The characterization of modern jet engines or stationary gas turbines running with lean combustion by means of swirl-stabilized ames necessitates seedingless optical field measurements of the ame transfer function, i.e. the ratio of the uctuating heat release rate inside the ame volume, the instationary ow velocity at the combustor outlet and the time average of both quantities. For this reason, a high-speed camera-based laser interferometric vibrometer is proposed for spatio-temporally resolved measurements of the ame transfer function inside a swirl-stabilized technically premixed ame. Each pixel provides line-of-sight measurements of the heat release rate due to the linear coupling to uctuations of the refractive index along the laser beam, which are based on density uctuations inside the ame volume. Additionally, field measurements of the instationary ow velocity are possible due to correlation of simultaneously measured pixel signals and the known distance between the measurement positions. Thus, the new system enables the spatially resolved detection of the ame transfer function and instationary ow behavior with a single measurement for the first time. The presented setup offers single pixel resolution with measurement rates up to 40 kHz at an maximum image resolution of 256 px x 128 px. Based on a comparison with reference measurements using a standard pointwise laser interferometric vibrometer, the new system is validated and a discussion of the measurement uncertainty is presented. Finally, the measurement of refractive index uctuations inside a ame volume is demonstrated.

Soil is one of the most precious resources on Earth because of its role in storing and recycling water and nutrients essential for life, providing a variety of ecosystem services. This vulnerable resource is at risk from degradation by erosion, salinity, contamination and other effects of mismanagement. Information from soil is therefore crucial for its sustainable management. While the demand for soil information is growing, the quantity of data collected in the field is reducing due to financial constraints. Digital Soil Mapping (DSM) supports the creation of geographically referenced soil databases generated by using field observations or legacy data coupled, through quantitative relationships, with environmental covariates. This enables the creation of soil maps at unexplored locations at reduced costs. The selection of an optimal scale for environmental covariates is still an unsolved issue affecting the accuracy of DSM. The overall aim of this research was to explore the effect of spatial scale alterations of environmental covariates in DSM. Three main targets were identified: assessing the impact of spatial scale alterations on classifying soil taxonomic units; investigating existing approaches from related scientific fields for the detection of scale patterns and finally enabling practitioners to find a suitable scale for environmental covariates by developing a new methodology for spatial scale analysis in DSM. Three study areas, covered by detailed reconnaissance soil survey, were identified in the Republic of Ireland. Their different pedological and geomorphological characteristics allowed to test scale behaviours across the spectrum of conditions present in the Irish landscape. The investigation started by examining the effects of scale alteration of the finest resolution environmental covariate, the Digital Elevation Model (DEM), on the classification of soil taxonomic units. Empirical approaches from related scientific fields were subsequently selected from the literature, applied to the study areas and compared with the experimental methodology. Wavelet analysis was also employed to decompose the DEMs into a series of independent components at varying scales and then used in DSM analysis of soil taxonomic units. Finally, a new multiscale methodology was developed and evaluated against the previously presented experimental results. The results obtained by the experimental methodology have proved the significant role of scale alterations in the classification accuracy of soil taxonomic units, challenging the common practice of using the finest available resolution of DEM in DSM analysis. The set of eight empirical approaches selected in the literature have been proved to have a detrimental effect on the selection of an optimal DEM scale for DSM applications. Wavelet analysis was shown effective in removing DEM sources of variation, increasing DSM model performance by spatially decomposing the DEM. Finally, my main contribution to knowledge has been developing a new multiscale methodology for DSM applications by combining a DEM segmentation technique performed by k-means clustering of local variograms parameters calculated in a moving window with an experimental methodology altering DEM scales. The newly developed multiscale methodology offers a way to significantly improve classification accuracy of soil taxonomic units in DSM. In conclusion, this research has shown that spatial scale analysis of environmental covariates significantly enhances the practice of DSM, improving overall classification accuracy of soil taxonomic units. The newly developed multiscale methodology can be successfully integrated in current DSM analysis of soil taxonomic units performed with data mining techniques, so advancing the practice of soil mapping. The future of DSM, as it successfully progresses from the early pioneering years into an established discipline, will have to include scale and in particular multiscale investigations in its methodology. DSM will have to move from a methodology of spatial data with scale to a spatial scale methodology. It is now time to consider scale as a key soil and modelling attribute in DSM.

