Jorge García-Gutiérrez

A comparison of machine learning regression techniques for LiDAR-derived estimation of forest variables

Light Detection and Ranging (LiDAR) is a remote sensor able to extract three-dimensional information. Environmental models in forest areas have been benefited by the use of LiDAR-derived information in the last years. A multiple linear regression (MLR) with previous stepwise feature selection is the most common method in the literature to develop those models. MLR defines the relation between the set of field measurements and the statistics extracted from a LiDAR flight. Machine learning has emerged as a suitable tool to improve classic stepwise MLR results on LiDAR. Unfortunately, few studies have been proposed to compare the quality of the multiple machine learning approaches. This paper presents a comparison between the classic MLR-based methodology and regression techniques in machine learning (neural networks, support vector machines, nearest neighbour, ensembles such as random forests) with special emphasis on regression trees. The selected techniques are applied to real LiDAR data from two areas in the province of Lugo (Galizia, Spain). The results confirm that classic MLR is outperformed by machine learning techniques and concretely, our experiments suggest that Support Vector Regression with Gaussian kernels statistically outperforms the rest of the techniques.

Evolutionary feature selection to estimate forest stand variables using LiDAR

Abstract Light detection and ranging (LiDAR) has become an important tool in forestry. LiDAR-derived models are mostly developed by means of multiple linear regression (MLR) after stepwise selection of predictors. An increasing interest in machine learning and evolutionary computation has recently arisen to improve regression use in LiDAR data processing. Although evolutionary machine learning has already proven to be suitable for regression, evolutionary computation may also be applied to improve parametric models such as MLR. This paper provides a hybrid approach based on joint use of MLR and a novel genetic algorithm for the estimation of the main forest stand variables. We show a comparison between our genetic approach and other common methods of selecting predictors. The results obtained from several LiDAR datasets with different pulse densities in two areas of the Iberian Peninsula indicate that genetic algorithms perform better than the other methods statistically. Preliminary studies suggest that a lack of parametric conditions in field data and possible misuse of parametric tests may be the main reasons for the better performance of the genetic algorithm. This research confirms the findings of previous studies that outline the importance of evolutionary computation in the context of LiDAR analisys of forest data, especially when the size of fieldwork datatasets is reduced.

On the evolutionary optimization of k-NN by label-dependent feature weighting

Different approaches of feature weighting and k-value selection to improve the nearest neighbour technique can be found in the literature. In this work, we show an evolutionary approach called k-Label Dependent Evolutionary Distance Weighting (kLDEDW) which calculates a set of local weights depending on each class besides an optimal k value. Thus, we attempt to carry out two improvements simultaneously: we locally transform the feature space to improve the accuracy of the k-nearest-neighbour rule whilst we search for the best value for k from the training data. Rigorous statistical tests demonstrate that our approach improves the general k-nearest-neighbour rule and several approaches based on local weighting.

Comparison of ALS based models for estimating aboveground biomass in three types of Mediterranean forest

Abstract This study aimed to develop ALS-based models for estimating stem, crown and aboveground biomass in three types of Mediterranean forest, based on low density ALS data. Two different modelling approaches were used: (i) linear models with different variable selection methods (Stepwise Selection [SS], Clustering/Exhaustive search [CE] and Genetic Algorithm [GA]), and (ii) previously Published Models (PM) applicable to diverse types of forest. Results indicated more accurate estimations of biomass components for pure Pinus pinea L. (rRMSE = 25.90-26.16%) than for the mixed (30.86-36.34%) and Quercus pyrenaica Willd. forests (32.78-34.84%). All the tested approaches were valuable, but SS and GA performed better than CE and PM in most cases.

