Sensitivity of spectral reflectance values to different burn and vegetation ratios: A multi-scale approach applied in a fire affected area

The aim of our study was to explore the spectral properties of fire-scorched (burned) and non fire-scorched (vegetation) areas, as well as areas with different burn/vegetation ratios, using a multisource multiresolution satellite data set. A case study was undertaken following a very destructive wildfire that occurred in Parnitha, Greece, July 2007, for which we acquired satellite images from LANDSAT, ASTER, and IKONOS. Additionally, we created spatially degraded satellite data over a range of coarser resolutions using resampling techniques. The panchromatic (1 m) and multispectral component (4 m) of IKONOS were merged using the Gram-Schmidt spectral sharpening method. This very high-resolution imagery served as the basis to estimate the cover percentage of burned areas, bare land and vegetation at pixel level, by applying the maximum likelihood classification algorithm. Finally, multiple linear regression models were fit to estimate each land-cover fraction as a function of surface reflectance values of the original and the spatially degraded satellite images.The main findings of our research were: (a) the Near Infrared (NIR) and Short-wave Infrared (SWIR) are the most important channels to estimate the percentage of burned area, whereas the NIR and red channels are the most important to estimate the percentage of vegetation in fire-affected areas; (b) when the bi-spectral space consists only of NIR and SWIR, then the NIR ground reflectance value plays a more significant role in estimating the percent of burned areas, and the SWIR appears to be more important in estimating the percent of vegetation; and (c) semi-burned areas comprising 45–55% burned area and 45–55% vegetation are spectrally closer to burned areas in the NIR channel, whereas those areas are spectrally closer to vegetation in the SWIR channel. These findings, at least partially, are attributed to the fact that: (i) completely burned pixels present low variance in the NIR and high variance in the SWIR, whereas the opposite is observed in completely vegetated areas where higher variance is observed in the NIR and lower variance in the SWIR, and (ii) bare land modifies the spectral signal of burned areas more than the spectral signal of vegetated areas in the NIR, while the opposite is observed in SWIR region of the spectrum where the bare land modifies the spectral signal of vegetation more than the burned areas because the bare land and the vegetation are spectrally more similar in the NIR, and the bare land and burned areas are spectrally more similar in the SWIR.     

Automated lithological mapping by integrating spectral enhancement techniques and machine learning algorithms using AVIRIS-NG hyperspectral data in Gold-bearing granite-greenstone rocks in Hutti,India

In this study, we proposed an automated lithological mapping approach by using spectral enhancement techniques and Machine Learning Algorithms (MLAs) using Airborne Visible Infrared Imaging Spectroradiometer-Next Generation (AVIRIS-NG) hyperspectral data in the greenstone belt of the Hutti area, India. We integrated spectral enhancement techniques such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA) transformation and different MLAs for an accurate mapping of rock types. A conjugate utilization of conventional geological map and spectral enhancement products derived from ASTER data were used for the preparation of a high-resolution reference lithology map. Feature selection and extraction methods were applied on the AVIRIS-NG data to derive different input dataset such as (a) all spectral bands, (b) shortwave infrared bands, (c) Joint Mutual Information Maximization (JMIM) based optimum bands, and (d) optimum bands using PCA, to choose optimum input dataset for automated lithological mapping. The comparative analysis of different MLAs shows that the Support Vector Machine (SVM) outperforms other Machine Learning (ML) models. The SVM achieved an Overall Accuracy (OA) and Kappa Coefficient (k) of 85.48% and 0.83, respectively, using JMIM based optimum bands. The JMIM based optimum bands were more suitable than other input datasets to classify most of the lithological units (i.e. metabasalt, amphibolite, granite, acidic intrusive and migmatite) within the study area . The sensitivity analysis performed in this study illustrates that the SVM is less sensitive to the number of samples and mislabeling in the model training than other MLAs. The obtained high-resolution classified map with accurate litho-contacts of amphibolite, metabasalt, and granite can be coupled with an alteration map of the area for targeting the potential zone of gold mineralization.     

