Weiyue Li

Nonlinear Dimensionality Reduction via the ENH-LTSA Method for Hyperspectral Image Classification

The problems of neglecting spatial features in hyperspectral imagery (HSI) and the high complexity of Local Tangent Space Alignment (LTSA) still exist in the nonlinear dimensionality reduction with LTSA for classification. Therefore, this paper proposes an innovative ENH-LTSA (Enhanced-Local Tangent Space Alignment) method to solve the two problems. First, random projection is introduced to preliminarily reduce the dimension of HSI data. It aims to improve the speed of neighbor searching and the local tangent space construction. Then, the new method presents the similarity measure via the adaptive weighted summation kernel (AWSK) distance. The AWSK distance considers both spectral and spatial features in HSI data, and attempts to ameliorate the k-nearest neighbors (KNNs) of each pixel. Furthermore, the adaptive spatial window is proposed to automatically estimate the proper window size for the description of spatial features. After that, fast approximate KNNs graph construction via Recursive Lanczos Bisection is incorporated into the new method to reduce the complexity of KNNs searching. When finishing constructing each local tangent space, the new method uses a fast low-rank approximate singular value decomposition to speed up eigenvalue decomposition of the global alignment matrix that is constituted with local manifold coordinates. Five groups of experiments with two different HSI datasets are designed to completely analyze and testify the ENH-LTSA method. Experimental results show that ENH-LTSA outperforms LTSA, both in classification results and in computational speed.

Low-rank and sparse matrix decomposition-based anomaly detection for hyperspectral imagery

Abstract A low-rank and sparse matrix decomposition (LRaSMD) detector has been proposed to detect anomalies in hyperspectral imagery (HSI). The detector assumes background images are low-rank while anomalies are gross errors that are sparsely distributed throughout the image scene. By solving a constrained convex optimization problem, the LRaSMD detector separates the anomalies from the background. This protects the background model from corruption. An anomaly value for each pixel is calculated using the Euclidean distance, and anomalies are determined by thresholding the anomaly value. Four groups of experiments on three widely used HSI datasets are designed to completely analyze the performances of the new detector. Experimental results show that the LRaSMD detector outperforms the global Reed-Xiaoli (GRX), the orthogonal subspace projection-GRX, and the cluster-based detectors. Moreover, the results show that LRaSMD achieves equal or better detection performance than the local support vector data description detector within a shorter computational time.

Model test study on monitoring dynamic process of slope failure through spatial sensor network

Landslides represent a major type of natural hazards worldwide. For development of risk mitigation capabilities, an effective system for monitoring dynamic process of slope failure, capable of gathering spatially distributed information before, during and after a landslide occurrence at real-time manner is essential. A spatial sensor network (SSN), which integrates the real-time communication infrastructure and observations from in situ sensors and remote sensing platforms, offers an efficient and effective approach for such purpose. In this paper, a SSN-based landslide monitoring system was designed and evaluated through a model test study conducted at Tongji University, China. This system, MUNOLD (MUlti-Sensor Network for Observing Landslide Disaster), has been designed as a comprehensive monitoring framework, including sensor observations, multi-channel wireless communication, remote data storage, visualization, data processing and data analysis. In this model test study, initial experimentation demonstrated the capabilities of the MUNOLD system for collecting real-time information about the dynamic process and propagation of slope failure. Innovatively, generated from the high-speed stereo images, the sequential surface deformation vector field can be created and may exhibit the dynamic process during the extremely critical and short period of the slope failure. After this model test study, the MUNOLD system is going to be further improved and extended in a landslide prone region in Sichuan Province, China.

