Inspired by theories of higher local order autocorrelation (HLAC), this paper presents a simple, novel, yet very powerful approach for wood recognition. The method is suitable for wood database applications, which are of great importance in wood related industries and administrations. At the feature extraction stage, a set of features is extracted from Mask Matching Image (MMI). The MMI features preserve the mask matching information gathered from the HLAC methods. The texture information in the image can then be accurately extracted from the statistical and geometrical features. In particular, richer information and enhanced discriminative power is achieved through the length histogram, a new histogram that embodies the width and height histograms. The performance of the proposed approach is compared to the state-of-the-art HLAC approaches using the wood stereogram dataset ZAFU WS 24. By conducting extensive experiments on ZAFU WS 24, we show that our approach significantly improves the classification accuracy.

With the fast development of new array technology and intelligent antenna, it is easier to obtain angle of arrival (AOA) measurements. Hybrid received signal strength (RSS) and AOA measurement techniques are proposed for the position computing in sensor networks. By converting the measurement equations and relaxing the optimization function, range-based square semidefinite programming (RLS-SDP) and squared range-based square semidefinite programming (SRLS-SDP) algorithms are put forward to obtain the source position estimate by considering the transmit power to be known or unknown. The proposed RLS-SDP and SRLS-SDP algorithms provide accurate solution to the source position estimate and avoid the initialization process of numerical calculation. The simulations show that the proposed RLS-SDP and SRLS-SDP algorithms perform better than the linear estimator and provide the accuracy performance which is very close to the Cramér-Rao Lower Bound (CRLB) of position estimation. The proposed SRLS-SDP algorithm shows its advantages in the computational complexity compared with the RLS-SDP, since the complexity of SRLS-SDP is independent of the number of anchor nodes. INDEX TERMS Localization, received signal strength, angle of arrival, semidefinite programming.

No abstract

Pore feature is important for hardwood identification. But it's difficult to segment pores from wood cross-section images since pore, fiber and longitudinal parenchyma in the image are similar in shapes but different only in size, and the different hardwood species varies in the size of pores. In order to segment pores automatically without parameters set manually, it is necessary to design an adaptive algorithm which may be applied for all kinds of hardwood cross-section images. In the paper, an adaptive method is proposed to evaluate the optimal threshold of closed region area for pore segmentation. The method sorts all closed regions according to the area and classifies closed regions into two classes with maximum between-class variance method. We implements the method based on genetic algorithm to overcome the drawback of being time-consuming. Experiment on images of hardwood species shows that the threshold obtained by the genetic algorithm is very close to but more efficient than the ordinary enumeration algorithm. Moreover, with the obtained threshold majority of pores can be extracted except for some very small ones.

The paper analyzes the key concepts, theories and methods of machine learning ensemble, and reviews the related studies on support vector machine (SVM) ensemble. The experiments on the remote sensing classification show that SVM ensemble is more accurate than single SVM. To obtain an effective SVM ensemble, we propose a selective SVM ensemble approach based on fuzzy clustering and discuss the issues on it.

Acceleration sensor is extensively used in the field of human activity recognition, since it provides better recognition rate of human activity. Based on the principle of molecular attribute, a simple and adaptive activity recognition method is proposed using the acceleration data flow, which constitutes a serial activity, when the acceleration data are treated as the material flow with certain molecular structure. Then five molecular attributes including relative molecular mass, density, internal forces in a molecule, molecule stability, and attraction between molecules are introduced to recognize six human activities, since the closer molecular attribute means the more similar activity. Based on the calculated molecular attributes, a reliability-based voting method for human activity recognition is developed. Since each activity has respective motion cycle, a sliding window with variable sizes is put forward to enhance the recognition rate. Furthermore, adaptive incremental learning is designed to adapt to the different users. The long-time experimental results show that the proposed method is rather accurate and robust for different crowds. The average recognition rate achieves 97.2% for six human activities including walking, jogging, running, going upstairs, going downstairs, and sitting down.

No abstract

Wood identification is a foundation of wood science and industry. The arrangement-of-pore(AOP) of wood cross-section is a important wood identification feature. The paper applies mathematical morphology methods such as dilation and erosion, open and close transform to wood cross-section image repairing, noise filtering and edge detection, thereby obtaining AOP feature. The study is a beneficial attempt to intelligent wood identification based on image processing technology.