength band selection scheme to detect important wavelengths. We evaluated spectral smoothing and band reduction approaches, given signal-to-noise ratio (SNR) concerns. Our regression results indicated that unsmoothed curves, with the range of wavelength bands from 450- 1000 nm, provided the highest model performance with R2 = 0.68 for cross-validation. Future work will include hyperspectral flight data in the short-wave infrared (SWIR; 1000-2500 nm) regime that were also collected. Ultimately, models will need validation in different vineyards with a full range of plant stress.

The use of small unmanned aircraft systems (sUAS) for applications in the field of precision agriculture has demonstrated the need to produce temporally consistent imagery to allow for quantitative comparisons. In order for these aerial images to be used to identify actual changes on the ground, conversion of raw digital count to reflectance, or to an atmospherically normalized space, needs to be carried out. This paper will describe an experiment that compares the use of reflectance calibration panels, for use with the empirical line method (ELM), against a newly proposed ratio of the target radiance and the downwelling radiance, to predict the reflectance of known targets in the scene. We propose that the use of an on-board downwelling light sensor (DLS) may provide the sUAS remote sensing practitioner with an approach that does not require the expensive and time consuming task of placing known reflectance standards in the scene. Three calibration methods were tested in this study: 2-Point ELM, 1-Point ELM, and At-altitude Radiance Ratio (AARR). Our study indicates that the traditional 2-Point ELM produces the lowest mean error in band effective reflectance factor, 0.0165. The 1-Point ELM and AARR produce mean errors of 0.0343 and 0.0287 respectively. A modeling of the proposed AARR approach indicates that the technique has the potential to perform better than the 2-Point ELM method, with a 0.0026 mean error in band effective reflectance factor, indicating that this newly proposed technique may prove to be a viable alternative with suitable on-board sensors.

Landsat-8 has been operating on-orbit for 5+ years. Its two sensors, the Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS), are continuing to produce high quality data. The OLI has been radiometrically stable at the better than 0.3% level on a band average basis for all but the shortest wavelength (443 nm) band, which has degraded about 1.3% since launch. All on-board calibration devices continue to perform well and consistently. No gaps in across track coverage exist as 100% operability of the detectors is maintained. The variability over time of detector responsivity within a band relative to the average is better than 0.05% (1 sigma), though there are occasional detectors that jump up to 1.5% in response in the Short-Wave InfraRed (SWIR) bands. Signal-to-Noise performance continues at 2-3x better than requirements, with a small degradation in the 443 nm band commensurate with the loss in sensitivity. Pre-launch error analysis, combined with the stability of the OLI indicates that the absolute reflectance calibration uncertainty is better than 3%; comparisons to ground measurements and comparisons to other sensors are consistent with this. The Landsat-8 TIRS is similarly radiometrically stable, showing changes of at most 0.3% over the mission. The uncertainty in the absolute calibration as well as the detector to detector variability are largely driven by the stray light response of TIRS. The current processing corrects most of the stray light effects, resulting in absolute uncertainties of ~1% and reduced striping. Efforts continue to further reduce the striping. Noise equivalent delta temperature is about 50 mK at typical temperatures and 100% detector operability is maintained. Landsat-9 is currently under development with a launch no earlier than December 2020. The nearly identical OLI-2 and upgraded TIRS-2 sensors have completed integration and are in the process of instrument level performance characterization including spectral, spatial, radiometric and geometric testing. Component and assembly level measurements of the OLI-2, which include spectral response, radiometric response and stray light indicate comparable performance to OLI. The first functional tests occurred in July 2018 and spatial performance testing in vacuum is scheduled for August 2018. Similarly, for TIRS-2, partially integrated instrument level testing indicated spectral and spatial responses comparable to TIRS, with stray light reduced by approximately an order of magnitude from TIRS.

This study examines the hyperspectral reflectance characteristics of vegetation stressed by the influence of low-level sub-terrainean methane leakage from buried pipelines. The purpose is to ascertain whether high-spatial resolution spectral imagery can be used to geolocate small methane leaks in imagery collected from small unmanned aerial systems (sUAS). This could lead to rapid detection of methane leaks by finding spectrally unique regions of stressed vegetation which might benefit a variety of industries including utility inspectors, grounds maintenance crews, and construction personnel. This document describes an experiment to manually stress vegetation by introducing methane at a low ow rate beneath a layer of turf, allowing it to percolate to the surface and affect the vitality of the overlying turf. For comparison, a turf plot was stressed by root rot caused by overwatering, as well as a sample of turf used as a control area (healthy grass). The three areas of vegetation were observed daily over the course of a one-month period with a ground spectrometer to determine the onset and time line of damage to the vegetation. High-spatial resolution spectral imagery was also collected each day to observe wavelength characteristics of the damage. First derivative analysis was used alongside physiology-based indices and logistic regression to detect differences between healthy and stressed vegetation. The hyperspectral data showed that as vegetation is stressed the red-edge slope decreases along with values through the near infrared (NIR) while the short wave infrared (SWIR) region increases. The normalized difference index (NDI) calculation of stressed vegetation in relation to healthy vegetation is maximum using a ratio of reflectance values at 750 and 1910 nm. Conclusions will be presented as to whether sUAS may be used to determine if vegetation stressed by methane can be easily detected and which spectral bands are most effective for spotting this particular stressor.

The Thermal Infrared Sensor (TIRS) instrument is the thermal-band imager on the Landsat-8 platform. The initial on-orbit calibration estimates of the two TIRS spectral bands indicated large average radiometric calibration errors, -0.29 and -0.51 W/m(2) sr mu m or -2.1K and -4.4K at 300K in Bands 10 and 11, respectively, as well as high variability in the errors, 0.87K and 1.67K (1-sigma), respectively. The average error was corrected in operational processing in January 2014, though, this adjustment did not improve the variability. The source of the variability was determined to be stray light from far outside the field of view of the telescope. An algorithm for modeling the stray light effect was developed and implemented in the Landsat-8 processing system in February 2017. The new process has improved the overall calibration of the two TIRS bands, reducing the residual variability in the calibration from 0.87K to 0.51K at 300K for Band 10 and from 1.67K to 0.84K at 300K for Band 11. There are residual average lifetime bias errors in each band: 0.04 W/m(2) sr mu m (0.30K) and -0.04 W/m(2) sr mu m (-0.29K), for Bands 10 and 11, respectively.

This paper deals with a calibration algorithm to be used with data from the Thermal Infrared Sensor (TIRS) on board Landsat 8. Some non-uniform banding calibration errors have been observed in TIRS data since it was launched in 2013. Investigations have shown that this artifact is due to out-of-field radiance that scatters onto the TIRS focal plane. A calibration algorithm which utilizes TIRS image data itself to correct the stray light error has been proposed and implemented. Preliminary experiments have indicated this methodology reduces stray light artifacts significantly. However, there are some cases in which the TIRS TIRS method may not optimally mitigate stray light. These are special cases where there is a large temperature contrast between the edge of the TIRS image and of out-of-field radiance. This paper outlines an alternative approach with near-coincident image data from an external satellite sensor and compares the correction results with the current operational method, in general, and for some of the out-of-field special cases.