High Spatial Resolution Hyperspectral Imagery Research Articles

Mapping invasive noxious weed species in the alpine grassland ecosystems using very high spatial resolution UAV hyperspectral imagery and a novel deep learning model

ABSTRACT The term “invasive noxious weed species” (INWS), which refers to noxious weed plants that invade native alpine grasslands, has increasingly become an ecological and economic threat in the alpine grassland ecosystem of the Qinghai-Tibetan Plateau (QTP). Both the INWS and native grass species are small in physical size and share a habitat. Using remote sensing data to distinguish INWS from native alpine grass species remains a challenge. High spatial resolution hyperspectral imagery provides an alternative for addressing this problem. Here, we explored the use of unmanned aerial vehicle (UAV) hyperspectral imagery and deep learning methods with a small sample size for mapping the INWS in mixed alpine grasslands. To assess the method, UAV hyperspectral data with a very high spatial resolution of 2 cm were collected from the study site, and a novel convolutional neural network (CNN) model called 3D&2D-INWS-CNN was developed to take full advantage of the rich information provided by the imagery. The results indicate that the proposed 3D&2D-INWS-CNN model applied to the collected imagery for mapping INWS and native species with small ground truth training samples is robust and sufficient, with an overall classification accuracy exceeding 95% and a kappa value of 98.67%. The F1 score for each native species and INWS ranged from 92% to 99%. In conclusion, our results highlight the potential of using very high spatial resolution UAV hyperspectral data combined with a state-of-the-art deep learning model for INWS mapping even with small training samples in degraded alpine grassland ecosystems. Studies such as ours can aid the development of invasive species management practices and provide more data for decision-making in controlling the spread of invasive species in similar grassland ecosystems or, more widely, in terrestrial ecosystems.

A method for detecting underground natural gas pipeline micro-leakage in vegetated areas using high spatial resolution hyperspectral imagery

Natural gas is an important clean energy source that is mainly transported through buried pipelines. However, slight leakage during transportation is inevitable and needs to be identified in time to reduce the risk of related accidents. For this purpose, remote sensing can be used as an indirect and non-destructive technique to detect natural gas leakage through vegetation. Abundant spectral and detailed spatial information enables the high spatial resolution hyperspectral imagery to capture subtle stress responses from vegetation. In this study, we designed a simulation experiment of natural gas leakage in vegetated areas, and collected the high spatial resolution hyperspectral images of three plant species. Considering the shadow problem contained in high spatial resolution images, a new vegetation pixel extraction method to eliminate the influence of background shadows was proposed. Then, two spectral transformation methods, continuum removal (CR) and variational mode decomposition (VMD), were integrated to analyze the vegetation spectra, and a spectral index which could reflect the degree of vegetation stress was designed based on the spectral characteristics, named CVI (Continuum removal - Variational mode decomposition Index). Finally, the threshold segmentation methods and the cumulative minimum circumscribed circle method were used to extract the range and location of natural gas leakage. The study demonstrated the potential of using high spatial hyperspectral imagery to map and monitor natural gas leakage through its stress on vegetation. Results of this research can be a reference for the fine identification of stressed plants and are expected to be applied to provide support for related departments and stakeholders.

Mapping native and invasive grassland species and characterizing topography-driven species dynamics using high spatial resolution hyperspectral imagery

Characterizing the distribution, mechanism, and behaviour of invasive species is crucial to implementing an effective plan to protect and manage native grassland ecosystems. Hyperspectral remote sensing has been used to map and monitor invasive species at various spatial and temporal scales. However, most studies focus either on invasive tree species mapping or on the landscape level using coarse-spatial resolution imagery. These coarse-resolution images are not fine enough to distinguish individual invasive grasses, especially in a heterogeneous environment where invasive species are small, fragmented, and co-existent with native plants with similar color and texture. To capture the small yet highly dynamic invasive plants at different stages of the growing season and under various topography and hydrological conditions, we use airborne high-resolution narrow-band hyperspectral imagery (HrHSI) to map invasive species in a heterogeneous grassland ecosystem in southern Ontario, Canada. The results show that there is high spectral and textural separability between two invasive species and between invasive and native plants, leading to an overall species classification accuracy of up to 89.6%. The combination of resultant species-level maps and the digital elevation model (DEM) showed that seasonality is the dominant factor that drives the distribution of invasive species at the landscape level, while small-scale topographic variations partially explain local patches of invasive species. This study provides insights into the feasibility of using HrHSI in mapping invasive species in a heterogeneous ecosystem and offers the means to understand the mechanism and behaviour of invasive species for a more effective grassland management strategy.

