Identifying urban building function plays a critical role in understanding the complexness of urban construction and improving the effectiveness of urban planning. The emergence of user generated ...contents has brought access to massive semantic information which complements the traditional remote sensing data for identifying urban building functions and exploring the spatial structure in urban environment. This article proposes a stepwise identification framework for urban building functions based on remote sensing imagery and point of interests (POIs) data, which merges the spatial similarity of buildings and kernel density to improve the identification accuracy and completeness. Taking Wuhan as an example, Google earth images and POI data were obtained to identify the seven primary categories for the individual buildings in the core urban area. The results suggest that the proposed stepwise framework is feasible to identify the urban building functions as the identification results exhibit the superiority in terms of accuracy and completeness. Our results suggest that the identification of urban building function is sensitive to the bandwidth of kernel density estimation and 200 meter is the optimal size. The findings also indicate that significant spatial agglomeration exists in residential and commercial buildings at both macro and microlevels.
•Integrate NetKDE and Moran’s I for accident hot spot detection in a network space.•Statistical significance is assessed with two kinds of Monte Carlo simulation.•NetKDE approach results in spatially ...more contiguous and fewer number of clusters.
Kernel density estimation (KDE) has long been used for detecting traffic accident hot spots and network kernel density estimation (NetKDE) has proven to be useful in accident analysis over a network space. Yet, both planar KDE and NetKDE are still used largely as a visualization tool, due to the missing of quantitative statistical inference assessment. This paper integrates NetKDE with local Moran’I for hot spot detection of traffic accidents. After density is computed for road segments through NetKDE, it is then used as the attribute for computing local Moran’s I. With an NetKDE-based approach, conditional permutation, combined with a 100-m neighbor for Moran’s I computation, leads to fewer statistically significant “high-high” (HH) segments and hot spot clusters. By conducting a statistical significance analysis of density values, it is now possible to evaluate formally the statistical significance of the extensiveness of locations with high density values in order to allocate limited resources for accident prevention and safety improvement effectively.
Data-based procedures for monitoring the operating performance of a PV system are proposed in this article. The only information required to apply the procedures is the availability of system ...measurements, which are routinely on-line collected via sensors. Here, kernel-based machine learning methods, including support vector regression (SVR) and Gaussian process regression (GPR), are used to model multivariate data from the PV system for fault detection because of their flexibility and capability to nonlinear approximation. Essentially, the SVR and GPR models are adopted to obtain residuals to detect and identify occurred faults. Then, residuals are passed through an exponential smoothing filter to reduce noise and improve data quality. In this work, a monitoring scheme based on kernel density estimation is used to sense faults by examining the generated residuals. Several different scenarios of faults were considered in this study, including PV string fault, partial shading, PV modules short-circuited, module degradation, and line-line faults on the PV array. Using data from a 20 MWp grid-connected PV system, the considered faults were successfully traced using the developed procedures. Also, it has been demonstrated that GPR-based monitoring procedures achieve better detection performance over SVRs to monitor PV systems.
The problem of computing empirical cumulative distribution functions (ECDF) efficiently on large, multivariate datasets, is revisited. Computing an ECDF at one evaluation point requires O(N) ...operations on a dataset composed of N data points. Therefore, a direct evaluation of ECDFs at N evaluation points requires a quadratic O(N2) operations, which is prohibitive for large-scale problems. Two fast and exact methods are proposed and compared. The first one is based on fast summation in lexicographical order, with a O(NlogN) complexity and requires the evaluation points to lie on a regular grid. The second one is based on the divide-and-conquer principle, with a O(Nlog(N)(d−1)∨1) complexity and requires the evaluation points to coincide with the input points. The two fast algorithms are described and detailed in the general d-dimensional case, and numerical experiments validate their speed and accuracy. Secondly, a direct connection between cumulative distribution functions and kernel density estimation (KDE) is established for a large class of kernels. This connection paves the way for fast exact algorithms for multivariate kernel density estimation and kernel regression. Numerical tests with the Laplacian kernel validate the speed and accuracy of the proposed algorithms. A broad range of large-scale multivariate density estimation, cumulative distribution estimation, survival function estimation and regression problems can benefit from the proposed numerical methods.
The vessel Automatic Identification System (AIS) data collected by satellites have the features of large coverage area and large data volume, and they are instantaneous discrete data rather than ...time-continuous data, so the data has large dispersion with many noise points. This poses a challenge for vessel sailing route extraction. This paper proposes a vessel sailing route extraction method which consists of the fast Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm and the Kernel Density Estimation-based Outlier Factor (KDE-based OF) noise reduction algorithm. The method in this paper firstly improves the clustering discrimination method in the DBSCAN algorithm to separate trajectories in different directions. Secondly, this paper extracts a fast clustering algorithm based on the density clustering algorithm to reduce its computing time overhead with satellite big data. Finally, this paper proposes the KDE-based OF processing algorithm, which calculates the outlier probability distribution value of the trajectory points through the algorithm to eliminate the edge trajectory points with low probability distribution. Based on the actual satellite vessel AIS data, this paper conducts multi-method comparisons and performance analysis experiments. Experiments show that the proposed method has the best stability and advancement.