Classifiers vs. input variables—The drivers in image classification for land cover mapping

The study investigates the performance of image classifiers for landscape-scale land cover mapping and the relevance of ancillary data for the classification success in order to assess and to quantify the importance of these components in image classification. Specifically tested are the performance of maximum likelihood classification (MLC), artificial neural networks (ANN) and discriminant analysis (DA) based on Landsat7 ETM+ spectral data in combination with topographic measures and NDVI. ANN produced high accuracies of more than 75% also with limited input information, while MLC and DA produced comparable results only by incorporating ancillary data into the classification process. The superiority of ANN classification was less pronounced on the level of the single land cover classes. The use of ancillary data generally increased classification accuracy and showed a similar potential for increasing classification accuracy than the selection of the classifier. Therefore, a stronger focus on the development of appropriate and optimised sets of input variables is suggested. Also the definition and selection of land cover classes has shown to be crucial and not to be simply adaptable from existing land cover class schemes. A stronger research focus towards discriminating land cover classes by their typical spectral, topographic or seasonal properties is therefore suggested to advance image classification.     

Landsat8和MODIS融合构建高时空分辨率数据识别秋粮作物

本文利用Wu等人提出的遥感数据时空融合方法 STDFA(Spatial Temporal Data Fusion Approach)以Landsat 8和MODIS为数据源构建高时间、空间分辨率的遥感影像数据。以此为基础,构建15种30 m分辨率分类数据集,然后利用支持向量机SVM(Support Vector Machine)进行秋粮作物识别,验证不同维度分类数据集进行秋粮作物识别的适用性。实验结果显示,不同分类数据集的秋粮作物分类结果均达到了较高的识别精度。综合各项精度指标分析,Red+Phenology数据组合对秋粮识别效果最好,水稻识别的制图精度和用户精度分别达到91.76%和82.49%,玉米识别的制图精度和用户精度分别达到85.80%和74.97%,水稻和玉米识别的总体精度达到86.90%。     

Enhancing the performance of regional land cover mapping

Different pixel-based, object-based and subpixel-based methods such as time-series analysis, decision-tree, and different supervised approaches have been proposed to conduct land use/cover classification. However, despite their proven advantages in small dataset tests, their performance is variable and less satisfactory while dealing with large datasets, particularly, for regional-scale mapping with high resolution data due to the complexity and diversity in landscapes and land cover patterns, and the unacceptably long processing time. The objective of this paper is to demonstrate the comparatively highest performance of an operational approach based on integration of multisource information ensuring high mapping accuracy in large areas with acceptable processing time. The information used includes phenologically contrasted multiseasonal and multispectral bands, vegetation index, land surface temperature, and topographic features. The performance of different conventional and machine learning classifiers namely Malahanobis Distance (MD), Maximum Likelihood (ML), Artificial Neural Networks (ANNs), Support Vector Machines (SVMs) and Random Forests (RFs) was compared using the same datasets in the same IDL (Interactive Data Language) environment. An Eastern Mediterranean area with complex landscape and steep climate gradients was selected to test and develop the operational approach. The results showed that SVMs and RFs classifiers produced most accurate mapping at local-scale (up to 96.85% in Overall Accuracy), but were very time-consuming in whole-scene classification (more than five days per scene) whereas ML fulfilled the task rapidly (about 10 min per scene) with satisfying accuracy (94.2–96.4%). Thus, the approach composed of integration of seasonally contrasted multisource data and sampling at subclass level followed by a ML classification is a suitable candidate to become an operational and effective regional land cover mapping method.     

Exploring image fusion of ALOS/PALSAR data and LANDSAT data to differentiate forest area

Remote sensing data utilize valuable information via various satellite sensors that have different specifications. Image fusion allows the user to combine different spatial and spectral resolutions to improve the information for purposes such as forest monitoring and land cover mapping. In this study, I assessed the contribution of dual-polarized Advanced Land Observing Satellite/Phased Array type L-band Synthetic Aperture Radar data to multispectral Landsat imagery. The research investigated the separability of forested areas using different image fusion techniques. Quality analysis of the fused images was conducted using qualitative and quantitative analyses. I applied the support vector machine image classification method for land cover mapping. Among all methods examined, the à trous wavelet transform method best differentiated the forested area with an overall accuracy (OA) of 94.316%, while Landsat had an OA of 92.626%. The findings of this study indicated that optical-SAR-fused images improve land cover classification, which results in higher quality forest inventory data and mapping.     