Correction of Sunglint Effects in High Spatial Resolution Hyperspectral Imagery Using SWIR or NIR Bands and Taking Account of Spectral Variation of Refractive Index of Water

Passive high spatial resolution hyperspectral and multispectral imaging systems in the solar spectral region from aircraft and satellite platforms are being increasingly used for remote sensing of coastal waters and inland lakes. However, the remotely sensed data are often contaminated by the specular reflection of solar radiation at the air/water interface. Thus, the effects of sunglint need to be corrected. The purpose of sunglint correction is to remove the undesired specular reflected portion of radiance that does not carry any information about the water body and its bottom surface properties and to obtain the water-leaving radiance from the measured total radiance. In general, adequate treatment of the spectral variation of the sunglint effect with due consideration of the physics and the spectroscopy principles has not been done in previously reported hyperspectral sunglint removing techniques. In this study, a practical sunglint removing technique using either a shortwave infrared (SWIR) band or a near-infrared (NIR) band has been developed, where the liquid water absorption is sufficiently large, and the water-leaving radiance (in units of reflectance) is negligible. In the proposed method, proper consideration has been taken for the spectral Fresnel reflection in the wavelength range of 0.35–2.5 mm. Four cases of sunglint removal from hyperspectral imaging data acquired from two airborne imaging spectrometers have been presented.

Characterization of Mineralogy in the Highland Valley Porphyry Cu District Using Hyperspectral Imaging, and Potential Applications

The Highland Valley Copper (HVC) district in British Columbia, Canada, is host to at least four major porphyry Cu systems: Bethlehem (~209 Ma), and Valley, Lornex, and Highmont (~208 to 207 Ma). High spatial resolution (0.2–1.0 mm/pixel) hyperspectral imagery in the shortwave infrared (SWIR) were acquired on 755 rock samples and 400 m of continuous drill core. Spectral metrics are used to measure the relative abundance of 12 minerals and an additional metric is derived to estimate white mica grain size. In the Valley and Lornex deposits, coarse-grained white mica is associated with mineralization and is detectable up to 4 km away from the deposits. Kaolinite is present within 2 km of the mineralized centers but does not necessarily occur within strongly mineralized intervals. Prehnite is ubiquitous from 4 to 8 km from the deposits. In the Bethlehem deposit, tourmaline and epidote are associated with mineralization. We propose a spectral alteration score based on these proximal hyperspectral SWIR mineralogical patterns to assist explorers in targeting porphyry Cu systems when using drill core, surface rock samples and potentially remote sensing imagery. In a production environment, this metric could serve to facilitate ore-sorting.

Land cover mapping in urban environments using hyperspectral APEX data: A study case in Baden, Switzerland

High spatial resolution hyperspectral imagery has shown considerable potential for deriving accurate land cover maps in urban areas. In this paper, a new classification framework for mapping land cover in urban environments using high spatial resolution hyperspectral data was proposed. The proposed classification scheme was applied to map urban land cover using APEX data in the city of Baden, Switzerland. We first used the NDWI and NDVI indices to separate the land cover in the scene into three main classes: water, vegetation and non-vegetated surface. Then we partitioned the scene into many superpixels and applied classification using a SVM separately on the vegetation and non-vegetated surfaces. Soil was classified both in vegetation and non-vegetated surface, and these two soil results were merged in the final classification map. Shadows were initially classified in shaded vegetation surfaces and shaded non-vegetated surfaces, and then they were further classified into meaningful land cover categories. Our experimental results demonstrate that the proposed classification framework is well suited for mapping land cover in urban environments using high resolution hyperspectral data. Although the proposed method performs better than traditional methods in terms of soil classification accuracy, our findings emphasize that the soil class should be interpreted with caution in urban land cover maps derived from remote sensing data, even when high spatial resolution hyperspectral data are used. Results from this study also demonstrate that although shaded surfaces are generally classified as a single category in urban environments, in high resolution hyperspectral data, the shadows can be further classified into meaningful land cover classes with an acceptable accuracy.

