This paper explores hyperspectral reflectance factor estimation using Gaussian process regression with multispectral- and trichromatic measurements. Estimations are performed in visible- (400-700 nm) or visible-near infrared (400-980 nm) wavelength ranges using the learning-based approach, where sensor and light spectral characteristics are not required. We first construct new estimation models via Gaussian processes, show connection to previous kernel-based models, and then evaluate new models by using marginal likelihood optimization within the probabilistic interpretation. By using standard spectral ensembles and several images in experiments, we evaluate new models with anisotropic radial- and combination kernels (process covariance), marginal likelihood optimization (parameter selection), as well as with input data transformations (pre-processing). Several new Gaussian process models provide spectral accuracy improvements for simulated and real data, when compared with the previous kernel-based models. Most versatile new model is using spectral subspace coordinate learning and combination kernels, and can be efficiently optimized via marginal likelihood. Preliminary results suggest that new models provide uncertainty estimates, which can be used for iterative training set augmentation.

In this letter, we evaluated the pixel-level and plot-level tree species classification of Scots Pine, Norway Spruce, and deciduous birch in a boreal forest using 64-band AisaEAGLE II hyperspectral data in a wavelength range of 400-1000 nm. First, band selection was performed using a sparse logistic regression-based feature selection algorithm with pixel-level and plot-level data in case of balanced and imbalanced training data. This resulted in 8-11 selected hyperspectral bands, depending on the properties of the data used. We evaluated a tree species classification with 8-11 selected hyperspectral bands directly for a least squares support vector machine (LS-SVM)-based pixel-level classification with a relatively small training set size (0.5%-1.5% of the total data) and obtained an accuracy and kappa of around 93.50% and 0.90, respectively. These results are around 0.53%-0.94% points lower than those obtained using all of the hyperspectral bands. Second, one important wavelength region highlight by the selected bands was used to modify the sensor sensitivity configuration in the Leica Airborne Digital Sensor 40 (ADS40) multispectral sensor. Using a simulation model and the hyperspectral data, the modified and standard Leica ADS40 sensor responses were simulated and compared, and the modified system simulated response indicates a 3%-5% point improvement in the pixel-level and plot-level LS-SVM classification accuracy compared with the simulated responses of the standard Leica ADS40 band configuration.

This paper focuses on the use of multispectral measurements to classify remotely sensed radiance and reflectance information into three tree species, Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies (L.) H. Karst.), and birch (Betula pubescens Ehrh., Betula pendula Roth), using a Support Vector Machine (SVM) algorithm. The features used for the classifier are radiometric involving different viewing angles, but without textural information. At-sensor radiance (ASR) signals used here were obtained using a four-band Leica ADS40-SH52 airborne sensor. The experiments were carried out in a forest area at Hyytiälä, in southern Finland (61°50' N, 24°20' E), which has been widely used for similar purposes, so that detailed tree-level information has been reported previously. The flight was carried out on August 23, 2008. ADS40 ASR measurements can be converted to ground reflectance signatures in two viewing directions using atmosphere and BRDF modeling implemented in Leica XPro 4.2 software. Taking into account the assumptions entailed in the radiometric model, the classification performance of the ground reflectance is evaluated only for the pixel values under sunlit conditions and is compared with the performance of the ASR data. The sunlit and shaded parts of the tree crown were extracted based on the use of LiDAR data for crown shape modeling. The classification results for the real multispectral measurements are compared with the earlier results obtained with simulated Leica ADS40 at-sensor radiance response values which were based on the ground-level high-resolution ground reflectance factor measurements using a single viewing direction. The simulated classification accuracy was 75-79% with the original four bands, while it was up to 85-88%, using the simulated fifth channel. It was found here that the classification accuracy using comparable real ADS40-SH52 four-band data and one viewing angle was 75-79% and increased to 78-82% with two viewing angles. The results show that the best-case classification accuracy with real data can reach 88% if trees are modeled as objects with sunlit and shaded areas, and multiple measurements are available for every tree. The results suggest that ground reflectance estimation with normalization of anisotropic reflectance behavior leads to similar classification performance to ASR data, but can in some cases improve the generalization properties of training data.

The information content of remotely sensed data depends primarily on the spatial and spectral properties of the imaging device. This paper focuses on the classification performance of the different spectral features (hyper- and multispectral measurements) with respect to three tree species. The Support Vector Machine was chosen as the classification algorithm for these features. A simulated optical radiation model was constructed to evaluate the identification performance of the given multispectral system for the tree species, and the effects of spectral-band selection and data preprocessing were studied in this setting. Simulations were based on the reflectance measurements of the pine (Pinus sylvestris L.), spruce [Picea abies (L.)H.Karst.], and birch trees (Betula pubescens Ehrh. andBetula pendula Roth). Leica ADS80 airborne sensor with four spectral bands (channels) was used as a fixed multispectral sensor system that leads to response values for the at-sensor radiance signal. Results suggest that this four-band system has inadequate classification performance for the three tree species. The simulations demonstrate on average a 5-15 percentage points improvement in classification performance when the Leica system is combined with one additional spectral band. It is also demonstrated for the Leica data that feature mapping through a Mahalanobis kernel leads to a 5-10 percentage points improvement in classification performance when compared with other kernels.