Unsupervised Deep Learning for Landslide Detection from Multispectral Sentinel-2 Imagery

This paper proposes a new approach based on an unsupervised deep learning (DL) model for landslide detection. Recently, supervised DL models using convolutional neural networks (CNN) have been widely studied for landslide detection. Even though these models provide robust performance and reliable results, they depend highly on a large labeled dataset for their training step. As an alternative, in this paper, we developed an unsupervised learning model by employing a convolutional auto-encoder (CAE) to deal with the problem of limited labeled data for training. The CAE was used to learn and extract the abstract and high-level features without using training data. To assess the performance of the proposed approach, we used Sentinel-2 imagery and a digital elevation model (DEM) to map landslides in three different case studies in India, China, and Taiwan. Using minimum noise fraction (MNF) transformation, we reduced the multispectral dimension to three features containing more than 80% of scene information. Next, these features were stacked with slope data and NDVI as inputs to the CAE model. The Huber reconstruction loss was used to evaluate the inputs. We achieved reconstruction losses ranging from 0.10 to 0.147 for the MNF features, slope, and NDVI stack for all three study areas. The mini-batch K-means clustering method was used to cluster the features into two to five classes. To evaluate the impact of deep features on landslide detection, we first clustered a stack of MNF features, slope, and NDVI, then the same ones plus with the deep features. For all cases, clustering based on deep features provided the highest precision, recall, F1-score, and mean intersection over the union in landslide detection.

Mixed Noise Estimation Model for Optimized Kernel Minimum Noise Fraction Transformation in Hyperspectral Image Dimensionality Reduction

Dimensionality reduction (DR) is of great significance for simplifying and optimizing hyperspectral image (HSI) features. As a widely used DR method, kernel minimum noise fraction (KMNF) transformation preserves the high-order structures of the original data perfectly. However, the conventional KMNF noise estimation (KMNF-NE) uses the local regression residual of neighbourhood pixels, which depends heavily on spatial information. Due to the limited spatial resolution, there are many mixed pixels in HSI, making KMNF-NE unreliable for noise estimation and leading to poor performance in KMNF for classification on HSIs with low spatial resolution. In order to overcome this problem, a mixed noise estimation model (MNEM) is proposed in this paper for optimized KMNF (OP-KMNF). The MNEM adopts the sequential and linear combination of the Gaussian prior denoising model, median filter, and Sobel operator to estimate noise. It retains more details and edge features, making it more suitable for noise estimation in KMNF. Experiments using several HSI datasets with different spatial and spectral resolutions are conducted. The results show that, compared with some other DR methods, the improvement of OP-KMNF in average classification accuracy is up to 4%. To improve the efficiency, the OP-KMNF was implemented on graphics processing units (GPU) and sped up by about 60× compared to the central processing unit (CPU) implementation. The outcome demonstrates the significant performance of OP-KMNF in terms of classification ability and execution efficiency.

Spectral-Spatial Hyperspectral Image Semisupervised Classification by Fusing Maximum Noise Fraction and Adaptive Random Multigraphs

Hyperspectral images (HSIs) contain large amounts of spectral and spatial information, and this provides the possibility for ground object classification. However, when using the traditional method, achieving a satisfactory classification result is difficult because of the insufficient labeling of samples in the training set. In addition, parameter adjustment during HSI classification is time-consuming. This paper proposes a novel fusion method based on the maximum noise fraction (MNF) and adaptive random multigraphs for HSI classification. Considering the overall spectrum of the object and the correlation of adjacent bands, the MNF was utilized to reduce the spectral dimension. Next, a multiscale local binary pattern (LBP) analysis was performed on the MNF dimension-reduced data to extract the spatial features of different scales. The obtained multiscale spatial features were then stacked with the MNF dimension-reduced spectral features to form multiscale spectral-spatial features (SSFs), which were sent into the RMG for HSI classification. Optimal performance was obtained by fusion. For all three real datasets, our method achieved competitive results with only 10 training samples. More importantly, the classification parameters corresponding to different hyperspectral data can be automatically optimized using our method.

Nonlinear Noise Reduction for the Airborne Transient Electromagnetic Method based on Kernel Minimum Noise Fraction

The airborne transient electromagnetic method has become a powerful tool to explore deep resource and tectonic structures. However, aircraft vibrations and flight environments produce very strong and complex nonlinear noise and result in poor data quality compared to ground transient electromagnetic methods. Consequently, the reduction of airborne electromagnetic noises is of vital importance to data inversion and imaging. To suppress and remove the nonlinear noise, we propose using kernel minimum noise fraction (KMNF), which is a nonlinear generalized method of minimum noise fraction. First, an adaptive variable window-width filtering algorithm is used to evaluate the noises and perform the preliminary denoising. Then, we adopt the two filter methods, which are minimum noise fraction (MNF) and KMNF to suppress the noise. The results show that these two methods can both suppress noise and make the decay curves smooth, but kernel MNF is more effective for the nonlinear characteristics of noise and it does not weaken the anomaly. Finally, field data from the Qinling mine area is processed, using the MNF and KMNF methods. The results show that nonlinear noise is suppressed by both methods but the results of KMNF are better than those of the linear MNF method.

Entomological Remote Dark Field Signal Extraction by Maximum Noise Fraction and Unsupervised Clustering for Species Identification

Characterization of flying insects in-situ measurement using remote sensing spectroscopy is an emerging research field. Also, most analysis techniques in remote sensing spectroscopy are based on the use of an intensity threshold which introduces indeterminacies in the number of detected specimens. In this manuscript, we investigated the possibility of analysing passive remote sensing spectroscopy measurement data using the maximum noise fraction method. The results obtained show that this analysis technique can help to overcome the measurement of background noise in spectroscopic measurements.

Alteration Mapping for Lead-Zinc-Barium Mineralization in Parts of the Southern Benue Trough, Nigeria, Using ASTER Multispectral Data

The lead-zinc-barium deposits of the southern Benue Trough, Nigeria belong to a suite of clastic dominated fracture filling hydrothermal vein deposits. The alteration types and spread are poorly known yet required to aid exploration. Band ratio composites (BRC), Principal Component Analysis (PCA), and Minimum Noise Fraction (MNF) were applied to a full scene Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) imagery covering the study area. Spectral analysis of sulphide minerals known in the area led to the development of the (B1+B3)/2 ratio, which provided a highly effective sulphide discriminant. PCA and MNF bands with high eigenvectors in the absorption features of target minerals qualified as colour composite candidates for alteration mapping. This study demonstrated the effectiveness of combining the BRC, PCA and MNF techniques in the discrimination of ferric-ferrous/sulphide and silica alteration zones in the Southern Benue Trough.

Superpixel-Based Minimum Noise Fraction Feature Extraction for Classification of Hyperspectral Images

In this paper, a novel feature extraction technique called SuperMNF is proposed, which is an extension of the minimum noise fraction (MNF) transformation. In SuperMNF, each superpixel has its own transformation matrix and MNF transformation is performed on each superpixel individually. The basic idea behind the SuperMNF is that each superpixel contains its specific signal and noise covariance matrices which are different from the adjacent superpixels. The extracted features, owning spatial-spectral content and provided in the lower dimension, are classified by maximum likelihood classifier and support vector machines. Experiments that are conducted on two real hyperspectral images, named Indian Pines and Pavia University, demonstrate the efficiency of SuperMNF since it yielded more promising results than some other feature extraction methods (MNF, PCA, SuperPCA, KPCA, and MMP).