Airborne Hyperspectral Evaluation of Maximum Gross Photosynthesis, Gravimetric Water Content, and CO2 Uptake Efficiency of the Mer Bleue Ombrotrophic Peatland

Peatlands cover a large area in Canada and globally (12% and 3% of the landmass, respectively). These ecosystems play an important role in climate regulation through the sequestration of carbon dioxide from, and the release of methane to, the atmosphere. Monitoring approaches, required to understand the response of peatlands to climate change at large spatial scales, are challenged by their unique vegetation characteristics, intrinsic hydrological complexity, and rapid changes over short periods of time (e.g., seasonality). In this study, we demonstrate the use of multitemporal, high spatial resolution (1 m2) hyperspectral airborne imagery (Compact Airborne Spectrographic Imager (CASI) and Shortwave Airborne Spectrographic Imager (SASI) sensors) for assessing maximum instantaneous gross photosynthesis (PGmax) in hummocks, and gravimetric water content (GWC) and carbon uptake efficiency in hollows, at the Mer Bleue ombrotrophic bog. We applied empirical models (i.e., in situ data and spectral indices) and we derived spatial and temporal trends for the aforementioned variables. Our findings revealed the distribution of hummocks (51.2%), hollows (12.7%), and tree cover (33.6%), which is the first high spatial resolution map of this nature at Mer Bleue. For hummocks, we found growing season PGmax values between 8 μmol m−2 s−1 and 12 μmol m−2 s−1 were predominant (86.3% of the total area). For hollows, our results revealed, for the first time, the spatial heterogeneity and seasonal trends for gravimetric water content and carbon uptake efficiency for the whole bog.

Spatio-Temporal Relationships between Optical Information and Carbon Fluxes in a Mediterranean Tree-Grass Ecosystem

Spatio-temporal mismatches between Remote Sensing (RS) and Eddy Covariance (EC) data as well as spatial heterogeneity jeopardize terrestrial Gross Primary Production (GPP) modeling. This article combines: (a) high spatial resolution hyperspectral imagery; (b) EC footprint climatology estimates; and (c) semi-empirical models of increasing complexity to analyze the impact of these factors on GPP estimation. Analyses are carried out in a Mediterranean Tree-Grass Ecosystem (TGE) that combines vegetation with very different physiologies and structure. Half-hourly GPP (GPPhh) were predicted with relative errors ~36%. Results suggest that, at EC footprint scale, the ecosystem signals are quite homogeneous, despite tree and grass mixture. Models fit using EC and RS data with high degree of spatial and temporal match did not significantly improved models performance; in fact, errors were explained by meteorological variables instead. In addition, the performance of the different models was quite similar. This suggests that none of the models accurately represented light use efficiency or the fraction of absorbed photosynthetically active radiation. This is partly due to model formulation; however, results also suggest that the mixture of the different vegetation types might contribute to hamper such modeling, and should be accounted for GPP models in TGE and other heterogeneous ecosystems.

Mapping Spartina alterniflora Biomass Using LiDAR and Hyperspectral Data

Large-scale coastal reclamation has caused significant changes in Spartina alterniflora (S. alterniflora) distribution in coastal regions of China. However, few studies have focused on estimation of the wetland vegetation biomass, especially of S. alterniflora, in coastal regions using LiDAR and hyperspectral data. In this study, the applicability of LiDAR and hypersectral data for estimating S. alterniflora biomass and mapping its distribution in coastal regions of China was explored to attempt problems of wetland vegetation biomass estimation caused by different vegetation types and different canopy height. Results showed that the highest correlation coefficient with S. alterniflora biomass was vegetation canopy height (0.817), followed by Normalized Difference Vegetation Index (NDVI) (0.635), Atmospherically Resistant Vegetation Index (ARVI) (0.631), Visible Atmospherically Resistant Index (VARI) (0.599), and Ratio Vegetation Index (RVI) (0.520). A multivariate linear estimation model of S. alterniflora biomass using a variable backward elimination method was developed with R squared coefficient of 0.902 and the residual predictive deviation (RPD) of 2.62. The model accuracy of S. alterniflora biomass was higher than that of wetland vegetation for mixed vegetation types because it improved the estimation accuracy caused by differences in spectral features and canopy heights of different kinds of wetland vegetation. The result indicated that estimated S. alterniflora biomass was in agreement with the field survey result. Owing to its basis in the fusion of LiDAR data and hyperspectral data, the proposed method provides an advantage for S. alterniflora mapping. The integration of high spatial resolution hyperspectral imagery and LiDAR data derived canopy height had significantly improved the accuracy of mapping S. alterniflora biomass.

