Refine
Document Type
Way of publication
- Open Access (1)
Language
- English (5)
Keywords
- Fog/low stratus top heights (1)
- MODIS (1)
- MTSAT (1)
- Principle component analysis (PCA) (1)
- Satellite retrieval (1)
- Sea fog (1)
- Yellow Sea (1)
- dimensionality reduction (1)
- robustness (1)
- weighted principal component analysis (WPCA) (1)
Institute
- FG IT-Sicherheit (2)
- FG Atmosphärische Prozesse (1)
- FG Mikroelektronik (1)
- JP Photovoltaik (1)
An efficiency-enchanced asynchronous buck converter with threshold compensated freewheeling diode
(2013)
Principal component analysis (PCA) is an important method for processing high-dimensional data. In recent years, PCA models based on various norms have been extensively studied to improve the robustness. However, on the one hand, these algorithms do not consider the relationship between reconstruction error and covariance; on the other hand, they lack the uncertainty of considering the principal component to the data description. Aiming at these problems, this paper proposes a new robust PCA algorithm. Firstly, the L2,p-norm is used to measure the reconstruction error and the description variance of the projection data. Based on the reconstruction error and the description variance, the adaptive probability error minimization model is established to calculate the uncertainty of the principal component's description of the data. Based on the uncertainty, the adaptive probability weighting PCA is established. The corresponding optimization method is designed. The experimental results of artificial data sets, UCI data sets and face databases show that RPCA-PW is superior than other PCA algorithms.
Nowadays the mobile phone has become an indispensable tool in the lives of many people. While facilitating people's lives, it also provides criminals with a very important tool for spreading the terrorist video. Traditional manual detection of the terrorist video has the problem of low accuracy and inefficiency. To address the issue, this paper proposes a terrorist video detection system based on Light Gradient Boosting Machine (LightGBM) and Faster Region-based Convolutional Neural Network (Faster R-CNN) for mobile phone forensics system, which is used to quickly detect whether there is a terrorist video in the suspect's mobile phone. The system uses a multi-model method for detection, which includes preliminary detection and deep detection in two stages. Experimental research shows that it can effectively and accurately detect terrorist videos in mobile phones, thereby helping criminal investigation personnel to quickly grasp criminal evidence and provide some clues for the detection of the case.
An accurate cloud top retrieval from geostationary (GEO) and low earth orbit (LEO) platforms is still a pending problem. This particularly holds for low level clouds. Furthermore, cloud top height is a crucial parameter to calculate cloud immersion of underlying terrain from GEO/LEO data and thus, for the discrimination between low level stratus and ground fog, where the latter is a main obstruction for air, land and sea traffic. All problems are particularly evident for ocean areas such as the Yellow Sea where no ground observations are available. In this paper, a novel method is presented to retrieve low stratus/fog top heights with special reference to the Yellow Sea and its surroundings, based on GEO data of MTSAT-1 and MTSAT-2 (JAMI sensor) and LEO data (MODIS sensor on Terra and Aqua) using the infrared (IR) water vapor and split-window bands. Two cases with very good data coverage are discussed where the retrieved low stratus/fog heights are compared to CALIPSO cloud top heights, and simulated data using the mesoscale model WRF. The comparison of JAMI retrievals with the spatial data sources used shows an encouraging accuracy (root-mean-square error, RMSE, around 300 m) in comparison to other retrieval schemes base on IR data hitherto published. A validation of the retrievals for the position of two radiosonde stations using available sounding data of seven foggy days revealed an even better performance with an average deviation of 184 m (standard deviation of 132 m). However, the validation revealed that the application of the underlying equations to retrieve inversion strength and thickness under foggy conditions would need some adjustments because the equations taken from the work of Liu and Key (2003) were originally developed for clear sky situations. Thus, the adaptation of the original scheme during future work should especially address cloudy conditions under moderate inversion strengths which could lead to an improvement of the retrieval accuracy.