Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Central to the looming paradigm shift toward data-intensive science, machine-learning techniques are becoming increasingly important. In particular, deep learning has proven to be both a major breakthrough and an extremely powerful tool in many fields. Shall we embrace deep learning as the key to everything? Or should we resist a black-box solution? These are controversial issues within the remote-sensing community. In this article, we analyze the challenges of using deep learning for remote-sensing data analysis, review recent advances, and provide resources we hope will make deep learning in remote sensing seem ridiculously simple. More importantly, we encourage remote-sensing scientists to bring their expertise into deep learning and use it as an implicit general model to tackle unprecedented, large-scale, influential challenges, such as climate change and urbanization.

/pdf/deep-learning-in-remote-sensing-a-comprehensive-review-and-3o9ov8r19z.pdf

Deep Learning in Remote Sensing: A Comprehensive Review and List of Resources

Deep-learning (DL) algorithms, which learn the representative and discriminative features in a hierarchical manner from the data, have recently become a hotspot in the machine-learning area and have been introduced into the geoscience and remote sensing (RS) community for RS big data analysis. Considering the low-level features (e.g., spectral and texture) as the bottom level, the output feature representation from the top level of the network can be directly fed into a subsequent classifier for pixel-based classification. As a matter of fact, by carefully addressing the practical demands in RS applications and designing the input?output levels of the whole network, we have found that DL is actually everywhere in RS data analysis: from the traditional topics of image preprocessing, pixel-based classification, and target recognition, to the recent challenging tasks of high-level semantic feature extraction and RS scene understanding.

Deep Learning for Remote Sensing Data: A Technical Tutorial on the State of the Art

Deep learning (DL) is a powerful state-of-the-art technique for image processing including remote sensing (RS) images. This letter describes a multilevel DL architecture that targets land cover and crop type classification from multitemporal multisource satellite imagery. The pillars of the architecture are unsupervised neural network (NN) that is used for optical imagery segmentation and missing data restoration due to clouds and shadows, and an ensemble of supervised NNs. As basic supervised NN architecture, we use a traditional fully connected multilayer perceptron (MLP) and the most commonly used approach in RS community random forest, and compare them with convolutional NNs (CNNs). Experiments are carried out for the joint experiment of crop assessment and monitoring test site in Ukraine for classification of crops in a heterogeneous environment using nineteen multitemporal scenes acquired by Landsat-8 and Sentinel-1A RS satellites. The architecture with an ensemble of CNNs outperforms the one with MLPs allowing us to better discriminate certain summer crop types, in particular maize and soybeans, and yielding the target accuracies more than 85% for all major crops (wheat, maize, sunflower, soybeans, and sugar beet).

Deep Learning Classification of Land Cover and Crop Types Using Remote Sensing Data

A Viking Lander 1 image was modeled as mixtures of reflectance spectra of palagonite dust, gray andesitelike rock, and a coarse rocklike soil. The rocks are covered to varying degrees by dust but otherwise appear unweathered. Rocklike soil occurs as lag deposits in deflation zones around stones and on top of a drift and as a layer in a trench dug by the lander. This soil probably is derived from the rocks by wind abrasion and/or spallation. Dust is the major component of the soil and covers most of the surface. The dust is unrelated spectrally to the rock but is equivalent to the global-scale dust observed telescopically. A new method was developed to model a multispectral image as mixtures of end-member spectra and to compare image spectra directly with laboratory reference spectra. The method for the first time uses shade and secondary illumination effects as spectral end-members; thus the effects of topography and illumination on all scales can be isolated or removed. The image was calibrated absolutely from the laboratory spectra, in close agreement with direct calibrations. The method has broad applications to interpreting multispectral images, including satellite images.

Spectral mixture modeling - A new analysis of rock and soil types at the Viking Lander 1 site. [on Mars]

Recently, convolutional neural networks have demonstrated excellent performance on various visual tasks, including the classification of common two-dimensional images. In this paper, deep convolutional neural networks are employed to classify hyperspectral images directly in spectral domain. More specifically, the architecture of the proposed classifier contains five layers with weights which are the input layer, the convolutional layer, the max pooling layer, the full connection layer, and the output layer. These five layers are implemented on each spectral signature to discriminate against others. Experimental results based on several hyperspectral image data sets demonstrate that the proposed method can achieve better classification performance than some traditional methods, such as support vector machines and the conventional deep learning-based methods.