Hierarchical feature learning with dropout k-means for hyperspectral image classification

A huge volume of high spatial resolution hyperspectral imagery (HSI) data sets can currently be acquired. However, making full use of the information within the HSI is still a huge problem. The exploitation of spatial information is playing a more and more important role in the classification of remote sensing data. How to efficiently extract the spatial feature for HSI has become a critical task. In this paper, we propose a dropout k-means based framework to extract an effective hierarchical spatial feature for HSI. This paper focuses on unsupervised hierarchical feature learning representation. The proposed framework was tested on two HSIs. The extensive experimental results clearly show that the proposed dropout k-means based framework achieves a superior classification performance.

Hyperspectral imagery super-resolution by compressive sensing inspired dictionary learning and spatial-spectral regularization.

Due to the instrumental and imaging optics limitations, it is difficult to acquire high spatial resolution hyperspectral imagery (HSI). Super-resolution (SR) imagery aims at inferring high quality images of a given scene from degraded versions of the same scene. This paper proposes a novel hyperspectral imagery super-resolution (HSI-SR) method via dictionary learning and spatial-spectral regularization. The main contributions of this paper are twofold. First, inspired by the compressive sensing (CS) framework, for learning the high resolution dictionary, we encourage stronger sparsity on image patches and promote smaller coherence between the learned dictionary and sensing matrix. Thus, a sparsity and incoherence restricted dictionary learning method is proposed to achieve higher efficiency sparse representation. Second, a variational regularization model combing a spatial sparsity regularization term and a new local spectral similarity preserving term is proposed to integrate the spectral and spatial-contextual information of the HSI. Experimental results show that the proposed method can effectively recover spatial information and better preserve spectral information. The high spatial resolution HSI reconstructed by the proposed method outperforms reconstructed results by other well-known methods in terms of both objective measurements and visual evaluation.

Assessing the Application of Cloud–Shadow Atmospheric Correction Algorithm on HICO

Several ocean color earth observation satellite sensors are presently collecting daily imagery, including the Hyperspectral Imager for the Coastal Ocean (HICO). HICO has been operating aboard the International Space Station since its installation on September 24, 2009. It provides high spatial resolution hyperspectral imagery optimized for the coastal ocean. Atmospheric correction, however, still remains a challenge for this sensor, particularly in optically complex coastal waters. In this paper, we assess the application of the cloud-shadow atmospheric correction approach on HICO data and validate the results with the in situ data. We also use multiple sets of cloud, shadow, and sunlit pixels to correct a single image multiple times and intercompare the results to assess variability in the retrieved reflectance spectra. Retrieved chlorophyll values from this intercomparison are similar and also agree well with the in situ chlorophyll measurements.

High spectral and spatial resolution hyperspectral imagery for quantifying Russian wheat aphid infestation in wheat using the constrained energy minimization classifier

The effects of insect infestation in agricultural crops are of major ecological and eco- nomic interest because of reduced yield, increased cost of pest control and increased risk of environmental contamination from insecticide application. The Russian wheat aphid (RWA, Diuraphis noxia) is an insect pest that causes damage to wheat (Triticum aestivum L.). We pro- posed that concentrated RWA feeding areas, referred to as hot spots, could be identified and isolated from uninfested areas within a field for site specific aphid management using remotely sensed data. Our objectives were to (1) investigate the reflectance characteristics of infested and uninfested wheat by RWA and (2) evaluate utility of airborne hyperspectral imagery with 1-m spatial resolution for detecting, quantifying, and mapping RWA infested areas in commercial winter wheat fields using the constrained energy minimization classifier. Percent surface reflec- tance from uninfested wheat was lower in the visible and higher in the near infrared portions of the spectrum when compared with RWA-infested wheat. The overall classification accuracies of >89% for damage detection were achieved. These results indicate that hyperspectral imagery can be effectively used for accurate detection and quantification of RWA infestation in wheat for site-specific aphid management. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI. (DOI: 10.1117/1.JRS.8.083661)

A neighbourhood-constrained k-means approach to classify very high spatial resolution hyperspectral imagery

In classifying very high spatial resolution (VHR) hyperspectral imagery, intra-class variation often adversely affects classification accuracy, mainly due to a low signal-to-noise ratio (SNR) and high spatial heterogeneity. To address this problem, this article develops a neighbourhood-constrained k-means (NC-k-means) algorithm by incorporating the pure neighbourhood index into the traditional k-means algorithm. The performance of the NC-k-means algorithm was assessed through a series of simulated images and a real hyperspectral image. The results indicate that the classification accuracy of NC-k-means algorithm is consistently better than that of the traditional k-means algorithm, in particular for the images with significant spatial autocorrelations among neighbouring pixels.

