Search

Large-area land cover maps are produced to satisfy different information needs. Land cover maps having partial or complete spatial and/or temporal overlap, different legends, and varying accuracies for similar classes, are increasingly common. To address these concerns and combine two 30-m resolution land cover products, we implemented a harmonization procedure using a Latent Dirichlet Allocation (LDA) model. The LDA model used regionalized class co-occurrences from multiple maps to generate a harmonized class label for each pixel by statistically characterizing land attributes from the class co-occurrences. We evaluated multiple harmonization approaches: using the LDA model alone and in combination with more commonly used information sources for harmonization (i.e. error matrices and semantic affinity scores). The results were compared with the benchmark maps generated using simple legend crosswalks and showed that using LDA outputs with error matrices performed better and increased harmonized map overall accuracy by 6–19% for areas of disagreement between the source maps. Our results revealed the importance of error matrices to harmonization, since excluding error matrices reduced overall accuracy by 4–20%. The LDA-based harmonization approach demonstrated in this paper is quantitative, transparent, portable, and efficient at leveraging the strengths of multiple land cover maps over large areas.

Licencia Creative Commons: Attribution 3.0 Unported (CC BY 3.0), This paper assesses the combined effect of field plot size and LiDAR density on the estimation of four forest structure attributes: volume, total biomass, basal area and canopy cover. A total of 21 different plot sizes were considered, obtained by decreasing the field measured plot radius value from 25 to 5 m with regular intervals of 1 m. LiDAR data densities were simulated by randomly removing LiDAR pulses until reaching nine different density values. In order to avoid influence of the digital terrain model spatial resolution, eight different resolutions were considered (from 0.25 to 2 m grid size) and tested. A set of per-plot LiDAR metrics was extracted for each parameter combination. Prediction models of forest attributes were defined using forward stepwise ordinary least-square regressions. Results show that the highest R 2 values are reached by combining large plot sizes and high LiDAR data density values. However, plot size has a greater effect than LiDAR point density. In general, minimum plot areas of 500–600 m2 are needed for volume, biomass and basal area estimates, and of 300–400 m2 for canopy cover. Larger plot sizes do not significantly increase the accuracy of the models, but they increase the cost of fieldwork.

This is an author's accepted manuscript of an article published in “Remote Sensing Letters", Volume 5, Issue 4, 2014; copyright Taylor & Francis; available online at: http://www.tandfonline.com/doi/abs/10.1080/2150704X.2014.903350, [EN] When processing scanning LiDAR data, it is commonly assumed that the extracted full-waveform LiDAR pulse registers truly vertical information of forest canopies. This assumption may lead to uncertain results for the spatiotemporal analysis of the waveforms due to off-nadir scanning angles and varying trajectories travelled by the pulses in overlapping strips. In this letter, we investigate these assumptions and undertake some preliminary analysis to overcome their impacts on forest-based LiDAR analyses. Our results demonstrate that for a standard LiDAR forest acquisition programme in Oregon, USA, most of the hits (83%) are acquired off-nadir, which leads to positional displacements on the ground of the full-waveforms of about 0.20 m for each 1-m height increment. We propose an approach to synthetize multiple waveform data into composite waveforms containing the vertical profile of vegetation for a given location. This approach is based on partitioning the aboveground vertical space into voxels and using the maximum full-waveform intensity value to construct new full-waveforms comprising the vertical information of the various waveforms crossing over a location. Our initial results indicate that deriving spatiotemporal metrics from the composite pseudo-vertical full-waveforms produces a more consistent response across adjacent height levels, which in turn enables a more complete characterization and more vegetation structure to be retrieved. We conclude that this type of pseudo-vertical full-waveform analysis is necessary to more fully understand the impact of the return signals from tree crowns

