Sergio Bernabé

GPU Implementation of an Automatic Target Detection and Classification Algorithm for Hyperspectral Image Analysis

The detection of (moving or static) targets in remotely sensed hyperspectral images often requires real-time responses for swift decisions that depend upon high computing performance of algorithm analysis. The automatic target detection and classification algorithm (ATDCA) has been widely used for this purpose. In this letter, we develop several optimizations for accelerating the computational performance of ATDCA. The first one focuses on the use of the Gram-Schmidt orthogonalization method instead of the orthogonal projection process adopted by the classic algorithm. The second one is focused on the development of a new implementation of the algorithm on commodity graphics processing units (GPUs). The proposed GPU implementation properly exploits the GPU architecture at low level, including shared memory, and provides coalesced accesses to memory that lead to very significant speedup factors, thus taking full advantage of the computational power of GPUs. The GPU implementation is specifically tailored to hyperspectral imagery and the special characteristics of this kind of data, achieving real-time performance of ATDCA for the first time in the literature. The proposed optimizations are evaluated not only in terms of target detection accuracy but also in terms of computational performance using two different GPU architectures by NVIDIA: Tesla C1060 and GeForce GTX 580, taking advantage of the performance of operations in single-precision floating point. Experiments are conducted using hyperspectral data sets collected by three different hyperspectral imaging instruments. These results reveal considerable acceleration factors while retaining the same target detection accuracy for the algorithm.

Hyperspectral Unmixing on GPUs and Multi-Core Processors: A Comparison

One of the main problems in the analysis of remotely sensed hyperspectral data cubes is the presence of mixed pixels, which arise when the spatial resolution of the sensor is not able to separate spectrally distinct materials. Due to this reason, spectral unmixing has become one of the most important tasks for hyperspectral data exploitation. However, unmixing algorithms can be computationally very expensive, a fact that compromises their use in applications under real-time constraints. For this purpose, in this paper we develop two efficient implementations of a full hyperspectral unmixing chain on two different kinds of high performance computing architectures: graphics processing units (GPUs) and multi-core processors. The proposed full unmixing chain is composed for three stages: (i) estimation of the number of pure spectral signatures or endmembers, (ii) automatic identification of the estimated endmembers, and (iii) estimation of the fractional abundance of each endmember in each pixel of the scene. The two computing platforms used in this work are inter-compared in the context of hyperspectral unmixing applications. The GPU implementation of the proposed methodology has been implemented using the compute devide unified architecture (CUDA) and the cuBLAS library, and tested on two different GPU architectures: NVidia™ GeForce GTX 580 and NVidia™ Tesla C1060. It provides real-time unmixing performance in two different analysis scenarios using hyperspectral data collected by NASAs Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS) over the Cuprite mining district in Nevada and the World Trade Center complex in New York City. The multi-core implementation, developed using the applications program interface (API) OpenMP and the Intel Math Kernel Library (MKL) used for matrix multiplications, achieved near real-time performance in the same scenarios. A comparison of both architectures in terms of performance, cost and mission payload considerations is given based on the results obtained in the two considered data analysis scenarios.

Spectral–Spatial Classification of Multispectral Images Using Kernel Feature Space Representation

Over the last few years, several new strategies have been proposed for spectral-spatial classification of remotely sensed image data, for cases when high spatial and spectral resolutions are available. In this letter, we focus on the possibility of performing advanced spectral-spatial classification of remote sensing images with limited spectral resolution (often called multispectral). A new strategy is proposed, where the spectral dimensionality of the multispectral data is first expanded by using nonlinear feature extraction with kernel methods such as kernel principal component analysis. Then, extended multiattribute profiles (EMAPs), built on the expanded set of spectral features, are used to include spatial information. This strategy allows us to first decompose different spectral clusters into different spectral features and further improve the spatial discrimination. The resulting EMAPs are used for classification using advanced classifiers such as support vector machines and random forests. We test our proposed methodology with different multispectral data sets obtaining state-of-the-art classification results.