Hyperspectral imagery super-resolution by sparse representation and spectral regularization

For the instrument limitation and imperfect imaging optics, it is difficult to acquire high spatial resolution hyperspectral imagery. Low spatial resolution will result in a lot of mixed pixels and greatly degrade the detection and recognition performance, affect the related application in civil and military fields. As a powerful statistical image modeling technique, sparse representation can be utilized to analyze the hyperspectral image efficiently. Hyperspectral imagery is intrinsically sparse in spatial and spectral domains, and image super-resolution quality largely depends on whether the prior knowledge is utilized properly. In this article, we propose a novel hyperspectral imagery super-resolution method by utilizing the sparse representation and spectral mixing model. Based on the sparse representation model and hyperspectral image acquisition process model, small patches of hyperspectral observations from different wavelengths can be represented as weighted linear combinations of a small number of atoms in pre-trained dictionary. Then super-resolution is treated as a least squares problem with sparse constraints. To maintain the spectral consistency, we further introduce an adaptive regularization terms into the sparse representation framework by combining the linear spectrum mixing model. Extensive experiments validate that the proposed method achieves much better results.

Balloon imagery verification of remotely sensed Phragmites australis expansion in an urban estuary of New Jersey, USA

The invasion of the exotic common reed ( Phragmites australis) is increasingly displacing local native species from northeastern coastal estuaries. This study evaluates the accuracy of a remote sensing technique to map the distribution of common reed, monitor the rate of invasion and determine areas of natural resistance to invasion. The current invasion footprint of Phragmites in the Hackensack Meadowlands District in Northern New Jersey was determined using high spectral and spatial resolution hyperspectral imagery. A tethered balloon-based imaging device with limited coverage area was used to assess the accuracy of the hyperspectral imagery classification. The accuracy assessment based on true color balloon images revealed that the hyperspectral classification technique from images covering hundreds of hectares was 90% accurate in separating the dominant common reed-invaded areas from the native vegetation. Furthermore, linear spectral un-mixing techniques for sub-pixel classification revealed that for mixed areas where Phragmites covered 75% or more of a pixel, the classification was correct 96% of the time. The accuracy dropped to 52% for pixels that contained 25% or less of Phragmites cover, and was only 4% for pixels where invasive and native species cover was the same (50–50%).

A first assessment of the use of high spatial resolution hyperspectral imagery in discriminating among animal species, and between animals and their surroundings

This paper presents the results of a first evaluation of the feasibility of using data from a high spatial resolution hyperspectral sensor to discriminate among animal species, and between animals and their surroundings. The evaluation was performed by simulating 1000 pixels produced by a shortwave infrared (1051–1789 nm) hyperspectral sensor with a pixel size of either 2 or 4 m, and a spectral resolution of 11.5 nm. The simulated pixels consisted of five classes: pure vegetation, and mixtures of four animal species (cattle, horses, sheep, and pigs) with vegetation. The data were processed by converting from reflectance to absorption, taking the first difference of the data, and then using stepwise discriminant analysis to create discriminant functions based on 250 of the 1000 pixels. These discriminant functions were then used to assign the remaining 750 pixels to classes. The results indicate that the spectra of the four animal species are quite different from each other, and that these differences include variations in the presence and strength of absorptions in the shortwave infrared portion of the spectrum. Overall accuracies were 90% based on the 2 m data and 61% based on the 4 m data. User's and producer's accuracies for individual classes based on the 2 m data ranged from 72 to 100%, but were lower for the 4 m data (40–90%). These results suggest that further study of the use of high spatial resolution hyperspectral imagery in locating and identifying animals is merited.