[EN] Urban spatial structures reflect local particularities produced during the development of a city. High spatial resolution imagery and LiDAR data are currently used to derive numerical attributes to describe in detail intra-urban structures and morphologies. Urban block boundaries have been frequently used to define the units for extracting metrics from remotely sensed data. In this paper, we propose to complement these metrics with a set of novel descriptors of the streets surrounding the urban blocks under consideration. These metrics numerically describe geometrical properties in addition to other distinctive aspects, such as presence and properties of vegetation and the relationship between the streets and buildings. For this purpose, we also introduce a methodology for partitioning the street area related to an urban block into polygons from which the street urban metrics are derived. We achieve the assessment of these metrics through application of a one-way ANOVA procedure, the winnowing technique, and a decision tree classifier. Our results suggest that street metrics, and particularly those describing the street geometry, are suitable for enhancing the discrimination of complex urban typologies and help to reduce the confusion between certain typologies. The overall classification accuracy increased from 72.7% to 81.1% after the addition street of descriptors. The results of this study demonstrate the usefulness of these metrics for describing street properties and complementing information derived from urban blocks to improve the description of urban areas. Street metrics are of particular use for the characterization of urban typologies and to study the dynamics of cities.

A plot-based approach is proposed to detect fruit trees from high spatial resolution aerial images and extract tree and plot-based parameters, such as fraction of tree cover, number of trees, and planting patterns. Each plot image, defined by the boundaries or polygons obtained from a cadastral database, is analyzed independently. The methodology is based on image processing methods: an unsupervised classification with the k-means algorithm is applied, followed by the automatic identification of the classes representing trees. Once extracted, each tree is individualized using a morphological process applied on the binary image of the trees. A set of parameters is calculated at tree and plot levels that produces a comprehensive description of the spectral and morphological aspects of the trees, as well as their spatial distribution in each plot. The methodology was tested on 0.5 m/pixel spatial resolution aerial images of 300 citrus orchard plots which included the three citrus tree typologies found in the Valencia region (Spain). The accuracy of the fruit tree extraction and the parameters calculated was evaluated by comparison with reference data obtained by manual delineation of the images. The automatically extracted fraction of tree cover was significantly related to the reference tree cover area (R2 = 0.96). In the case of the number of detected trees, the R2 values were always higher than 0.90 for the three typologies. Tree location was estimated with an average positional error of 40 cm. The error obtained in the characterization of the planting pattern was less than 50 cm. The proposed methodology may be applied to large agricultural databases, and the derived information combined with precision agricultural techniques could improve the efficiency of various irrigation and agricultural management tasks such as handling per-plot water requirements and distribution.

he benchmark problem proposed in this paper is to identify regions in aerial or satellite images with geometric patterns and describe the geometric properties of the constituent elements of the pattern and their spatial distribution. This is a relevant topic in image analysis processes where spatial regular patterns are studied. This paper first presents two approaches based on multi-directional semivariograms for reducing the processing time required to compute omnidirectional semivariograms. A set of parameters for describing the structure of a semivariogram, introduced by Balaguer et al. (2010), is extracted from an experimental semivariogram and analysed to quantify the heterogeneity of the distribution of elements (trees) with periodic patterns in images of orchards. An assessment is made using four image datasets. The first dataset is composed of synthetic images that simulate regularly spaced tree crops and real images, and is used to evaluate the influence that the orientation of elements (regularly spaced trees) in the objects (crop plots) has in the descriptive parameter values. This dataset is also used to compare different semivariogram computational approaches. The other three are also composed of synthetic images and are used to evaluate the semivariogram parameters under different spatial heterogeneity conditions, and are generated by varying patterns and tree characteristics, i.e., existence or absence of faults, regular/irregular distributions, and size of the elements. Finally, the proposed methodology is applied to real aerial orthoimages of orchard plots.

[ES] Se propone una metodología para automatizar la detección y la localización de árboles en zonas forestales a través de datos LIDAR y ortofotos. Para la automatización se han empleado diversos algoritmos para las diferentes fases como son el procesado de datos LiDAR y la detección y la localización de los árboles con watershed y máximos locales. Los resultados obtenidos son satisfactorios, consiguiendo tasas de aciertos en la detección y localización que pueden ser útiles para ciertas aplicaciones prácticas en inventarios forestales.

