Crop stage classification of hyperspectral data using unsupervised techniques

The presence of a large number of spectral bands in the hyperspectral images increases the capability to distinguish between various physical structures. However, they suffer from the high dimensionality of the data. Hence, the processing of hyperspectral images is applied in two stages: dimensionality reduction and unsupervised classification techniques. The high dimensionality of the data has been reduced with the help of Principal Component Analysis (PCA). The selected dimensions are classified using Niche Hierarchical Artificial Immune System (NHAIS). The NHAIS combines the splitting method to search for the optimal cluster centers using niching procedure and the merging method is used to group the data points based on majority voting. Results are presented for two hyperspectral images namely EO-1 Hyperion image and Indian pines image. A performance comparison of this proposed hierarchical clustering algorithm with the earlier three unsupervised algorithms is presented. From the results obtained, we deduce that the NHAIS is efficient.

Data processing of remotely sensed airborne hyperspectral data using the Airborne Processing Library (APL): Geocorrection algorithm descriptions and spatial accuracy assessment

Remote sensing airborne hyperspectral data are routinely used for applications including algorithm development for satellite sensors, environmental monitoring and atmospheric studies. Single flight lines of airborne hyperspectral data are often in the region of tens of gigabytes in size. This means that a single aircraft can collect terabytes of remotely sensed hyperspectral data during a single year. Before these data can be used for scientific analyses, they need to be radiometrically calibrated, synchronised with the aircraft's position and attitude and then geocorrected. To enable efficient processing of these large datasets the UK Airborne Research and Survey Facility has recently developed a software suite, the Airborne Processing Library (APL), for processing airborne hyperspectral data acquired from the Specim AISA Eagle and Hawk instruments. The APL toolbox allows users to radiometrically calibrate, geocorrect, reproject and resample airborne data. Each stage of the toolbox outputs data in the common Band Interleaved Lines (BILs) format, which allows its integration with other standard remote sensing software packages. APL was developed to be user-friendly and suitable for use on a workstation PC as well as for the automated processing of the facility; to this end APL can be used under both Windows and Linux environments on a single desktop machine or through a Grid engine. A graphical user interface also exists. In this paper we describe the Airborne Processing Library software, its algorithms and approach. We present example results from using APL with an AISA Eagle sensor and we assess its spatial accuracy using data from multiple flight lines collected during a campaign in 2008 together with in situ surveyed ground control points.

Data processing of remotely sensed airborne hyperspectral data using the Airborne Processing Library (APL): Geocorrection algorithm descriptions and spatial accuracy assessment

Remote sensing airborne hyperspectral data are routinely used for applications including algorithm development for satellite sensors, environmental monitoring and atmospheric studies. Single flight lines of airborne hyperspectral data are often in the region of tens of gigabytes in size. This means that a single aircraft can collect terabytes of remotely sensed hyperspectral data during a single year. Before these data can be used for scientific analyses, they need to be radiometrically calibrated, synchronised with the aircraft's position and attitude and then geocorrected. To enable efficient processing of these large datasets the UK Airborne Research and Survey Facility has recently developed a software suite, the Airborne Processing Library (APL), for processing airborne hyperspectral data acquired from the Specim AISA Eagle and Hawk instruments. The APL toolbox allows users to radiometrically calibrate, geocorrect, reproject and resample airborne data. Each stage of the toolbox outputs data in the common Band Interleaved Lines (BILs) format, which allows its integration with other standard remote sensing software packages. APL was developed to be user-friendly and suitable for use on a workstation PC as well as for the automated processing of the facility; to this end APL can be used under both Windows and Linux environments on a single desktop machine or through a Grid engine. A graphical user interface also exists. In this paper we describe the Airborne Processing Library software, its algorithms and approach. We present example results from using APL with an AISA Eagle sensor and we assess its spatial accuracy using data from multiple flight lines collected during a campaign in 2008 together with in situ surveyed ground control points.

Bathymetric map of Heron Reef, Australia, derived from airborne hyperspectral data at 1 m resolution

A simple method for efficient inversion of arbitrary radiative transfer models for image analysis is presented. The method operates by representing the shape of the function that maps model parameters to spectral reflectance by an adaptive look-up tree (ALUT) that evenly distributes the discretization error of tabulated reflectances in spectral space. A post-processing step organizes the data into a binary space partitioning tree that facilitates an efficient inversion search algorithm. In an example shallow water remote sensing application, the method performs faster than an implementation of previously published methodology and has the same accuracy in bathymetric retrievals. The method has no user configuration parameters requiring expert knowledge and minimizes the number of forward model runs required, making it highly suitable for routine operational implementation of image analysis methods. For the research community, straightforward and robust inversion allows research to focus on improving the radiative transfer models themselves without the added complication of devising an inversion strategy.

