Abstract
PhenoSat is an experimental software tool that extracts phenological information from satellite vegetation index time-series. Temporal satellite NDVI data provided by VEGETATION sensor from three different vegetation types (Vineyard, Closed Deciduous Forest and Deciduous Shrubland with Sparse Trees) and for different geographical locations were used to test the ability of the software in extracting vegetation dynamics information. Six noise reduction filters were tested: piecewise-logistic, Savitzky-Golay, cubic smoothing splines, Gaussian models, Fourier series and polynomial curve fitting. The results showed that PhenoSat is an useful tool to extract phenological NDVI metrics, providing similar results to those obtained from field measurements. The best results presented correlations of 0.89 (n=6; p<0.01) and 0.71 (n=6; p<0.06) for the green-up and maximum stages, respectively. In the fitting process, the polynomial and Gaussian algorithms over smoothed the peak related with a double-growth season, the opposite to the other methods that could detect more accurately this peak.

Abstract

Abstract
A fiber optic sensor for simultaneous measurement of refractive index and temperature is presented. The sensing probe is achieved by introducing multimode interference inside a high birefringence fiber loop mirror resulting in a configuration capable of refractive index and temperature discrimination. The multimode interference peak is sensitive to the surrounding refractive index (90 nm/RIU) and slightly responsive to the temperature (0.005 nm/degrees C). On the other hand, the birrefringent fiber loop mirror is highly sensitive to temperature (2.39 nm/degrees C) and has no response to refractive index. Therefore, a temperature independent refractive index measurement can be made with a resolution of +/- 2.5x10(-5).

Abstract
Field-Programmable Gate Arrays (FPGAs) are becoming increasingly important in embedded and high-performance computing systems. They allow performance levels close to the ones obtained with Application-Specific Integrated Circuits, while still keeping design and implementation flexibility. However, to efficiently program FPGAs, one needs the expertise of hardware developers in order to master hardware description languages (HDLs) such as VHDL or Verilog. Attempts to furnish a high-level compilation flow (e.g., from C programs) still have to address open issues before broader efficient results can be obtained. Bearing in mind an FPGA available resources, it has been developed LALP (Language for Aggressive Loop Pipelining), a novel language to program FPGA-based accelerators, and its compilation framework, including mapping capabilities. The main ideas behind LALP are to provide a higher abstraction level than HDLs, to exploit the intrinsic parallelism of hardware resources, and to allow the programmer to control execution stages whenever the compiler techniques are unable to generate efficient implementations. Those features are particularly useful to implement loop pipelining, a well regarded technique used to accelerate computations in several application domains. This paper describes LALP, and shows how it can be used to achieve high-performance computing solutions.

Abstract
Given the large amount of data mining algorithms, their combinations (e.g. ensembles) and possible parameter settings, finding the most adequate method to analyze a new dataset becomes an ever more challenging task. This is because in many cases testing all possibly useful alternatives quickly becomes prohibitively expensive. In this paper we propose a novel technique, called active testing, that intelligently selects the most useful cross-validation tests. It proceeds in a tournament-style fashion, in each round selecting and testing the algorithm that is most likely to outperform the best algorithm of the previous round on the new dataset. This 'most promising' competitor is chosen based on a history of prior duels between both algorithms on similar datasets. Each new cross-validation test will contribute information to a better estimate of dataset similarity, and thus better predict which algorithms are most promising on the new dataset. We have evaluated this approach using a set of 292 algorithm-parameter combinations on 76 UCI datasets for classification. The results show that active testing will quickly yield an algorithm whose performance is very close to the optimum, after relatively few tests. It also provides a better solution than previously proposed methods. © 2012 Springer-Verlag.

Abstract
This paper presents a visual localization approach that is suitable for domestic and industrial environments as it enables accurate, reliable and robust pose estimation. The mobile robot is equipped with a single camera which update sits pose whenever a landmark is available on the field of view. The innovation presented by this research focuses on the artificial landmark system which has the ability to detect the presence of the robot, since both entities communicate with each other using an infrared signal protocol modulated in frequency. Besides this communication capability, each landmark has several high intensity light-emitting diodes (LEDs) that shine only for some instances according to the communication, which makes it possible for the camera shutter and the blinking of the LEDs to synchronize. This synchronization increases the system tolerance concerning changes in brightness in the ambient lights over time, independently of the landmarks location. Therefore, the environment's ceiling is populated with several landmarks and an Extended Kalman Filter is used to combine the dead-reckoning and landmark information. This increases the flexibility of the system by reducing the number of landmarks required. The experimental evaluation was conducted in a real indoor environment with an autonomous wheelchair prototype.