Study on the utility of IRS-P6 AWiFS SWIR band for crop discrimination and classification

This present study was conducted to find out the usefulness of SWIR (Short Wave Infra Red) band data in AWiFS (Advanced Wide Field Sensor) sensor of Resourcesat 1, for the discrimination of different Rabi season crops (rabi rice, groundnut and vegetables) and other vegetations of the undivided Cuttack district of Orissa state. Four dates multi-spectral AWiFS data during the period from 10 December 2003 to 2 May 2004 were used. The analysis was carried out using various multivariate statistics and classification approaches. Principal Component Analysis (PCA) and separability measures were used for selection of best bands for crop discrimination. The analysis showed that, for discrimination of the crops in the study area, NIR was found to be the best band, followed by SWIR and Red. The results of the supervised MXL classification showed that inclusion of SWIR band increased the overall accuracy and kappa coefficient. The ‘Three Band Ratio’ index, which incorporated Red, NIR and SWIR bands, showed improved discrimination in the multi-date dataset classification, compared to other SWIR based indices.     

Analyzing fine-scale wetland composition using high resolution imagery and texture features

In order to monitor natural and anthropogenic disturbance effects to wetland ecosystems, it is necessary to employ both accurate and rapid mapping of wet graminoid/sedge communities. Thus, it is desirable to utilize automated classification algorithms so that the monitoring can be done regularly and in an efficient manner. This study developed a classification and accuracy assessment method for wetland mapping of at-risk plant communities in marl prairie and marsh areas of the Everglades National Park. Maximum likelihood (ML) and Support Vector Machine (SVM) classifiers were tested using 30.5 cm aerial imagery, the normalized difference vegetation index (NDVI), first and second order texture features and ancillary data. Additionally, appropriate window sizes for different texture features were estimated using semivariogram analysis. Findings show that the addition of NDVI and texture features increased classification accuracy from 66.2% using the ML classifier (spectral bands only) to 83.71% using the SVM classifier (spectral bands, NDVI and first order texture features).     

应用遥感及地理信息系统进行植被制图

利用卫星遥感数据制作复杂地形环境的植被图面临的最主要问题是精度，单纯对遥感数据（ＴＭ或ＳＰＯＩ）进行监督或非监督分类的精度低于５０％。本文选择美国亚利桑那州ＳａｎｔａＣａｔａｌｉｎａ山脉的ＰｕｓｃｈＲｉｄｇｅ作为研究区，分析地理信息系统模型在改善植被分类精度中的作用。结果表明，通过结合辅助数据和应用地理信息系统模型，其精度可以从３７．４１％提高到７１．６７％（ＳＰＯＴ数据，非监督分类），或从５０．０７％提高到６１．５０％（ＴＭ数据，监督分类）。同时表明用ＳＰＯＴ数据进行山区植被制图的效果好于ＴＭ数据。     

Fusion of RADARSAT-2 and multispectral optical remote sensing data for LULC extraction in a tropical agricultural area

In this study, we investigated the performance of different fusion and classification techniques for land cover mapping in Hilir Perak, Peninsula Malaysia using RADAR and Landsat-8 images in a predominantly agricultural area. The fusion methods used are Brovey Transform, Wavelet Transform, Ehlers and Layer Stacking and their results classified into seven different land cover classes which include (1) pixel-based classifiers (spectral angle mapper (SAM), maximum likelihood (ML), support vector machine (SVM)) and (2) Object-based (rule-based and standard nearest neighbour (NN)) classifiers. The result shows that pixel-based classification achieved maximum accuracy of the optical data classification using SVM in Landsat-8 with 74.96% accuracy compared to SAM and ML. For multisource data classification, the highest overall accuracy recorded for layer stacking (SVM) was 79.78%, Ehlers fusion (SVM) with 45.57%, Brovey fusion (SVM) with 63.70% and Wavelet fusion (SVM) 61.16%. And for object-based classifiers, the overall classification accuracy is 95.35% for rule-based and 76.33% for NN classifier, respectively. Based on the analysis of their performances, object-based and the rule-based classifiers produced the best classification accuracy from the fused images.     