On independent component analysis applied to unmixing hyperspectral data

One of the most challenging task underlying many hyperspectral imagery applications is the spectral unmixing, which decomposes a mixed pixel into a collection of reectance spectra, called endmember signatures, and their corresponding fractional abundances. Independent Component Analysis (ICA) have recently been proposed as a tool to unmix hyperspectral data. The basic goal of ICA is to nd a linear transformation to recover independent sources (abundance fractions) given only sensor observations that are unknown linear mixtures of the unobserved independent sources. In hyperspectral imagery the sum of abundance fractions associated to each pixel is constant due to physical constraints in the data acquisition process. Thus, sources cannot be independent. This paper address hyperspectral data source dependence and its impact on ICA performance. The study consider simulated and real data. In simulated scenarios hyperspectral observations are described by a generative model that takes into account the degradation mechanisms normally found in hyperspectral applications. We conclude that ICA does not unmix correctly all sources. This conclusion is based on the a study of the mutual information. Nevertheless, some sources might be well separated mainly if the number of sources is large and the signal-to-noise ratio (SNR) is high.

Applying ISOMAP to the learning of hyperspectral image

In this paper, we present the application of a non-linear dimensionality reduction technique for the learning and probabilistic classification of hyperspectral image. Hyperspectral image spectroscopy is an emerging technique for geological investigations from airborne or orbital sensors. It gives much greater information content per pixel on the image than a normal colour image. This should greatly help with the autonomous identification of natural and manmade objects in unfamiliar terrains for robotic vehicles. However, the large information content of such data makes interpretation of hyperspectral images time-consuming and userintensive. We propose the use of Isomap, a non-linear manifold learning technique combined with Expectation Maximisation in graphical probabilistic models for learning and classification. Isomap is used to find the underlying manifold of the training data. This low dimensional representation of the hyperspectral data facilitates the learning of a Gaussian Mixture Model representation, whose joint probability distributions can be calculated offline. The learnt model is then applied to the hyperspectral image at runtime and data classification can be performed.

System-on-chip field-programmable gate array design for onboard real-time hyperspectral unmixing

Hyperspectral instruments have been incorporated in satellite missions, providing large amounts of data of high spectral resolution of the Earth surface. This data can be used in remote sensing applications that often require a real-time or near-real-time response. To avoid delays between hyperspectral image acquisition and its interpretation, the last usually done on a ground station, onboard systems have emerged to process data, reducing the volume of information to transfer from the satellite to the ground station. For this purpose, compact reconfigurable hardware modules, such as field-programmable gate arrays (FPGAs), are widely used. This paper proposes an FPGA-based architecture for hyperspectral unmixing. This method based on the vertex component analysis (VCA) and it works without a dimensionality reduction preprocessing step. The architecture has been designed for a low-cost Xilinx Zynq board with a Zynq-7020 system-on-chip FPGA-based on the Artix-7 FPGA programmable logic and tested using real hyperspectral data. Experimental results indicate that the proposed implementation can achieve real-time processing, while maintaining the methods accuracy, which indicate the potential of the proposed platform to implement high-performance, low-cost embedded systems, opening perspectives for onboard hyperspectral image processing.

Hyperspectral field measurements to determine dry matter yield and nutritive value of legume-grass swards

Summary: Productivity, botanical composition and forage quality of legume-grass swards are important factors for successful arable farming in both organic and conventional farming systems. As these attributes can vary considerably within a field, a non-destructive method of detection while doing other tasks would facilitate a more targeted management of crops, forage and nutrients in the soil-plant-animal system. This study was undertaken to explore the potential of field spectral measurements for a non destructive prediction of dry matter (DM) yield, legume proportion in the sward, metabolizable energy (ME), ash content, crude protein (CP) and acid detergent fiber (ADF) of legume-grass mixtures. Two experiments were conducted in a greenhouse under controlled conditions which allowed collecting spectral measurements which were free from interferences such as wind, passing clouds and changing angles of solar irradiation. In a second step this initial investigation was evaluated in the field by a two year experiment with the same legume-grass swards. Several techniques for analysis of the hyperspectral data set were examined in this study: four vegetation indices (VIs): simple ratio (SR), normalized difference vegetation index (NDVI), enhanced vegetation index (EVI) and red edge position (REP), two-waveband reflectance ratios, modified partial least squares (MPLS) regression and stepwise multiple linear regression (SMLR). The results showed the potential of field spectroscopy and proved its usefulness for the prediction of DM yield, ash content and CP across a wide range of legume proportion and growth stage. In all investigations prediction accuracy of DM yield, ash content and CP could be improved by legume-specific calibrations which included mixtures and pure swards of perennial ryegrass and of the respective legume species. The comparison between the greenhouse and the field experiments showed that the interaction between spectral reflectance and weather conditions as well as incidence angle of light interfered with an accurate determination of DM yield. Further research is hence needed to improve the validity of spectral measurements in the field. Furthermore, the developed models should be tested on varying sites and vegetation periods to enhance the robustness and portability of the models to other environmental conditions.