Mapping Downy Brome (Bromus tectorum) Using Multidate AVIRIS Data

Invasive plants impose threats to both natural and managed ecosystems. Downy brome is among the most aggressive invasive weeds that has infested the shrub-steppe ecoregion of eastern Washington. Hyperspectral remote sensing has potential for early detection and for monitoring the spread of downy brome—information that is essential for developing effective management strategies. Two airborne hyperspectral Advanced Visible Infrared Imaging Spectrometer (AVIRIS) images (electromagnetic spectrum ranging from 400 to 2,500 nm) were acquired at a nominal 4-m ground resolution over a study area in south-central Washington on July 27, 2000 and May 5, 2003. We used a mixture-tuned matched filtering (MTMF) algorithm to classify downy brome and predict its percent cover in each dataset plus a merged multiseasonal dataset using the transformed bands from a minimum noise fraction (MNF) output. The correlation coefficient was 0.79, calculated for the multidate MTMF predicted downy brome abundance, compared to 0.41 and 0.51 derived from the July 2000 and May 2003 data, respectively. Although this study used high spatial resolution (∼3 to 4 m) hyperspectral imagery, this result shows that data acquired in different seasons is more effective for detection of downy brome invasion, compared to single-date datasets. These results support expanded use of multitemporal data for weed mapping to capitalize on spectral differences between seasons for weeds, in this case downy brome, and the surrounding environment.

Detection of estuarine and tidal river hydromorphology using hyper-spectral and LiDAR data: Forth estuary, Scotland

High spatial resolution hyper-spectral imagery (CASI) and light detection and ranging (LiDAR) imagery acquired for the tidal River Carron and Forth estuary, Scotland, were used in conjunction with field surveys to assess the feasibility of monitoring hydromorphology and human alterations with satellite and airborne remote sensing data. The study was undertaken in the context of the European Union Water Framework Directive (WFD) that requires member states to monitor hydromorphological elements as a component of the ecological status of rivers, estuaries and shorelines. Visual assessment and automated classifications of the imagery were compared with field survey data for an estuarine reach comprising saline waters, mudflats, a tidal reach of a tributary river and an urban/industrialised shoreline. The morphology of the estuary and inflowing tidal waters together with most artificial features of interest could be clearly seen in the CASI imagery at 1 m spatial resolution. Supervised classification of the imagery produced an overall accuracy value of 72%. Downgrading the imagery to simulate the spatial resolution of 4 m IKONOS satellite data surprisingly improved the accuracy to 74%. Simulation of 10 m SPOT imagery resulted in an image where many artificial features of interest such as roads, pipelines and jetties were rendered invisible. Adding LiDAR data as an additional data set aided manual and automated identification of features and visualisation of the hydromorphology of the rivers and estuaries in the study area. Shadows cast from tall objects were a feature of the winter imagery and reduced automated classification accuracy. Overall, the study demonstrates that high spatial resolution remotely sensed digital imagery has the potential to be a useful tool for panoptic mapping of the geomorphology and human impact on tidal rivers and estuaries. In the context of the WFD, remote sensing provides a potential way forward for monitoring the physical status of rivers and estuaries at the national scale. The possibilities and constraints, in light of the findings of this study, are discussed.

High spatial resolution hyperspectral mapping of in-stream habitats, depths, and woody debris in mountain streams

This article evaluates the potential of 1-m resolution, 128-band hyperspectral imagery for mapping in-stream habitats, depths, and woody debris in third- to fifth-order streams in the northern Yellowstone region. Maximum likelihood supervised classification using principal component images provided overall classification accuracies for in-stream habitats (glides, riffles, pools, and eddy drop zones) ranging from 69% for third-order streams to 86% for fifth-order streams. This scale dependency of classification accuracy was probably driven by the greater proportion of transitional boundary areas in the smaller streams. Multiple regressions of measured depths ( y) versus principal component scores ( x 1, x 2,…, x n ) generated R 2 values ranging from 67% for high-gradient riffles to 99% for glides in a fifth-order reach. R 2 values were lower in third-order reaches, ranging from 28% for runs and glides to 94% for pools. The less accurate depth estimates obtained for smaller streams probably resulted from the relative increase in the number of mixed pixels, where a wide range of depths and surface turbulence occurred within a single pixel. Matched filter (MF) mapping of woody debris generated overall accuracies of 83% in the fifth-order Lamar River. Accuracy figures for the in-stream habitat and wood mapping may have been misleadingly low because the fine-resolution imagery captured fine-scale variations not mapped by field teams, which in turn generated false “misclassifications” when the image and field maps were compared. The use of high spatial resolution hyperspectral (HSRH) imagery for stream mapping is limited by the need for clear water to measure depth, by any tree cover obscuring the stream, and by the limited availability of airborne hyperspectral sensors. Nonetheless, the high accuracies achieved in northern Yellowstone streams indicate that HSRH imagery can be a powerful tool for watershed-wide mapping, monitoring, and modeling of streams.