Several studies have addressed the biomass and volume of trees using Airborne Light Detection and Ranging (LiDAR) data. However, little research has been conducted into shrub vegetation, which covers a high percentage of Mediterranean forest. We used LiDAR data and an airborne image to estimate biomass and volume of shrub vegetation. Field data were collected in 29 square plots of 100 m2. In each plot, the percentage of the surface covered was measured in the field. Shrub vegetation inside 3 stands for each plot was clear cut to calculate the biomass and volume of the 29 plots. To find the best fit between LiDARspectral data and field measurements, stepwise regressions were performed using previously selected variables. The highest accuracy was found when variables derived from LiDAR data and the airborne image were combined (R2 values of 0.78 and 0.84 for biomass and volume, respectively). Biomass and volume were predicted using variables from height metrics of LiDAR data (median and standard deviation); density metrics (percentage of points whose height was between 0.50 m and 0.75 m, 0.75 me1 m, and higher than 1 m); and spectral data (standard deviation of green band, mean of the vegetation index NDVI). These results revealed the potential of LiDAR and spectral data to characterize shrub structure and make it possible to estimate and map the biomass and volume of this vegetation.

A methodology for mapping urban land-use types integrating information from multiple data sources (high spatial resolution imagery, LiDAR data, and cadastral plots) is presented. A large set of complementary descriptive features that allow distinguishing different urban structures (historical, urban, residential, and industrial) is extracted and, after a selection process, a plot-based image classification approach applied, facilitating to directly relate the classification results and the urban descriptive parameters computed to the existent land-use/land-cover units in geospatial databases. The descriptive features are extracted by considering different hierarchical scale levels with semantic meaning in urban environments: buildings, plots, and urban blocks. Plots are characterised by means of image-based (spectral and textural), three-dimensional, and geometrical features. In addition, two groups of contextual features are defined: internal and external. Internal contextual features describe the main land cover types inside the plot (buildings and vegetation). External contextual features describe each object in terms of the properties of the urban block to which it belongs. After the evaluation in an heterogeneous Mediterranean urban area, the land-use classification accuracy values obtained show that the complementary descriptive features proposed improve the characterisation of urban typologies. A progressive introduction of the different groups of descriptive features in the classification tests show how the subsequent addition of internal and external contextual features have a positive effect by increasing the final accuracy of the urban classes considered in this study. © 2012 Elsevier B.V.

[EN] The abandonment of agricultural plots entails a low economic productivity of the land and a higher vulnerability to wildfires and degradation of affected areas. In this sense, the local government of Galicia is promoting new methodologies based on high-resolution images in order to classify the territory in basic and generic land uses. This procedure will be used to control the sustainable management of plots belonging to the Land Bank. This paper presents an application study for maintaining and updating land use/land cover geospatial databases using parcel-oriented classification. The test is performed over two geographic areas of Galicia, in the northwest of Spain. In this region, forest and shrublands in mountain environments are very heterogeneous with many private unproductive plots, some of which are in a high state of abandonment. The dataset is made of high spatial resolution multispectral imagery, cadastral cartography employed to define the image objects (plots), and field samples used to define evaluation and training samples. A set of descriptive features is computed quantifying different properties of the objects, i.e. spectral, texture, structural, and geometrical. Additionally, the effect on the classification and updating processes of the historical land use as a descriptive feature is tested. Three different classification methodologies are analyzed: linear discriminant analysis, decision trees, and support vector machine. The overall accuracies of the classifications obtained are always above 90 % and support vector machine method is proved to provide the best performance. Forest and shrublands areas are especially undefined, so the discrimination between these two classes is low. The results enable to conclude that the use of automatic parcel-oriented classification techniques for updating tasks of land use/land cover geospatial databases, is effective in the areas tested, particularly when broad and well defined classes are required.

