A three-dimensional representation method for noisy point clouds based on growing self-organizing maps accelerated on GPUs

The research described in this thesis was motivated by the need of a robust model capable of representing 3D data obtained with 3D sensors, which are inherently noisy. In addition, time constraints have to be considered as these sensors are capable of providing a 3D data stream in real time. This thesis proposed the use of Self-Organizing Maps (SOMs) as a 3D representation model. In particular, we proposed the use of the Growing Neural Gas (GNG) network, which has been successfully used for clustering, pattern recognition and topology representation of multi-dimensional data. Until now, Self-Organizing Maps have been primarily computed offline and their application in 3D data has mainly focused on free noise models, without considering time constraints. It is proposed a hardware implementation leveraging the computing power of modern GPUs, which takes advantage of a new paradigm coined as General-Purpose Computing on Graphics Processing Units (GPGPU). The proposed methods were applied to different problem and applications in the area of computer vision such as the recognition and localization of objects, visual surveillance or 3D reconstruction.

Nuevas herramientas para el estudio de la interacción entre especies en el espacio y en el tiempo

Las funciones de segundo orden son cada vez más empleadas en el análisis de procesos ecológicos. En este trabajo presentamos dos funciones de 2º orden desarrolladas recientemente que permiten analizar la interacción espacio-temporal entre dos especies o tipos funcionales de individuos. Estas funciones han sido desarrolladas para el estudio de interacciones entre especies en masas forestales a partir de la actual distribución diamétrica de los árboles. La primera de ellas es la función bivariante para procesos de puntos con marca Krsmm, que permite analizar la correlación espacial de una variable entre los individuos pertenecientes a dos especies en función de la distancia. La segunda es la función de reemplazo , que permite analizar la asociación entre los individuos pertenecientes a dos especies en función de la diferencia entre sus diámetros u otra variable asociada a dichos individuos. Para mostrar el comportamiento de ambas funciones en el análisis de sistemas forestales en los que operan diferentes procesos ecológicos se presentan tres casos de estudio: una masa mixta de Pinus pinea L. y Pinus pinaster Ait. en la Meseta Norte, un bosque de niebla de la Región Tropical Andina y el ecotono entre las masas de Quercus pyrenaica Willd. y Pinus sylvestris L. en el Sistema Central, en los que tanto la función Krsmm como la función r se utilizan para analizar la dinámica forestal a partir de parcelas experimentales con todos los árboles localizados y de parcelas de inventario.

Energy consumption in an urban wastewater treatment plant: the case of Murcia Region (Spain)

Of all the costs associated with the operation and maintenance of wastewater treatment plants (WWTPs), those associated with energy use tend to be the most significant. From this point of view, it is hence logical that energy efficiency and saving strategies should be one of the current focuses of debate amongst those involved with the management of WWTPs. The present study's objective is to determine the correlation between size and energy consumption for a WWTP. To this end, 90 WWTPs currently in service were analysed and their energetic impact quantified in terms of kWh/m3 of water treated. The results obtained demonstrate that energy consumption ratio increases as the size of WWTPs decreases, either in terms of treatment volume or population equivalent served.

Detection and mapping of burnt areas from time series of MODIS-derived NDVI data in a Mediterranean region

Moderate resolution remote sensing data, as provided by MODIS, can be used to detect and map active or past wildfires from daily records of suitable combinations of reflectance bands. The objective of the present work was to develop and test simple algorithms and variations for automatic or semiautomatic detection of burnt areas from time series data of MODIS biweekly vegetation indices for a Mediterranean region. MODIS-derived NDVI 250m time series data for the Valencia region, East Spain, were subjected to a two-step process for the detection of candidate burnt areas, and the results compared with available fire event records from the Valencia Regional Government. For each pixel and date in the data series, a model was fitted to both the previous and posterior time series data. Combining drops between two consecutive points and 1-year average drops, we used discrepancies or jumps between the pre and post models to identify seed pixels, and then delimitated fire scars for each potential wildfire using an extension algorithm from the seed pixels. The resulting maps of the detected burnt areas showed a very good agreement with the perimeters registered in the database of fire records used as reference. Overall accuracies and indices of agreement were very high, and omission and commission errors were similar or lower than in previous studies that used automatic or semiautomatic fire scar detection based on remote sensing. This supports the effectiveness of the method for detecting and mapping burnt areas in the Mediterranean region.