Involvement of p21 (waf1) in merlin deficient sporadic vestibular schwannomas

Previous studies have demonstrated that merlin acts as a tumor suppressor by blocking Ras-mediated signaling. However, the mechanism by which merlin controls cell proliferation has remained obscure. Here we show that merlin deficient tumors exhibited loss of p21, concomitant with elevated CDKs/cyclin D1 levels in sporadic vestibular schwannomas (VS) from clinic patients. Likewise, silencing of merlin gene expression in the cell lines resulted in down-regulation of p21. Furthermore, we find that merlin-enhanced p21 protein stability, rather than increased RNA accumulation, was responsible for the elevated p21 levels. Interestingly, p21 was required to maintain merlin levels and the inhibitory effect of merlin on Ras signaling was partially overridden by knockdown of p21. Consistent with the observation that over-expression of merlin arrested cell growth at G1-phase, the current study indicates that merlin exerts its antiproliferative effect, at least in part, by maintaining p21 expression, and loss of p21 is a prominent feature of merlin deficient schwannomas.

Architecture planning and geo-disasters assessment mapping of landslide by using airborne lidar data and UAV images

After the operation of GPS/IMU direct geo-referencing, segmentation, filtering, classification of scattered point data and aerial triangulation on airborne LiDAR(Light Detection and Ranging) data, the accurate and high-resolution DEM of the study area in the west part of Zengcheng city, Guangdong, China was constructed. In addition, unmanned aerial vehicle (UAV) images were used for ground objects identification. Landslides occur frequently in summer in the city because of heavy rainfall. The LiDAR data (point cloud) and the mosaic images were then combined to produce the suitability distribution maps by considering Several factors, such as slope gradient, slope aspect, on-the-spot investigation data etc The maps can then be used to analyze the potential risk of landslides and assess the risk level around some buildings. The experiment results show that the method based on LiDAR data and UAV images can rapidly and accurately survey the terrain of the study area and also provides useful information for architectural design.

A public Cloud-based China’s Landslide Inventory Database (CsLID): development, zone, and spatiotemporal analysis for significant historical events, 1949-2011

Landslide inventory plays an important role in recording landslide events and showing their temporal-spatial distribution. This paper describes the development, visualization, and analysis of a Chinas Landslide Inventory Database (CsLID) by utilizing Google’s public cloud computing platform. Firstly, CsLID (Landslide Inventory Database) compiles a total of 1221 historical landslide events spanning the years 1949-2011 from relevant data sources. Secondly, the CsLID is further broken down into six zones for characterizing landslide cause-effect, spatiotemporal distribution, fatalities, and socioeconomic impacts based on the geological environment and terrain. The results show that among all the six zones, zone V, located in Qinba and Southwest Mountainous Area is the most active landslide hotspot with the highest landslide hazard in China. Additionally, the Google public cloud computing platform enables the CsLID to be easily accessible, visually interactive, and with the capability of allowing new data input to dynamically augment the database. This work developed a cyber-landslide inventory and used it to analyze the landslide temporal-spatial distribution in China.

Building a water feature extraction model by integrating aerial image and lidar point clouds

An innovative model for extracting water regions from aerial images fused with light detection and ranging (lidar) data is proposed in this article. This model extracts water features from coarse to fine levels of accuracy by considering special spectral bands of existing airborne lidar systems and their spectral characteristics. The particular model consists of two parts, namely inexact water region recognition and precise water extraction. (1) A strategy of using a triangulated irregular network (TIN) is introduced to describe point clouds with a particular structure. A TIN coarsely divides the network into water and non-water regions through a threshold, which can be determined through an equation by inputting the minimum width and point density of water regions. The coarsely defined water region can be detected through overlay analysis between the aerial image and the raster surface generated from the TIN. (2) An improved mean-shift algorithm is used to remove most land pixels from the roughly recognized water to obtain precise water edges from coarse water. A new empirical formula to describe distance between multi-dimensional data is adopted. Using the mean-shift algorithm and empirical distance function, accurate water edge features are extracted from inexact water region(s). In addition, the classification field of lidar point clouds is used to remove land pixels from water features. A case study based on a point cloud data set and an aerial image is conducted to evaluate the feasibility and accuracy of the proposed model. Spatial distances between checkpoints and extracted water edges, as well as the confusion matrix of mean-shift classification, are adopted as measurements of accuracy for the extracted water edges in two case regions. Evaluation results show that the proposed model achieved continuous water-edge features, and that spatial accuracy of water edges is 0.3 to 0.4 m, at approximately the 1–2-pixel level, which is more than four times better than the maximum-likelihood classification method. General accuracy of the confusion matrix shows that mean-shift classification in the proposed model is better than 95%, which indicates excellent results.