FPGA Design of an Automatic Target Generation Process for Hyperspectral Image Analysis

Onboard processing of remotely sensed hyper spectral data is a highly desirable goal in many applications. For this purpose, compact reconfigurable hardware modules such as field programmable gate arrays (FPGAs) are widely used. In this paper, we develop a new implementation of an automatic target generation process (ATGP) for hyper spectral images. Our implementation is based on a design methodology that starts from a high-level description in Matlab (or alternative C/C++) and obtains a register transfer level (RTL) description that can be ported to FPGAs. In order to validate our new implementation, we develop a quantitative and comparative study using two different FPGA architectures: Xilinx Virtex-5 and Altera Stratix-III Altera. Experimental results have been obtained in the context of a real application focused on the detection of mineral components over the Cup rite mining district (Nevada), using hyper spectral data collected by NASAs Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS). Our experimental results indicate that the proposed implementation can achieve peak frequency designs above 200MHz in the considered FPGAs, in addition to satisfactory results in terms of target detection accuracy and parallel performance. This represents a step forward towards the design of real-time onboard implementations of hyper spectral image analysis algorithms.

Unmixing-based content retrieval system for remotely sensed hyperspectral imagery on GPUs

This paper presents a new unmixing-based retrieval system for remotely sensed hyperspectral imagery. The need for this kind of system is justified by the exponential growth in the volume and number of remotely sensed data sets from the surface of the Earth. This is particularly the case for hyperspectral images, which comprise hundreds of spectral bands at different (almost contiguous) wavelength channels. To deal with the high computational cost of extracting the spectral information needed to catalog new hyperspectral images in our system, we resort to efficient implementations of spectral unmixing algorithms on commodity graphics processing units (GPUs). Spectral unmixing is a very popular approach for interpreting hyperspectral data with sub-pixel precision. This paper particularly focuses on the design of the proposed framework as a web service, as well as on the efficient implementation of the system on GPUs. In addition, we present a comparison of spectral unmixing algorithms available in the system on both CPU and GPU architectures.

Cloud Implementation of a Full Hyperspectral Unmixing Chain Within the NASA Web Coverage Processing Service for EO-1

The launch of the NASA Earth Observing 1 (EO-1) platform in November 2000 marked the establishment of spaceborne hyperspectral technology for land imaging. The Hyperion sensor onboard EO-1 operates in the 0.4-2.5 micrometer spectral range, with 10 nanometer spectral resolution and 30-meter spatial resolution. Spectral unmixing has been one of the most successful approaches to analyze Hyperion data since its launch. It estimates the abundance of spectrally pure constituents (endmembers) in each observation collected by the sensor. Due to the high spectral dimensionality of Hyperion data, unmixing is a very time-consuming operation. In this paper, we develop a cloud implementation of a full hyperspectral unmixing chain made up of the following steps: 1) dimensionality reduction; 2) automatic endmember identification; and 3) fully constrained abundance estimation. The unmixing chain will be available online within the Web Coverage Processing Service (WCPS), an image processing framework that can run on the cloud, as part of the NASA SensorWeb suite of web services. The proposed implementation has been demonstrated using the EO-1 Hyperion imagery. Our experimental results with a hyperspectral scene collected over the Okavango Basin in Botswana suggest the (present and future) potential of spectral unmixing for improved exploitation of spaceborne hyperspectral data. The integration of the unmixing chain in the WCPS framework as part of the NASA SensorWeb suite of web services is just the start of an international collaboration in which many more processing algorithms will be made available to the community through this service. This paper is not so much focused on the theory and results of unmixing (widely demonstrated in other contributions) but about the process and added value of the proposed contribution for ground processing on the cloud and onboard migration of those algorithms to support the generation of low-latency products for new airborne/spaceborne missions.