This master thesis presents the process of designing and implementing a CNN-based architecture for image recognition included in a larger project in the field of fashion recommendation with deep learning. Concretely, the presented network aims to perform localization and segmentation tasks. Therefore, an accurate analysis of the most well-known localization and segmentation networks in the state of the art has been performed. Afterwards, a multi-task network performing RoI pixel-wise segmentation has been created. This proposal solves the detected weaknesses of the pre-existing networks in the field of application, i.e. fashion recommendation. These weaknesses are basically related with the lack of a fine-grained quality of the segmentation and problems with computational efficiency. When it comes to improve the details of the segmentation, this network proposes to work pixel- wise, i.e. performing a classification task for each of the pixels of the image. Thus, the network is more suitable to detect all the details presented in the analysed images. However, a pixel-wise task requires working in pixel resolution, which implies that the number of operations to perform is usually large. To reduce the total number of operations to perform in the network and increase the computational efficiency, this pixel-wise segmentation is only done in the meaningful regions of the image (Regions of Interest), which are also computed in the network (RoI masks). Then, after a study of the more recent deep learning libraries, the network has been successfully implemented. Finally, to prove the correct operation of the design, a set of experiments have been satisfactorily conducted. In this sense, it must be noted that the evaluation of the results obtained during testing phase with respect to the most well-known architectures is out of the scope of this thesis as the experimental conditions, especially in terms of dataset, have not been suitable for doing so. Nevertheless, the proposed network is totally prepared to perform this evaluation in the future, when the required experimental conditions are available. / Denna examensarbete presenterar processen för att designa och implementera en CNN-baserad arkitektur för bildigenkänning som ingår i ett större projekt inom moderekommendation med djup inlärning. Konkret, det presenterade nätverket syftar till att utföra lokaliseringsoch segmenteringsuppgifter. Därför har en noggrann analys av de mest kända lokaliseringsoch segmenteringsnätena utförts inom den senaste tekniken. Därefter har ett multi-task-nätverk som utför RoI pixel-wise segmentering skapats. Detta förslag löser de upptäckta svagheterna hos de befintliga näten inom tillämpningsområdet, dvs modeanbefaling. Dessa svagheter är i grund och botten relaterade till bristen på en finkornad kvalitet på segmenteringen och problem med beräkningseffektivitet. När det gäller att förbättra detaljerna i segmenteringen, föreslår detta nätverk att arbeta pixelvis, dvs att utföra en klassificeringsuppgift för var och en av bildpunkterna i bilden. Nätverket är sålunda lämpligare att detektera alla detaljer som presenteras i de analyserade bilderna. En pixelvis uppgift kräver dock att man arbetar med pixelupplösning, vilket innebär att antalet operationer som ska utföras är vanligtvis stor. För att minska det totala antalet operationer som ska utföras i nätverket och öka beräkningseffektiviteten görs denna pixelvisa segmentering endast i de meningsfulla regionerna i bilden (intressanta regioner), som också beräknas i nätverket (RoI-masker) . Sedan, efter en studie av de senaste djuplärningsbiblioteken, har nätverket framgångsrikt implementerats. Slutligen, för att bevisa korrekt funktion av konstruktionen, har en uppsättning experiment genomförts på ett tillfredsställande sätt. I detta avseende måste det noteras att utvärderingen av de resultat som uppnåtts under testfasen i förhållande till de mest kända arkitekturerna ligger utanför denna avhandling, eftersom de experimentella förhållandena, särskilt vad gäller dataset, inte har varit lämpliga För att göra det. Ändå är det föreslagna nätverket helt beredd att utföra denna utvärdering i framtiden när de nödvändiga försöksvillkoren är tillgängliga. / En aquest treball de fi de màster es presenta el disseny i la implementació d’una arquitectura pel reconeixement d’imatges fent ús de CNN. Aquesta xarxa es troba inclosa en un projecte de major envergadura en el camp de la recomanació de moda. En concret, la xarxa presentada en aquest document s’encarrega de realitzar les tasques de localització i segmentació. Després d’un estudi a consciència de les xarxes més conegudes de l’estat de l’art, s’ha dissenyat una xarxa multi-tasca encarregada de realitzar una segmentació a resolució de píxel de les regions d’interès de la imatge, les quals han sigut prèviament calculades i emmascarades. Aquesta proposta soluciona les mancances detectades en les xarxes ja existents pel que fa a la tasca de recomanació de moda. Aquestes mancances es basen en la obtenció d’una segmentació sense prou nivell de detalls i en una rellevant complexitat computacional. Pel que fa a la qualitat de la segmentació, aquesta tesi proposa treballar en resolució de píxel, classificant tots els píxels de la imatge de forma individual, per tal de poder adaptar-se a tots els detalls que puguin aparèixer a la imatge analitzada. No obstant, treballar píxel a píxel implica la realització d’una gran quantitat d’operacions. Per reduir-les, proposem fer la segmentació píxel a píxel només a les regions d’interès de la imatge. A continuació, després d’un estudi detallat de les llibreries de deep learnign més destacades, el disseny ha sigut implementat. Finalment s’han dut a terme una sèrie d’experiments per provar el correcte funcionament del disseny. En aquest sentit és important destacar que aquesta tesi no té com a objectiu avaluar el disseny respecte d’altres xarxes ja existents. La raó és que les condicions d’experimentació, sobretot pel que fa a la base de dades, no són adequades per aquesta tasca. No obstant, la xarxa està perfectament preparada per fer aquesta avaluació un cop les condicions d’experimentació així ho permetin.