Clinical utility of ocular residual astigmatism and topographic disparity vector indexes in subclinical and clinical keratoconus

Purpose. We aimed to characterize the distribution of the vector parameters ocular residual astigmatism (ORA) and topography disparity (TD) in a sample of clinical and subclinical keratoconus eyes, and to evaluate their diagnostic value to discriminate between these conditions and healthy corneas. Methods. This study comprised a total of 43 keratoconic eyes (27 patients, 17–73 years) (keratoconus group), 11 subclinical keratoconus eyes (eight patients, 11–54 years) (subclinical keratoconus group) and 101 healthy eyes (101 patients, 15–64 years) (control group). In all cases, a complete corneal analysis was performed using a Scheimpflug photography-based topography system. Anterior corneal topographic data was imported from it to the iASSORT software (ASSORT Pty. Ltd), which allowed the calculation of ORA and TD. Results. Mean magnitude of the ORA was 3.23 ± 2.38, 1.16 ± 0.50 and 0.79 ± 0.43 D in the keratoconus, subclinical keratoconus and control groups, respectively (p < 0.001). Mean magnitude of the TD was 9.04 ± 8.08, 2.69 ± 2.42 and 0.89 ± 0.50 D in the keratoconus, subclinical keratoconus and control groups, respectively (p < 0.001). Good diagnostic performance of ORA (cutoff point: 1.21 D, sensitivity 83.7 %, specificity 87.1 %) and TD (cutoff point: 1.64 D, sensitivity 93.3 %, specificity 92.1 %) was found for the detection of keratoconus. The diagnostic ability of these parameters for the detection of subclinical keratoconus was more limited (ORA: cutoff 1.17 D, sensitivity 60.0 %, specificity 84.2 %; TD: cutoff 1.29 D, sensitivity 80.0 %, specificity 80.2 %). Conclusion. The vector parameters ORA and TD are able to discriminate with good levels of precision between keratoconus and healthy corneas. For the detection of subclinical keratoconus, only TD seems to be valid.

Fine-scale estimations of bioclimatic change in the Valencia region, Spain

Recent advances in statistical downscaling have allowed the reconstruction of temperatures for the complete 1948–2011 period in a spatial resolution of 90 m and without gaps for the Valencian Community (Spain) and bordering areas. It presently enables analyses in this region, which allows the determination of recent temperature changes at subregional and local scales. The present work focuses on obtaining the thermicity index according to Rivas-Martínez, a well-known indicator of different thermotypes associated with bioclimatic horizons. The change in this index, which has happened in the region between 1948 and 2011, was calculated by generating fine-scale maps of the potential extension of different thermotypes. The results show a greater regression for the thermotypes in a finicolous position, e.g. Orotemperate, Supratemperate and Supramediterranean horizons, which herein indicate greater potential vulnerability in climate change. In the absence of, and given the need for, such fine-scale information, this work should be useful for specialized researchers to spatially limit the potentially most vulnerable biotopes to climate change.

Características de ecossistema e sustentabilidade: estudo sobre a vegetação original e remanescente da bacia hidrográfica do Rio Canoas

Com o objetivo de comparar em termos de paisagem, a vegetação florestal nativa com os fragmentos da vegetação remanescente na bacia hidrográfica do Rio Canoas, situada na divisa dos Estados de São Paulo (à Nordeste) e Minas Gerais (à Centro-Oeste), busca-se dar subsídios aos atores locais para trabalhar políticas públicas com vistas à sustentabilidade do uso dos recursos naturais na região. Foram mapeados sete tipos de vegetação, com destaque para o cerrado e a floresta estacional semidecidual. O método do estudo é exploratório, realizado em fontes secundárias, pesquisa bibliográfica e documental, baseado no estudo de mapas e suas simulações, dados do IBGE, Embrapa, Ministério do Meio Ambiente e trabalhos de autores, como: Cogo (1987), Coutinho (2000) e Bozini (2005). Observou-se que a bacia hidrográfica objeto em estudo é degradada historicamente pelos diversos usos e ocupações do solo e sua maior parte ainda está sujeita à um elevado nível de perturbação, seja pela pecuária, ou mesmo pelas plantações de laranja, milho, eucalipto, cana-de-açúcar e, principalmente, pela cultura do café.

Model Probability in Self-organising Maps

Growing models have been widely used for clustering or topology learning. Traditionally these models work on stationary environments, grow incrementally and adapt their nodes to a given distribution based on global parameters. In this paper, we present an enhanced unsupervised self-organising network for the modelling of visual objects. We first develop a framework for building non-rigid shapes using the growth mechanism of the self-organising maps, and then we define an optimal number of nodes without overfitting or underfitting the network based on the knowledge obtained from information-theoretic considerations. We present experimental results for hands and we quantitatively evaluate the matching capabilities of the proposed method with the topographic product.