Linear Hyperspectral Unmixing Using L0-norm Approximations and Nonnegative Matrix Factorization

Spectral unmixing (SU) is a technique to characterize mixed pixels of the hyperspectral images measured by remote sensors. Most of the existing spectral unmixing algorithms are developed using the linear mixing models. Since the number of endmembers/materials present at each mixed pixel is normally scanty compared with the number of total endmembers (the dimension of spectral library), the problem becomes sparse. This thesis introduces sparse hyperspectral unmixing methods for the linear mixing model through two different scenarios. In the first scenario, the library of spectral signatures is assumed to be known and the main problem is to find the minimum number of endmembers under a reasonable small approximation error. Mathematically, the corresponding problem is called the $\ell_0$-norm problem which is NP-hard problem. Our main study for the first part of thesis is to find more accurate and reliable approximations of $\ell_0$-norm term and propose sparse unmixing methods via such approximations. The resulting methods are shown considerable improvements to reconstruct the fractional abundances of endmembers in comparison with state-of-the-art methods such as having lower reconstruction errors. In the second part of the thesis, the first scenario (i.e., dictionary-aided semiblind unmixing scheme) will be generalized as the blind unmixing scenario that the library of spectral signatures is also estimated. We apply the nonnegative matrix factorization (NMF) method for proposing new unmixing methods due to its noticeable supports such as considering the nonnegativity constraints of two decomposed matrices. Furthermore, we introduce new cost functions through some statistical and physical features of spectral signatures of materials (SSoM) and hyperspectral pixels such as the collaborative property of hyperspectral pixels and the mathematical representation of the concentrated energy of SSoM for the first few subbands. Finally, we introduce sparse unmixing methods for the blind scenario and evaluate the efficiency of the proposed methods via simulations over synthetic and real hyperspectral data sets. The results illustrate considerable enhancements to estimate the spectral library of materials and their fractional abundances such as smaller values of spectral angle distance (SAD) and abundance angle distance (AAD) as well.

FPGA-based architecture for hyperspectral endmember extraction

Hyperspectral instruments have been incorporated in satellite missions, providing data of high spectral resolution of the Earth. This data can be used in remote sensing applications, such as, target detection, hazard prevention, and monitoring oil spills, among others. In most of these applications, one of the requirements of paramount importance is the ability to give real-time or near real-time response. Recently, onboard processing systems have emerged, in order to overcome the huge amount of data to transfer from the satellite to the ground station, and thus, avoiding delays between hyperspectral image acquisition and its interpretation. For this purpose, compact reconfigurable hardware modules, such as field programmable gate arrays (FPGAs) are widely used. This paper proposes a parallel FPGA-based architecture for endmember’s signature extraction. This method based on the Vertex Component Analysis (VCA) has several advantages, namely it is unsupervised, fully automatic, and it works without dimensionality reduction (DR) pre-processing step. The architecture has been designed for a low cost Xilinx Zynq board with a Zynq-7020 SoC FPGA based on the Artix-7 FPGA programmable logic and tested using real hyperspectral data sets collected by the NASA’s Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS) over the Cuprite mining district in Nevada. Experimental results indicate that the proposed implementation can achieve real-time processing, while maintaining the methods accuracy, which indicate the potential of the proposed platform to implement high-performance, low cost embedded systems, opening new perspectives for onboard hyperspectral image processing.