Copyright 2011 - American Society of Agricultural and Biological Engineers ISSN 2151-0032, [EN] There is need to identify and quantify the contribution of different sources to airborne particulate matter (PM) emissions from animal houses. To this end, we compared the chemical and morphological characteristics of fine and coarse PM from known sources collected from animal houses with the characteristics of on‐farm fine and coarse airborne PM using two methods: classification rules based on decision trees and multiple linear regression. Fourteen different farms corresponding to seven different housing systems for poultry and pigs were sampled during winter. A total of 28 fine and 28coarse on‐farm airborne PM samples were collected, together with a representative sample of each known source per farm (56 known source samples in total). Source contributions were calculated as relative percentage contributions in particle numbers and then estimated in particle mass. Based on particle numbers, results showed that in poultry houses, most on‐farm airborne PM originated from feathers (ranging from 4% to 43% in fine PM and from 6% to 35% in coarse PM) and manure (ranging from 9% to 85% in fine PM and from 30% to 94% in coarse PM). For pigs, most on‐farm airborne PM originated from manure (ranging from 70% to 98% in fine PM and from 41% to 94% in coarse PM). Based on particle mass, for poultry most on‐farm airborne PM still originated from feathers and manure; for pigs, however, most PM originated from skin and manure. Feed had a negligible contribution to on‐farm airborne PM compared with other sources. Results presented in this study improve the understanding of sources of PM in different animal housing systems, which may be valuable when choosing optimal PM reduction techniques.

[EN] This paper analyzes the effect of the addition of historical land-use as a descriptive feature in plot-based image classification when updating land-use/land-cover geospatial databases. Several historical databases have been simulated to assess the influence and significance of this feature in the classification. The causes, nature, and evolution of classification errors as the database currency varies are analyzed; and the impact of these errors on change detection during the updating process is evaluated. The results show that the addition of historical land-use information increases the overall accuracy of image classifications. During a database updating process, changes are detected by comparing the historical land-use with the classification results. The main drawback of employing historical land-use as a descriptive feature in image classification for change detection is that the percentage of undetectable errors significantly increases as more accurate is the database information.

A software application for automatic descriptive feature extraction from image-objects, FETEX 2.0, is presented and described in this paper. The input data include a multispectral high resolution digital image and a vector file in shapefile format containing the polygons or objects, usually extracted from a geospatial database. The design of the available descriptive features or attributes has been mainly focused on the description of agricultural parcels, providing a variety of information: spectral information from the different image bands; textural descriptors of the distribution of the intensity values based on the grey level co-occurrence matrix, the wavelet transform and a factor of edgeness; structural features describing the spatial arrangement of the elements inside the objects, based on the semivariogram curve and the Hough transform; and several descriptors of the object shape. The output file is a table that can be produced in four alternative formats, containing a vector of features for every object processed. This table of numeric values describing the objects from different points of view can be externally used as input data for any classification software. Additionally, several types of graphs and images describing the feature extraction procedure are produced, useful for interpretation and understanding the process. A test of the processing times is included, as well as an application of the program in a real parcel-based classification problem, providing some results and analyzing the applicability, the future improvement of the methodologies, and the use of additional types of data sets. This software is intended to be a dynamic tool, integrating further data and feature extraction algorithms for the progressive improvement of land use/land cover database classification and agricultural database updating processes. © 2011 Elsevier B.V.

In this paper, two main approaches for automatic building detection and localization using high spatial resolution imagery and LiDAR data are compared and evaluated: thresholding-based and object-based classification. The thresholding-based approach is founded on the establishment of two threshold values: one refers to the minimum height to be considered as building, defined using the LiDAR data, and the other refers to the presence of vegetation, which is defined according to the spectral response. The other approach follows the standard scheme of object-based image classification: segmentation, feature extraction and selection, and classification, here performed using decision trees. In addition, the effect of the inclusion in the building detection process of contextual relations with the shadows is evaluated. Quality assessment is performed at two different levels: area and object. Area-level evaluates the building delineation performance, whereas object-level assesses the accuracy in the spatial location of individual buildings. The results obtained show a high efficiency of the evaluated methods for building detection techniques, in particular the thresholding-based approach, when the parameters are properly adjusted and adapted to the type of urban landscape considered. © 2011 by the authors.

The knowledge on the contribution of individual sources to particulate matter (PM) in different size fractions is essential to improve PM reduction from livestock houses. We investigated which input data (particle chemical, morphological or combined characteristics) were best to distinguish amongst specific potential sources of PM. We used a cross-validation procedure with classification rules based on decision trees and analyzed misclassification errors. The PM from two livestock species (poultry and pigs), and in two different fractions (fine and coarse) was studied. A total of 618 particles were analyzed from poultry houses for fine, and 805 for coarse PM. A total of 317 particles were analyzed from pig houses for fine, and 337 for coarse PM. Results showed the selection of the best input data varies depending on the sources, which depend on livestock species.