Ground validation of an algorithm for estimating surface suspended sediment concentrations from multi-spectral reflectance data

Pollution of water resources by sediments eroded from degraded watersheds is a critical concern around the world. Current methods for locating these eroding areas and off-site damage to water resources through visual observations and field sampling with subsequent laboratory analysis are time consuming and expensive. There is thus, a justified interest in developing algorithms for quick estimation of suspended sediment concentrations in large water-bodies from remotely sensed data. This paper presents the results of a ground validation study on characterization and quantification of surface suspended sediment concentrations (SSC) in sediment laden water bodies through an n-waveband specific numerical index, total information content. A comparison of SSC-predictive potential of the proposed new index, derived from four broad (100–300 nm) Landsat MSS, five broad (40–300 nm) Landsat TM and eight narrow (20–40 nm) IRS-P4 OCM spectral bands, with that of the conventional (NIR-Red and NIR+Red) indices, computed from the same spectral band data, is also presented. The study reveaied that at SSCs 250 mg/1, the proposed index (derived from either broad / narrow landsat MSS/TM or IRS-P4 OCM spectral data) could lead to SSC predictions (with mean errors within 20%) comparable with those obtained with the conventional indices (derived from the same spectral band data). It could further be observed that, in general, lower sediment concentrations (i.e. SSCs 150 mg/1) were associated with higher prediction inaccuracies. A comparison of the mean errors of predictions associated with the proposed and the conventional (NIR-Red and NIR+Red) indices computed from broad and narrow band data for SSCs 150 mg/I, revealed that an increase in number of wavebands (from 4 MSS to 5 TM or 8 OCM bands) and a decrease in the bandwidth of these wavebands (from broad MSS/ TM bands to narrow OCM bands) led to a significant increase in the prediction accuracy of the proposed new index. These prediction accuracies were observed to be the highest with the proposed index calculated from narrow OCM-P4 spectral data. However this could not be observed with the conventional indices at any of the SSC ranges and with the proposed index at SSCs 250 mg/l. This shows that the lower SSC-predictive potential of proposed index was a significant function of both the number and the bandwidth of spectral bands used for its computation. In fact in one of the cases, lower SSC (150 mg/l) -predictive accuracy of the proposed index was found to be significantly higher than that of the conventional (NIR+R) index. The proposed algorithm could thus compress the information contained in the entire reflectance spectrum of the sediment laden water bodies to their sediment type and concentration specific characteristic values. This characteristic of the proposed index was not shared by any of the conventional indices, based on only two waveband data. In fact the proposed index appears to be the only mean of completely compressing and quantifying the information contained in all the information channels of a narrow band spectrometer (consisting of 200 wavebands) to be shortly launched by ISRO for satellite based inventory of natural resources.     

Evaluation of Landsat TM data for forest cover type and landuse classification in subtropical forests of Kumaon Himalaya (U.P.)

LANDSAT-TM has been evaluated for forest cover type and landuse classification in subtropical forests of Kumaon Himalaya (U.P.) Comparative evaluation of false colour composite generated by using various band combinations has been made. Digital image processing of Landsat-TM data on VIPS-32 RRSSC computer system has been carried out to stratify vegetation types. Conventional band combination in false colour composite is Bands 2, 3 and 4 in Red/Green/Blue sequence of Landsat TM for landuse classification. The present study however suggests that false colour combination using Landsat TM bands viz., 4, 5 and 3 in Red/Green/Blue sequence is the most suitable for visual interpretation of various forest cover types and landuse classes. It is felt that to extract full information from increased spatial and spectral resolution of Landsat TM, it is necessary to process the data digitally to classify land cover features like vegetation. Supervised classification using maximum likelihood algorithm has been attemped to stratify the forest vegetation. Only four bands are sufficient enough to classify vegetaton types. These bands are 2,3,4 and 5. The classification results were smoothed digitaly to increase the readiability of the map. Finally, the classification carred out using digital technique were evaluated using systematic sampling design. It is observed that forest cover type mapping can be achieved upto 80% overall mapping accuracy. Monospecies stand Chirpine can be mapped in two density classes viz., dense pine (<40%) with more than 90% accuracy. Poor accuracy (66%) was observed while mapping pine medium dense areas. The digital smoothening reduced the overall mapping accuracy. Conclusively, Landsat-TM can be used as operatonal sensor for forest cover type mapping even in complex landuse-terrain of Kumaon Himalaya (U.P.)     

Feature reduction using a singular value decomposition for the iterative guided spectral class rejection hybrid classifier

Feature reduction in a remote sensing dataset is often desirable to decrease the processing time required to perform a classification and improve overall classification accuracy. This paper introduces a feature reduction method based on the singular value decomposition (SVD). This SVD-based feature reduction method reduces the storage and processing requirements of the SVD by utilizing a training dataset. This feature reduction technique was applied to training data from two multitemporal datasets of Landsat TM/ETM+ imagery acquired over a forested area in Virginia, USA and Rondônia, Brazil. Subsequent parallel iterative guided spectral class rejection (pIGSCR) forest/non-forest classifications were performed to determine the quality of the feature reduction. The classifications of the Virginia data were five times faster using SVD-based feature reduction without affecting the classification accuracy. Feature reduction using the SVD was also compared to feature reduction using principal components analysis (PCA). The highest average accuracies for the Virginia dataset (88.34%) and for the Rondônia dataset (93.31%) were achieved using the SVD. The results presented here indicate that SVD-based feature reduction can produce statistically significantly better classifications than PCA.     