An evolutionary-weighted majority voting and support vector machines applied to contextual classification of LiDAR and imagery data fusion

Data classification is a critical step to convert remotely sensed data into thematic information. Environmental researchers have recently maximized the synergy between passive sensors and LiDAR (Light Detection and Ranging) for land cover classification by means of machine learning. Although object-based paradigm is frequently used to classify high resolution imagery, it often requires a high level of expertise and time effort. Contextual classification may lead to similar results with a decrease in time and costs for non-expert users. This work shows a novel contextual classifier based on a Support Vector Machine (SVM) and an Evolutionary Majority Voting (SVM-EMV) to develop thematic maps from LiDAR and imagery data. Subsequently, the performance of SVM-EMV is compared to that achieved by a pixel-based SVM as well as to a contextual classified based on SVM and MRF. The classifiers were tested over three different areas of Spain with well differentiated environmental characteristics. Results show that SVM-EMV statistically outperforms the rest (SVM, SVM-MRF) for the three datasets obtaining a 77%, 91% and 92% of global accuracy for Trabada, Huelva and Alto Tajo, respectively.

Automatic environmental quality assessment for mixed-land zones using lidar and intelligent techniques

Research highlights? LiDAR can provide excellent information to improve thematic maps in especially interesting areas. ? Intelligent techniques are a key factor to provide fast and accurate results when lidar is applied to the study of the natural environment. ? Our experimentation on real data from a riparian area in the south of Spain shows decision trees (C4.5) provide the best results with the highest level of clarity for the final model. Human impact on the natural environment is an evident global fact. Natural, industrial and touristic areas coexist in a more than delicate balance. In Andalusia, in the south of Spain, the Regional Ministry for the Environment is responsible for the control and preservation of natural resources. This task bears a high cost in time and money. Remote sensing and the use of intelligent techniques are excellent tools to reduce such costs. This work explores the joint use of the lidar sensor, which provides a great quantity of information describing three dimensional space, and the application of intelligent techniques for rapid and efficient land use and land cover classification with the objective of differentiating urban land from natural ground close to protected areas of Huelva province. For this, seven types of land use and land cover have been studied for a riparian area next to the mouth of the rivers Tinto and Odiel, extracting 33 distinct features from the lidar point cloud. Subsequently, a supervised learning algorithm is applied to construct a model which, with a resolution of 4m2, obtained relative precision between 71% and 100% and an average total precision of 85%.

Biclusters evaluation based on shifting and scaling patterns

Microarray techniques have motivated the develop of different methods to extract useful information from a biological point of view. Biclustering algorithms obtain a set of genes with the same behaviour over a group of experimental conditions from gene expression data. In order to evaluate the quality of a bicluster, it is useful to identify specific tendencies represented by patterns on data. These patterns describe the behaviour of a bicluster obtained previously by an adequate biclustering technique from gene expression data. In this paper a new measure for evaluating biclusters is proposed. This measure captures a special kind of patterns with scaling trends which represents quality patterns. They are not contemplated with the previous evaluating measure accepted in the literature. This work is a first step to investigate methods that search biclusters based on the concept of shift and scale invariance. Experimental results based on the yeast cell cycle and the human B-cell lymphoma datasets are reported. Finally, the performance of the proposed technique is compared with an optimization method based on the Nelder-Mead Simplex search algorithm.

Quantifying biomass consumption and carbon release from the California Rim fire by integrating airborne LiDAR and Landsat OLI data

Abstract Quantifying biomass consumption and carbon release is critical to understanding the role of fires in the carbon cycle and air quality. We present a methodology to estimate the biomass consumed and the carbon released by the California Rim fire by integrating postfire airborne LiDAR and multitemporal Landsat Operational Land Imager (OLI) imagery. First, a support vector regression (SVR) model was trained to estimate the aboveground biomass (AGB) from LiDAR‐derived metrics over the unburned area. The selected model estimated AGB with an R 2 of 0.82 and RMSE of 59.98 Mg/ha. Second, LiDAR‐based biomass estimates were extrapolated to the entire area before and after the fire, using Landsat OLI reflectance bands, Normalized Difference Infrared Index, and the elevation derived from LiDAR data. The extrapolation was performed using SVR models that resulted in R 2 of 0.73 and 0.79 and RMSE of 87.18 (Mg/ha) and 75.43 (Mg/ha) for the postfire and prefire images, respectively. After removing bias from the AGB extrapolations using a linear relationship between estimated and observed values, we estimated the biomass consumption from postfire LiDAR and prefire Landsat maps to be 6.58 ± 0.03 Tg (1012 g), which translate into 12.06 ± 0.06 Tg CO2e released to the atmosphere, equivalent to the annual emissions of 2.57 million cars.

