Advanced Materials Research Vols. 756-759

Abstract: In this paper, we address the angle estimation problem in linear array with some ill sensors (partially-well sensors), which only work well randomly. The output of the array will miss some values, and this can be regarded as a low-rank matrix completion problem due to the property that the number of sources is smaller than the number of the total sensors. The output of the array, which is corrupted by the missing values and the noise, can be complete via the Optspace method, and then the angles can be estimated according to the complete output. The proposed algorithm works well for the array with some ill sensors; moreover, it is suitable for non-uniform linear array. Simulation results illustrate performance of the algorithm.

Abstract: With the increasing availability and mobile application of LBS (Location-Based Services), large scale spatial objects remind challenge in cloud environments. In order to retrieve a few data items within a very large structured data set, skyline queries are utilized to optimize a single respectively multiple criteria. In this paper, we develop a new pre-clustering-based skyline queries technique to address the skewed distribution problem. We also present distributed approaches that construct grid index and process skyline queries. We evaluate the effectiveness of our algorithms with extensive experiments using real data sets. The results demonstrate the efficiency and scalability of our skyline queries algorithms based on pre-clustering.

Abstract: Remote sensing satellite images can intuitively reflect the information of the Earth's surface. The computer image processing system is of the advantages of high-precision and low-cost. It has a strong application value to study the computer processing system of remote sensing satellite image. The paper first discussed the design principles of the computer processing system and the implementation of its workflow, and then the application of the image processing system is briefly analyzed.

Abstract: In order to get further improvement of the rendering speed, and the establishment of a high precision terrain at the same time, this paper discusses the high precision terrain rendering algorithms of the specific content and characteristics based on the fractal theory. The analysis shows that: in order to draw out more realistic topography, there must be the magnitude of the terrain data as the foundation in the present terrain simulation field,. Even the terrain data is rich enough, it also have certain defective, which can cause untruthfulness of the vision. This paper can improve the rendering speed and establish a high precision terrain based on the fractal theory at the same time.

Abstract: Machine scheduling is a central task in production planning. In general it means the problem of scheduling job operations on a given number of available machines. In this paper we consider a machine scheduling problem with one machine, or the Single Machine Total Tardiness Problem. To solve this NP-hard problem, we develop an improved Tabu Search Algorithm, which is tested to have the ability to find good results by an example.

Abstract: This paper describes the basic ideas of several commonly used intelligent optimization algorithms, summarizes their essential features, classifies them, and points out the improvement directions.

Abstract: With the widely application of face recognition and the rapid development of Android OS, technique of face detection and recognition based on Android platform becomes increasingly attractive. This paper presents a real-time face recognition system on Android platform. The system realizes face detection by applying AdaBoost algorithm and face recognition by utilizing Eigenfaces. This paper also came up with some methods to speed up the face detection and recognition process and improve the correct rate of face recognition. Experimental results show that this system is able to realize real-time face detection and recognition on Android smart phones. In addition, all the work is completed on the smart phone without using any other terminals or tools.

Abstract: Anti-collision is a key technology in systems and applications in which Radio Frequency Identification (RFID) is used since it determines the efficiency of RFID tag identification. Although binary-tree search algorithms can effectively resolve collisions, they may not be very efficient in systems with a large number of RFID tags. In this paper, we propose a new flat-tree anti-collision algorithm and show that the proposed algorithm is more suitable for resolving collisions involving a massive number of RFID tags. Through analysis and experiment, we show that the flat-tree algorithm outperforms the binary-tree search algorithms when the number of RFID tags reaches a massive scale, i.e., exceeds a certain number.

Abstract: Prognostic capability is a very important character for PHM distinguishing from other diagnosis systems, and fault prognostic method based on LS-SVM was researched on in this paper. In order to solve kernel function parameters, penalty coefficient and insensitivity loss coefficient of LS-SVM, improved QPSO was put forward to train LS-SVM in this paper. And then a certain power supply combination in beam control system of a certain control and guide radar was taken as an example to collect voltage signal and forecasted, and its precision was very perfect and could satisfy prognostic demand.

Abstract: Computer vision is a diverse and relatively new field of study. Object tracking plays a crucial role as a preliminary step for high-level image processing in the field of computer vision. However, mean shift algorithm in the target tracking has some defects, such as: the application of fixed bandwidth for probability density estimation usually causes lack of smooth or too smooth; moving target often appears partial occlusion or complete occlusion due to the complexity of the background; background pixels in object model will induce localization error of object tracking, and so on. Therefore, this paper elaborates several elegant algorithms to solve some of the problems. After discussing the application of Mean shift in the field of target tracking, this paper presented an improved Mean shift algorithm by combining Mean Shift and Kalman Filter.