Assessing the potential of multi-seasonal WorldView-2 imagery for mapping West African agroforestry tree species

High resolution satellite systems enable efficient and detailed mapping of tree cover, with high potential to support both natural resource monitoring and ecological research. This study investigates the capability of multi-seasonal WorldView-2 imagery to map five dominant tree species at the individual tree crown level in a parkland landscape in central Burkina Faso. The Random Forest algorithm is used for object based tree species classification and for assessing the relative importance of WorldView-2 predictors. The classification accuracies from using wet season, dry season and multi-seasonal datasets are compared to gain insights about the optimal timing for image acquisition. The multi-seasonal dataset produced the most accurate classifications, with an overall accuracy (OA) of 83.4%. For classifications based on single date imagery, the dry season (OA = 78.4%) proved to be more suitable than the wet season (OA = 68.1%). The predictors that contributed most to the classification success were based on the red edge band and visible wavelengths, in particular green and yellow. It was therefore concluded that WorldView-2, with its unique band configuration, represents a suitable data source for tree species mapping in West African parklands. These results are particularly promising when considering the recently launched WorldView-3, which provides data both at higher spatial and spectral resolution, including shortwave infrared bands.     

Remote sensing of Red Mediterranean soils: a case study in the viticultural southern Rhone Valley using SPOT satellite imagery

The aim of this work was to map Red Mediterranean soils, for which no previous mapping approach exists, using optical multispectral satellite remote sensing data. This case study explores the use of SPOT XS images over the viticultural Southern Rhone Valley, France, to map exposed vineyard soils. Field spectral measurements were used to distinguish Red Mediterranean soil surfaces during Spring 1999. A supervised maximum likelihood classification was applied to sparsely vegetated and unvegetated surfaces of two spring images from 1995 and 1997, drawing on the field training set and available soil data. Similar global spatial segmentation was obtained despite different soil surface states on these dates. Classification performances were higher than 84% in both images. Mean classification accuracies of Red Mediterranean soils at seven reference surfaces were 60% in 1995 and 70% in 1997. This suggests that the direct use of optical remote sensing data at medium resolution can be useful for mapping bare Red Mediterranean soils.     

Assessing the performance of two unsupervised dimensionality reduction techniques on hyperspectral APEX data for high resolution urban land-cover mapping

Despite the high richness of information content provided by airborne hyperspectral data, detailed urban land-cover mapping is still a challenging task. An important topic in hyperspectral remote sensing is the issue of high dimensionality, which is commonly addressed by dimensionality reduction techniques. While many studies focus on methodological developments in data reduction, less attention is paid to the assessment of the proposed methods in detailed urban hyperspectral land-cover mapping, using state-of-the-art image classification approaches. In this study we evaluate the potential of two unsupervised data reduction techniques, the Autoassociative Neural Network (AANN) and the BandClust method – the first a transformation based approach, the second a feature-selection based approach – for mapping of urban land cover at a high level of thematic detail, using an APEX 288-band hyperspectral dataset. Both methods were tested in combination with four state-of-the-art machine learning classifiers: Random Forest (RF), AdaBoost (ADB), the multiple layer perceptron (MLP), and support vector machines (SVM). When used in combination with a strong learner (MLP, SVM) BandClust produces classification accuracies similar to or higher than obtained with the full dataset, demonstrating the method’s capability of preserving critical spectral information, required for the classifier to successfully distinguish between the 22 urban land-cover classes defined in this study. In the AANN data reduction process, on the other hand, important spectral information seems to be compromised or lost, resulting in lower accuracies for three of the four classifiers tested. Detailed analysis of accuracies at class level confirms the superiority of the SVM/Bandclust combination for accurate urban land-cover mapping using a reduced hyperspectral dataset. This study also demonstrates the potential of the new APEX sensor data for detailed mapping of land cover in spatially and spectrally complex urban areas.     