Performance versus energy consumption of hyperspectral unmixing algorithms on multi-core platforms

Hyperspectral imaging is a growing area in remote sensing in which an imaging spectrometer collects hundreds of images (at different wavelength channels) for the same area on the surface of the Earth. Hyperspectral images are extremely high-dimensional, and require on-board processing algorithms able to satisfy near real-time constraints in applications such as wildland fire monitoring, mapping of oil spills and chemical contamination, etc. One of the most widely used techniques for analyzing hyperspectral images is spectral unmixing, which allows for sub-pixel data characterization. This is particularly important since the available spatial resolution in hyperspectral images is typically of several meters, and therefore it is reasonable to assume that several spectrally pure substances (called endmembers in hyperspectral imaging terminology) can be found within each imaged pixel. There have been several efforts towards the efficient implementation of hyperspectral unmixing algorithms on architectures susceptible of being mounted onboard imaging instruments, including field programmable gate arrays (FPGAs) and graphics processing units (GPUs). While FPGAs are generally difficult to program, GPUs are difficult to adapt to onboard processing requirements in spaceborne missions due to its extremely high power consumption. In turn, with the increase in the number of cores, multi-core platforms have recently emerged as an easier to program platform compared to FPGAs, and also more tolerable radiation and power consumption requirements. However, a detailed assessment of the performance versus energy consumption of these architectures has not been conducted as of yet in the field of hyperspectral imaging, in which it is particularly important to achieve processing results in real-time. In this article, we provide a thoughtful perspective on this relevant issue and further analyze the performance versus energy consumption ratio of different processing chains for spectral unmixing when implemented on multi-core platforms.

FPGA Implementation of an Algorithm for Automatically Detecting Targets in Remotely Sensed Hyperspectral Images

Timely detection of targets continues to be a relevant challenge for hyperspectral remote sensing capability. The automatic target-generation process using an orthogonal projection operator (ATGP-OSP) has been widely used for this purpose. Hyperspectral target-detection applications require timely responses for swift decisions, which depend upon (near) real-time performance of algorithm analysis. Reconfigurable field-programmable gate arrays (FPGAs) are promising platforms that allow hardware/software codesign and the potential to provide powerful onboard computing capabilities and flexibility at the same time. In this paper, we present an FPGA implementation for the ATGP-OSP algorithm. Our system includes a direct memory access module and implements a prefetching technique to hide the latency of the input/output communications. The proposed method has been implemented on a Virtex-7 XC7VX690T FPGA and tested using real hyperspectral data collected by NASAs Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS) over the Cuprite mining district in Nevada and the World Trade Center in New York. Experimental results demonstrate that our hardware version of the ATGP-OSP algorithm can significantly outperform a software version, which makes our reconfigurable system appealing for onboard hyperspectral data processing.

A Web-Based System for Classification of Remote Sensing Data

The availability of satellite imagery has expanded over the past few years, and the possibility to perform fast processing of massive databases comprising this kind of imagery data has opened ground-breaking perspectives in many different fields. This paper describes a web-based system (available online: http://hypergim.ceta-ciemat.es), which allows an inexperienced user to perform unsupervised classification of satellite/airborne images. The processing chain adopted in this work has been implemented in C language and integrated in our proposed tool, developed with HTML5, JavaScript, Php, AJAX and other web programming languages. Image acquisition with the applications programmer interface (API) is fast and efficient. An important added functionality of the developed tool is its capacity to exploit a remote server to speed up the processing of large satellite/airborne images at different zoom levels. The ability to process images at different zoom levels allows the tool an improved interaction with the user, who is able to supervise the final result. The previous functionalities are necessary to use efficient techniques for the classification of images and the incorporation of content-based image retrieval (CBIR). Several experimental validation types of the classification results with the proposed system are performed by comparing the classification accuracy of the proposed chain by means of techniques available in the well-known Environment for Visualizing Images (ENVI) software package.

A new system to perform unsupervised and supervised classification of satellite images from Google Maps

In this paper, we describe a new system for unsupervised and supervised classification of satellite images from Google Maps. The system has been developed using the SwingX-WS library, and incorporates functionalities such as unsupervised classification of image portions selected by the user (at the maximum zoom level) using ISODATA and k-Means, and supervised classification using the Minimum Distance and Maximum Likelihood, followed by spatial post-processing based on majority voting. Selected regions in the classified portion are used to train a maximum likelihood classifier able to map larger image areas in a manner transparent to the user. The system also retrieves areas containing regions similar to those already classified. An experimental validation of the proposed system has been conducted by comparing the obtained classification results with those provided by commercial software, such as the popular Research Systems ENVI package.