A Comparison of Machine Learning Techniques Applied to Landsat-5 TM Spectral Data for Biomass Estimation

Abstract. Machine learning combines inductive and automated techniques for recognizing patterns. These techniques can be used with remote sensing datasets to map aboveground biomass (AGB) with an acceptable degree of accuracy for evaluation and management of forest ecosystems. Unfortunately, statistically rigorous comparisons of machine learning algorithms are scarce. The aim of this study was to compare the performance of the 3 most common nonparametric machine learning techniques reported in the literature, vis., Support Vector Machine (SVM), k-nearest neighbor (kNN) and Random Forest (RF), with that of the parametric multiple linear regression (MLR) for estimating AGB from Landsat-5 Thematic Mapper (TM) spectral reflectance data, texture features derived from the Normalized Difference Vegetation Index (NDVI), and topographical features derived from a digital elevation model (DEM). The results obtained for 99 permanent sites (for calibration/validation of the models) established during the winter of 2011 by systematic sampling in the state of Durango (Mexico), showed that SVM performed best once the parameterization had been optimized. Otherwise, SVM could be outperformed by RF. However, the kNN yielded the best overall results in relation to the goodness-of-fit measures. The findings confirm that nonparametric machine learning algorithms are powerful tools for estimating AGB with datasets derived from sensors with medium spatial resolution. Résumé. L’apprentissage automatique combine des techniques inductives et automatisées pour la reconnaissance des formes. Ces techniques peuvent être utilisées avec des ensembles de données de télédétection pour cartographier la biomasse aérienne « aboveground biomass » (AGB) avec un degré de précision acceptable pour l’évaluation et la gestion des écosystèmes forestiers. Malheureusement, des comparaisons statistiquement rigoureuses des algorithmes d’apprentissage automatique sont rares. Le but de cette étude était de comparer les performances des 3 méthodes d’apprentissage automatique non paramétriques les plus fréquemment rapportées dans la littérature, vis., les machines à vecteurs de support « Support Vector Machine » (SVM), les k plus proches voisins « k-nearest neighbor » (kNN) et les forêts aléatoires « Random Forest » (RF), avec celle de la régression linéaire multiple paramétrique (MLR) pour l’estimation de l’AGB provenant des données de réflectance spectrale de Landsat-5 Thematic Mapper (TM), des caractéristiques de texture dérivées de l’indice de végétation par différence normalisée « Normalized Difference Vegetation Index » (NDVI) et des caractéristiques topographiques dérivées d’un modèle numérique de terrain « digital elevation model » (DEM).Les résultats obtenus pour 99 sites permanents (pour la calibration/validation des modèles) établis au cours de l’hiver 2011 par l’échantillonnage systématique dans l’État de Durango (Mexique), ont montré que les SVM montrent leurs meilleures performances une fois que le paramétrage a été optimisé. Par ailleurs, les SVM pourraient être surpassées par les RF. Cependant, les kNN ont donné les meilleurs résultats globaux par rapport aux mesures d’ajustement. Les résultats confirment que les algorithmes d’apprentissage automatique non paramétriques sont des outils puissants pour l’estimation de l’AGB avec des ensembles de données provenant de capteurs avec une résolution spatiale moyenne.

A comparative study between two regression methods on LiDAR data: a case study

Airborne LiDAR (Light Detection and Ranging) has become an excellent tool for accurately assessing vegetation characteristics in forest environments. Previous studies showed empirical relationships between LiDAR and field-measured biophysical variables. Multiple linear regression (MLR) with stepwise feature selection is the most common method for building estimation models. Although this technique has provided very interesting results, many other data mining techniques may be applied. The overall goal of this study is to compare different methodologies for assessing biomass fractions at stand level using airborne Li-DAR data in forest settings. In order to choose the best methodology, a comparison between two different feature selection techniques (stepwise selection vs. genetic-based selection) is presented. In addition, classical MLR is also compared with regression trees (M5P). The results when each methodology is applied to estimate stand biomass fractions from an area of northern Spain show that genetically-selected M5P obtains the best results.