Spectral angle mapper and object-based classification combined with hyperspectral remote sensing imagery for obtaining land use/cover mapping in a Mediterranean region

In this study, we test the potential of two different classification algorithms, namely the spectral angle mapper (SAM) and object-based classifier for mapping the land use/cover characteristics using a Hyperion imagery. We chose a study region that represents a typical Mediterranean setting in terms of landscape structure, composition and heterogeneous land cover classes. Accuracy assessment of the land cover classes was performed based on the error matrix statistics. Validation points were derived from visual interpretation of multispectral high resolution QuickBird-2 satellite imagery. Results from both the classifiers yielded more than 70% classification accuracy. However, the object-based classification clearly outperformed the SAM by 7.91% overall accuracy (OA) and a relatively high kappa coefficient. Similar results were observed in the classification of the individual classes. Our results highlight the potential of hyperspectral remote sensing data as well as object-based classification approach for mapping heterogeneous land use/cover in a typical Mediterranean setting.     

Mapping forest leaf area index using reflectance and textural information derived from WorldView-2 imagery in a mixed natural forest area in Florida,US

The leaf area index (LAI) of plant canopies is an important structural parameter that controls energy, water, and gas exchanges of plant ecosystems. Remote sensing techniques may offer an alternative for measuring and mapping forest LAI at a landscape scale. Given the characteristics of high spatial/spectral resolution of the WorldView-2 (WV2) sensor, it is of significance that the textural information extracted from WV2 multispectral (MS) bands will be first time used in estimating and mapping forest LAI. In this study, LAI mapping accuracies would be compared from (a) spatial resolutions between 2-m WV2 MS data and 30-m Landsat TM imagery, (b) the nature of variables between spectrum-based features and texture-based features, and (c) sensors between TM and WV2. Therefore spectral/textural features (SFs) were first selected and tested; then a canonical correlation analysis was performed with different data sets of SFs and LAI measurement; and finally linear regression models were used to predict and map forest LAI with canonical variables calculated from image data. The experimental results demonstrate that for estimating and mapping forest LAI, (i) using high resolution data (WV2) is better than using relatively low resolution data (TM); (ii) extracted from the same WV2 data, texture-based features have higher capability than that of spectrum-based features; (iii) a combination of spectrum-based features with texture-based features could lead to even higher accuracy of mapping forest LAI than their either one separately; and (iv) WV2 sensor outperforms TM sensor significantly. However, we need to address the possible overfitting phenomenon that might be brought in by using more input variables to develop models. In addition, the experimental results also indicate that the red-edge band in WV2 was the worst on estimating LAI among WV2 MS bands and the WV2 MS bands in the visible range had a much higher correlation with ground measured LAI than that red-edge and NIR bands did.     

Satellite Data Classification Using Open Source Support

In this study we explored the potential of open source data mining software support to classify freely available Landsat image. The study identified several major classes that can be distinguished using Landsat data of 30 m spatial resolution. Decision tree classification (DTC) using Waikato environment for knowledge analysis (WEKA), open source software is used to prepare land use land cover (LULC) map and the result is compared with supervised (maximum likelihood classifier – MLC) and unsupervised (Iterative self-organizing data analysis technique - ISODATA clustering) classification techniques. The accuracy assessment indicates highest accuracy of the map prepared using DTC with overall accuracy (OA) 92 % (kappa = 0.90) followed by MLC with OA 88 % (kappa = 0.84) and ISODATA OA 76 % (kappa = 0.69). Results indicate that data set with a good definition of training sites can produce LULC map having good overall accuracy using decision tree. The paper demonstrates utility of open source system for information extraction and importance of DTC algorithm.     

Evaluating Hyperspectral Imager for the Coastal Ocean (HICO) data for seagrass mapping in Indian River Lagoon,FL

Differentiation between benthic habitats, particularly seagrass and macroalgae, using satellite data is complicated because of water column effects plus the presence of chlorophyll-a in both seagrass and algae that result in similar spectral patterns. Hyperspectral imager for the coastal ocean data over the Indian River Lagoon, Florida, USA, was used to develop two benthic classification models, Slope_RED and Slope_NIR. Their performance was compared with iterative self-organizing data analysis technique and spectral angle mapping classification methods. The slope models provided greater overall accuracies (63–64%) and were able to distinguish between seagrass and macroalgae substrates more accurately compared to the results obtained using the other classifications methods.