[EN] In this paper we propose and evaluate different mathematical parameters extracted from the experimental semivariogram for land use/land cover classification using high-resolution images and cadastral mapping limits for the definition of the objects of analysis. First, we describe the process of calculating the semivariogram from the gray level values in an image object. In order to optimize the computation time we present two pixel selection techniques that preserve the original shape of the semivariogram. Several parameters are then extracted from the semivariogram. Finally, we use various statistical techniques to select the most discriminant parameters. Last section shows the results obtained using aerial digital images of an agricultural area on the Mediterranean coast of Spain. The study of the practical application presented in this paper facilitates the understanding of the relationship between the behaviour of the experimental semivariogram and the variability of the intensity values in a digital image. In order to follow the development of this work, the reader should know some basis of classification methods and digital image processing techniques., [ES] En este trabajo se proponen y evalúan diferentes parámetros matemáticos extraídos del semivariograma experimental para la clasificación de los usos del suelo mediante imágenes de alta resolución, usando los límites catastrales para la definición de los objetos de análisis. En primer lugar, se describe el proceso de cálculo del semivariograma a partir de los valores de niveles de gris del objeto imagen. Con el fin de optimizar el tiempo de cálculo se presentan dos técnicas de selección de píxeles que conservan la forma original del semivariograma. A continuación se definen varios parámetros del semivariograma. Final- mente, se usan diferentes técnicas estadísticas para la selección de los parámetros más discriminantes. La última sección muestra los resultados obtenidos con las imágenes digitales aéreas de una zona agrícola en la costa mediterránea de España. El estudio de la aplicación práctica que se presenta facilita la comprensión de la relación entre el comportamiento del semivariograma experimental y la variabilidad de los valores de intensidad en una imagen digital. Con el fin de seguir el desarrollo de este trabajo, el lector debe conocer algunos métodos estadísticos de clasificación y algunas técnicas de procesamiento digital de imágenes.

There is need to identify and quantify the contribution of different sources to airborne particulate matter (PM) emissions from animal houses. To this end, we compared the chemical and morphological characteristics of fine and coarse PM from known sources collected from animal houses with the characteristics of on-farm fine and coarse airborne PM using two methods: classification rules based on decision trees and multiple linear regression. Fourteen different farms corresponding to seven different housing systems for poultry and pigs were sampled during winter. A total of 28 fine and 28 coarse on-farm airborne PM samples were collected, together with a representative sample of each known source per farm (56 known source samples in total). Source contributions were calculated as relative percentage contributions in particle numbers and then estimated in particle mass. Based on particle numbers, results showed that in poultry houses, most on-farm airborne PM originated from feathers (ranging from 4% to 43% in fine PM and from 6% to 35% in coarse PM) and manure (ranging from 9% to 85% in fine PM and from 30% to 94% in coarse PM). For pigs, most on-farm airborne PM originated from manure (ranging from 70% to 98% in fine PM and from 41% to 94% in coarse PM). Based on particle mass, for poultry most on-farm airborne PM still originated from feathers and manure; for pigs, however, most PM originated from skin and manure. Feed had a negligible contribution to on-farm airborne PM compared with other sources. Results presented in this study improve the understanding of sources of PM in different animal housing systems, which may be valuable when choosing optimal PM reduction techniques