Signal subspace identification in hyperspectral imagery

Terrestrial remote sensing imagery involves the acquisition of information from the Earth's surface without physical contact with the area under study. Among the remote sensing modalities, hyperspectral imaging has recently emerged as a powerful passive technology. This technology has been widely used in the fields of urban and regional planning, water resource management, environmental monitoring, food safety, counterfeit drugs detection, oil spill and other types of chemical contamination detection, biological hazards prevention, and target detection for military and security purposes [2-9]. Hyperspectral sensors sample the reflected solar radiation from the Earth surface in the portion of the spectrum extending from the visible region through the near-infrared and mid-infrared (wavelengths between 0.3 and 2.5 µm) in hundreds of narrow (of the order of 10 nm) contiguous bands [10]. This high spectral resolution can be used for object detection and for discriminating between different objects based on their spectral xharacteristics [6]. However, this huge spectral resolution yields large amounts of data to be processed. For example, the Airbone Visible/Infrared Imaging Spectrometer (AVIRIS) [11] collects a 512 (along track) X 614 (across track) X 224 (bands) X 12 (bits) data cube in 5 s, corresponding to about 140 MBs. Similar data collection ratios are achieved by other spectrometers [12]. Such huge data volumes put stringent requirements on communications, storage, and processing. The problem of signal sbspace identification of hyperspectral data represents a crucial first step in many hypersctral processing algorithms such as target detection, change detection, classification, and unmixing. The identification of this subspace enables a correct dimensionality reduction (DR) yelding gains in data storage and retrieval and in computational time and complexity. Additionally, DR may also improve algorithms performance since it reduce data dimensionality without losses in the useful signal components. The computation of statistical estimates is a relevant example of the advantages of DR, since the number of samples required to obtain accurate estimates increases drastically with the dimmensionality of the data (Hughes phnomenon) [13].

Hyperspectral unmixing with simultaneous dimensionality estimation

This paper is an elaboration of the simplex identification via split augmented Lagrangian (SISAL) algorithm (Bioucas-Dias, 2009) to blindly unmix hyperspectral data. SISAL is a linear hyperspectral unmixing method of the minimum volume class. This method solve a non-convex problem by a sequence of augmented Lagrangian optimizations, where the positivity constraints, forcing the spectral vectors to belong to the convex hull of the endmember signatures, are replaced by soft constraints. With respect to SISAL, we introduce a dimensionality estimation method based on the minimum description length (MDL) principle. The effectiveness of the proposed algorithm is illustrated with simulated and real data.

Hyperspectral sensing to detect the impact of herbicide drift on cotton growth and yield

Yield loss in crops is often associated with plant disease or external factors such as environment, water supply and nutrient availability. Improper agricultural practices can also introduce risks into the equation. Herbicide drift can be a combination of improper practices and environmental conditions which can create a potential yield loss. As traditional assessment of plant damage is often imprecise and time consuming, the ability of remote and proximal sensing techniques to monitor various bio-chemical alterations in the plant may offer a faster, non-destructive and reliable approach to predict yield loss caused by herbicide drift. This paper examines the prediction capabilities of partial least squares regression (PLS-R) models for estimating yield. Models were constructed with hyperspectral data of a cotton crop sprayed with three simulated doses of the phenoxy herbicide 2,4-D at three different growth stages. Fibre quality, photosynthesis, conductance, and two main hormones, indole acetic acid (IAA) and abscisic acid (ABA) were also analysed. Except for fibre quality and ABA, Spearman correlations have shown that these variables were highly affected by the chemical. Four PLS-R models for predicting yield were developed according to four timings of data collection: 2, 7, 14 and 28 days after the exposure (DAE). As indicated by the model performance, the analysis revealed that 7 DAE was the best time for data collection purposes (RMSEP = 2.6 and R2 = 0.88), followed by 28 DAE (RMSEP = 3.2 and R2 = 0.84). In summary, the results of this study show that it is possible to accurately predict yield after a simulated herbicide drift of 2,4-D on a cotton crop, through the analysis of hyperspectral data, thereby providing a reliable, effective and non-destructive alternative based on the internal response of the cotton leaves.

Spectral and thermal sensing for nitrogen and water status in rainfed and irrigated wheat environments

Variable-rate technologies and site-specific crop nutrient management require real-time spatial information about the potential for response to in-season crop management interventions. Thermal and spectral properties of canopies can provide relevant information for non-destructive measurement of crop water and nitrogen stresses. In previous studies, foliage temperature was successfully estimated from canopy-scale (mixed foliage and soil) temperatures and the multispectral Canopy Chlorophyll Content Index (CCCI) was effective in measuring canopy-scale N status in rainfed wheat (Triticum aestivum L.) systems in Horsham, Victoria, Australia. In the present study, results showed that under irrigated wheat systems in Maricopa, Arizona, USA, the theoretical derivation of foliage temperature unmixing produced relationships similar to those in Horsham. Derivation of the CCCI led to an r2 relationship with chlorophyll a of 0.53 after Zadoks stage 43. This was later than the relationship (r2 = 0.68) developed for Horsham after Zadoks stage 33 but early enough to be used for potential mid-season N fertilizer recommendations. Additionally, ground-based hyperspectral data estimated plant N (g kg)1) in Horsham with an r2 = 0.86 but was confounded by water supply and N interactions. By combining canopy thermal and spectral properties, varying water and N status can potentially be identified eventually permitting targeted N applications to those parts of a field where N can be used most efficiently by the crop.