Spatio-temporal analysis and simulation on shallow rainfall-induced landslides in China using landslide susceptibility dynamics and rainfall I-D thresholds

An empirical simulation method to simulate the possible position of shallow rainfall-induced landslides in China has been developed. This study shows that such a simulation may be operated in real-time to highlight those areas that are highly prone to rainfall-induced landslides on the basis of the landslide susceptibility index and the rainfall intensity-duration (I-D) thresholds. First, the study on landslide susceptibility in China is introduced. The entire territory has been classified into five categories, among which high-susceptibility regions (Zone 4- ‘High’ and 5-‘Very high’) account for 4.15% of the total extension of China. Second, rainfall is considered as an external triggering factor that may induce landslide initiation. Real-time satellite-based TMPA 3B42 products may provide real rainfall spatial and temporal patterns, which may be used to derive rainfall duration time and intensity. By using a historical record of 60 significant past landslides, the rainfall I-D equation has been calibrated. The rainfall duration time that may trigger a landslide has resulted between 3 hours and 45 hours. The combination of these two aspects can be exploited to simulate the spatiotemporal distribution of rainfall-induced landslide hazards when rainfall events exceed the rainfall I-D thresholds, where the susceptibility category is ‘high’ or ‘very high’. This study shows a useful tool to be part of a systematic landslide simulation methodology, potentially providing useful information for a theoretical basis and practical guide for landslide prediction and mitigation throughout China.

Rainstorm-induced shallow landslides process and evaluation – a case study from three hot spots, China

ABSTRACT The critical stage in the evaluation of rainfall-induced landslide failure is in formulating reasonable models to better simulate spatiotemporal changes of slopes in the hilly terrains. A physically based model can take into account the contribution of rainfall infiltration and shear strength of saturated soil layer, and therefore help revealing the landslide formation mechanisms. This paper presents a physically based approach to simulate the landslide process triggered by rainstorm. On the basis of previous solutions, we select the simplified infiltration model Slope-Infiltration-Distributed Equilibrium (SLIDE) to illustrate the dynamical relations between factor of safety (FS) and accumulation of rainfall over time. This model is tested with three representative landslide events in the southwest, southeast, and south central of China during rainstorm. Results show that the time of landslide failure predicted from the SLIDE model is consistent with the reality. Meanwhile, this paper illustrates the differences of FS among the different slope gradients in the vicinity of same soil texture and relationship between FS and rainfall accumulation. This work formulates a methodology of rainstorm-induced landslide evaluation and improves upon the existing landslide prediction methods.

Assessment of Regional Shallow Landslide Stability Based on Airborne Laser Scanning Data in the Yingxiu Area of Sichuan Province (China)

Abstract This study focuses on analyzing the slope stabilities in a landslide-prone area of Yingxiu Town, Sichuan Province (China). Airborne Laser Scanning (ALS) data were acquired to derive a Digital Elevation Model (DEM) with sufficient accuracy and resolution, as an input for the regional landslide stability analysis. The one-dimensional hydrological model—Stability Index Map (SINMAP), functioning with topographic data, geological settings, and rainfall conditions, was used as a simplified model for slope stability mapping. In this study, the investigated region was classified into six stability levels, and data reliability was subsequently checked with reference to recent landslide inventories. Several experiments have shown that the quality of ALS data played a key-role in the slope stability inside the SINMAP model regarding the point cloud density and the random error. Higher point cloud density may construct higher precision of DEM, however, it may also produce more noises. Although with these uncertainties, using ALS data and its derived high precision DEM, the physically-based SINMAP model is expected to provide a solid basis for further landslide susceptibility mapping at regional scale.