This is an Accepted Manuscript of an article published by Taylor & Francis in International Journal of Digital Earth on 2011, available online: http://wwww.tandfonline.com/10.1080/17538947.2010.533201, [EN] The creation of a quality Digital Terrain Model (DTM) is essential for representing and analyzing the Earth in a digital form. The continuous improvements in the acquisition and the potential of airborne Light Detection and Ranging (LiDAR) data are increasing the range of applications of this technique to the study of the Earth surface. The aim of this study was to determine the optimal parameters for calculating a DTM by using an iterative algorithm to select minimum elevations from LiDAR data in a steep mountain area with shrub vegetation. The parameters were: input data type, analysis window size, and height thresholds. The effects of slope, point density, and vegetation on DTM accuracy were also analyzed. The results showed that the lowest root mean square error (RMSE) was obtained with an analysis window size of 10 m, 5 m, and 2.5 m, rasterized data as input data, and height thresholds equal to or greater than 1.5 m. These parameters showed a RMSE of 0.19 m. When terrain slope varied from 0-10% to 50-60%, the RMSE increased by 0.11 m. The RMSE decreased by 0.06 m when point density was increased from 4 to 8 points/m2, and increased by 0.05 m in dense vegetation areas. © 2011 Taylor & Francis.

In this paper, a comprehensive set of texture features extracted from the experimental semivariogram of specific image objects is proposed and described, and their usefulness for land use classification of high resolution images is evaluated. Fourteen features are defined and categorized into three different groups, according to the location of their respective parameters in the semivariogram curve: (i) features that use parameters close to the origin of the semivariogram, (ii) the parameters employed extend to the first maximum, and (iii) the parameters employed are extracted from the first to the second maximum. A selection of the most relevant features has been performed, combining the analysis and interpretation of redundancies, and using statistical discriminant analysis methods. The suitability of the proposed features for object-based image classification has been evaluated using digital aerial images from an agricultural area on the Mediterranean coast of Spain. The performance of the selected semivariogram features has been compared with two different sets of texture features: those derived from the grey level co-occurrence matrix, and the values of raw semivariance directly extracted from the semivariogram at different positions. As a result of the tests, the classification accuracies obtained using the proposed semivariogram features are, in general, higher and more balanced than those obtained using the other two sets of standard texture features.

There is need to identify and quantify the contribution of different sources to airborne particulate matter (PM) emissions from animal houses. To this end, we compared the chemical and morphological characteristics of fine and coarse PM from known sources collected from animal houses with the characteristics of on-farm fine and coarse airborne PM using two methods: classification rules based on decision trees and multiple linear regression. Fourteen different farms corresponding to seven different housing systems for poultry and pigs were sampled during winter. A total of 28 fine and 28 coarse on-farm airborne PM samples were collected, together with a representative sample of each known source per farm (56 known source samples in total). Source contributions were calculated as relative percentage contributions in particle numbers and then estimated in particle mass. Based on particle numbers, results showed that in poultry houses, most on-farm airborne PM originated from feathers (ranging from 4% to 43% in fine PM and from 6% to 35% in coarse PM) and manure (ranging from 9% to 85% in fine PM and from 30% to 94% in coarse PM). For pigs, most on-farm airborne PM originated from manure (ranging from 70% to 98% in fine PM and from 41% to 94% in coarse PM). Based on particle mass, for poultry most on-farm airborne PM still originated from feathers and manure; for pigs, however, most PM originated from skin and manure. Feed had a negligible contribution to on-farm airborne PM compared with other sources. Results presented in this study improve the understanding of sources of PM in different animal housing systems, which may be valuable when choosing optimal PM reduction techniques

[EN] A fourth-order non-linear interpolation procedure based on the ENO (Essentially Non-Oscillatory) methodology is presented and evaluated, with the purpose of increasing the geometric accuracy of edge detection in digital images. Two possible cases are considered one that considers that each pixel of the image represents a point value, the other that the pixel is an average value of a function. After image interpolation to obtain a finer grid of pixels, the Canny edge detection algorithm is applied, with the objective of improving the localization and geometry of the edges at a subpixel level. The results are compared with other schemes based on fourth order two-dimensional interpolation methods, such as a centered scheme based on a cubic convolution, a fourth order non-centered lineal scheme and a centered cubic convolution based on local gradient features. The evaluation is performed using visual and analytical techniques applied over aerial and satellite images, analyzing the positional errors of the detected edges, as well as the errors due to changes in scale and orientation. In addition to the subpixel edge detection, the quality of the interpolated images is tested. We conclude that the proposed methodology based on ENO interpolation improves the detection of edges in images as compared to other fourth-order methods