/pdf/deep-convolutional-neural-networks-for-hyperspectral-image-1da7yalxyy.pdf

Deep Convolutional Neural Networks for Hyperspectral Image Classification

In this paper, an efficient statistical model, called generalized Gamma distribution (GΓD), for the empirical modeling of synthetic aperture radar (SAR) images is proposed. The GΓD forms a large variety of alternative distributions (especially including Rayleigh, exponential, Nakagami, Gamma, Weibull, and log-normal distributions commonly used for the probability density function (pdf) of SAR images as special cases), and is flexible to model the SAR images with different land-cover typologies. Moreover, based on second-kind cumulants, a closed-form estimator for GΓD parameters is derived by exploiting the second-order approximation for Polygamma function. Without involving the numerical iterative process for solutions, this estimator is computationally efficient and, hence, can make the GΓD convenient for applications in the online SAR image processing. Finally, experimental results from tests carried out with actual SAR images demonstrate that the GΓD can achieve better goodness of fit than the state-of-the-art pdfs.

On the Empirical-Statistical Modeling of SAR Images With Generalized Gamma Distribution

Spectral unmixing aims at estimating the fractional abundances of a set of pure spectral materials (endmembers) in each pixel of a hyperspectral image. The wide availability of large spectral libraries has fostered the role of sparse regression techniques in the task of characterizing mixed pixels in remotely sensed hyperspectral images. A general solution for sparse unmixing methods consists of using the $\ell _{1}$ regularizer to control the sparsity, resulting in a very promising performance but also suffering from sensitivity to large and small sparse coefficients. A recent trend to address this issue is to introduce weighting factors to penalize the nonzero coefficients in the unmixing solution. While most methods for this purpose focus on analyzing the hyperspectral data by considering the pixels as independent entities, it is known that there exists a strong spatial correlation among features in hyperspectral images. This information can be naturally exploited in order to improve the representation of pixels in the scene. In order to take advantage of the spatial information for hyperspectral unmixing, in this paper, we develop a new spectral–spatial weighted sparse unmixing (S2WSU) framework, which uses both spectral and spatial weighting factors, further imposing sparsity on the solution. Our experimental results, conducted using both simulated and real hyperspectral data sets, illustrate the good potential of the proposed S2WSU, which can greatly improve the abundance estimation results when compared with other advanced spectral unmixing methods.

Spectral–Spatial Weighted Sparse Regression for Hyperspectral Image Unmixing

Anomaly detection is an important task in hyperspectral remote sensing. Most widely used detectors, such as Reed–Xiaoli (RX), have been developed only using original spectral signatures, which may lack the capability of signal enhancement and noise suppression. In this article, an effective alternative approach, fractional Fourier entropy (FrFE)-based hyperspectral anomaly detection method, is proposed. First, fractional Fourier transform (FrFT) is employed as preprocessing, which obtains features in an intermediate domain between the original reflectance spectrum and its Fourier transform with complementary strengths by space-frequency representations. It is desirable for noise removal so as to enhance the discrimination between anomalies and background. Furthermore, an FrFE-based step is developed to automatically determine an optimal fractional transform order. With a more flexible constraint, i.e., Shannon entropy uncertainty principle on FrFT, the proposed method can significantly distinguish signal from background and noise. Finally, the proposed FrFE-based anomaly detection method is implemented in the optimal fractional domain. Experimental results obtained on real hyperspectral datasets demonstrate that the proposed method is quite competitive.

Hyperspectral Anomaly Detection by Fractional Fourier Entropy

One fundamental problem in Earth Vision is to accurately find the locations and identify the categories of the interesting objects in the aerial images, for which oriented bounding boxes (OBBs) are usually employed to depict better the objects emerging with arbitrary orientations. However, the regression of the OBBs always suffers from the ambiguous problem in the definition of the regression targets, which often reduces the convergency efficiency and decreases the detection accuracy. Although there are some methods like the binary segmentation map that can handle this problem, it brings a new problem of ambiguous background pixels in the OBBs. In this article, we propose to cast the OBB regression as a center-probability-map (CenterMap)-prediction problem, thus largely eliminating the ambiguities on the target definitions and the background pixels. The predicted CenterMaps are then used to generate the OBBs. The CenterMap OBB representation is simple, yet effective. Furthermore, to distinguish better the interesting objects from the cluttered background, a weighted pseudosegmentation-guided attention network is adopted to provide the object-level features for predicting the horizontal bounding boxes and the OBBs. The experimental results on three widely used data sets, i.e., DOTA, HRSC2016, and UCAS-AOD, demonstrate the effectiveness of our proposed method.

Heng-Chao Li

Papers

Deep Convolutional Neural Networks for Hyperspectral Image Classification

On the Empirical-Statistical Modeling of SAR Images With Generalized Gamma Distribution

Spectral–Spatial Weighted Sparse Regression for Hyperspectral Image Unmixing

Hyperspectral Anomaly Detection by Fractional Fourier Entropy

Learning Center Probability Map for Detecting Objects in Aerial Images