•A method for extracting vessel sailing routes algorithm is proposed.•An algorithm suitable for clustering cross-channel trajectories is proposed.•Experiments demonstrate the proposed method performs better than other methods.
Multilevel Monte Carlo (MLMC) method, enhanced by a smoothing technique based on Kernel Density Estimation, is coupled with the Finite-Difference Time-Domain (FDTD) algorithm in order to estimate the ...probability distribution of any quantity of interest, for uncertainty quantification in electromagnetic problems. It is shown that such enhanced MLMC-FDTD is faster than conventional Monte Carlo FDTD while inheriting its advantages of robustness, simplicity and generality, unlike other uncertainty analysis methods, such as the perturbation and the moment methods that cannot be used to straightforwardly estimate probability distribution, or the polynomial chaos method that suffers from the curse of dimensionality problem or even fails.
•Migratory corridors and core stopover sites were identified to guide conservation.•Movement patterns and utilization intensity en route have seasonal differences.•Seasonal migratory differences ...affected the efficiency of protection systems.•High selection in artificial surfaces calls for eco-friendly land use mode.•Satellite tracking and remote sensing facilitate the study of movement ecology.
Migratory species interact with different ecosystems in different regions during migration, making them more environmentally sensitive and therefore more vulnerable to extinction. Long migration routes and limited conservation resources desire clear identification of conservation priorities to improve the allocation efficiency of conservation resources. Clarifying the spatio-temporal heterogeneity of the utilization intensity during migration is an effective way to guide the conservation areas and priority. 12 Oriental White Storks (Ciconia boyciana), listed as an “endangered” species by the IUCN, were equipped with satellite-tracking loggers to record their hourly location throughout the year. Then, combined with remote sensing and dynamic Brownian Bridge Movement Model (dBBMM), characteristics and differences between spring and autumn migration were identified and compared. Our findings revealed that: (1) the Bohai Rim has always been the core stopover area for the Storks’ spring and autumn migration, but the utilization intensity has spatial differences; (2) differences in habitat selection resulted in differences in the Storks’ spatial distribution, thus affecting the efficiency of existing conservation systems; (3) the shift of habitat from natural wetlands to artificial surfaces calls for the development of eco-friendly land use mode; (4) the development of satellite tracking, remote sensing, and advanced data analysis methods have greatly facilitated movement ecology, even though they are still under development.
•A new data structure, Mino Vector (MV), is designed to serve the whole framework.•Concepts of dynamic nature (DN) and dynamic rank (DR) are defined for the first time.•A normal distribution is ...chosen as the pilot estimate for the Varying KDE.•The threshold of each pixel is adaptively controlled by each pixel's own DR.•A novel update mechanism called Tetris update scheme (TUS) is proposed.
In this paper, we propose a novel image background subtraction framework based on KDE. Firstly a new data structure called Mino Vector (MV) is designed for each pixel; we define dynamic nature (DN) for pixels of a scene and rank them in terms of DN for getting quantized results named dynamic rank (DR). Then, the varying KDE is adopted and implemented which significantly improves the estimation accuracy. Unlike using a global threshold in literature, we adaptively set a threshold for each pixel according to its DR. Inspired by the popular computer game Tetris, we present a Tetris update scheme (TUS) to update the background model in which the bottom row will be cleared, so do noises when the update condition is met. In experiments, we evaluate our framework on a well-known video dataset, CDnet 2012. Our results indicate that our framework achieves competitive results when compared with the state-of-the-art methods.
Earthquake-stricken areas are characterized by frequent and long-lasting geological disasters. Therefore, a scientific and reasonable study on the spatial prediction of post-earthquake landslide ...hazards is beneficial to the sustainable development of these areas. Firstly, the spatial effects of landslide hazards are analyzed using KDE, and the singularity index is used to delineate the area where negative samples can be selected to reduce the impact of uncertainty on the model. Secondly, eight landslide influencing factors are selected, and the weights of each influencing factor are calculated by using a geodetector. Finally, the traditional binary dependent variable approach is changed, and a KDE-MDBN landslide hazard spatial susceptibility assessment model is constructed. The research indicates that the assessment model based on KDE-MDBN can accurately reflect the characteristics of the spatial distribution of landslide hazards in the area. The results of presented in this paper can provide reference information for subsequent susceptibility assessments.
•A quantification method to delineate the selection area of negative samples is proposed.•The landslide hazard effect is considered in the construction of the landslide susceptibility assessment model.•The KDE-MDBN susceptibility assessment model can improve the spatial characteristic of landslide hazards.•The most important factors for spatial landslide hazards were determined.
We develop a method for obtaining safe initial policies for reinforcement learning via approximate dynamic programming (ADP) techniques for uncertain systems evolving with discrete-time dynamics. We ...employ the kernelized Lipschitz estimation to learn multiplier matrices that are used in semidefinite programming frameworks for computing admissible initial control policies with provably high probability. Such admissible controllers enable safe initialization and constraint enforcement while providing exponential stability of the equilibrium of the closed-loop system.