Earticle

Flower classification is a useful way for grouping a flower in certain class using specific features. This research propose a new method of flower classification system using combination of color and texture features. The first phase is getting the crown of the flower, which is localized from a flower image by using pillbox filtering and OTSU’s thresholding. In the next phase, color and texture features are extracted from the crown. The color features are extracted by removing L channel in L*a*b* color space, and taking only a* and b* channel, because of ignoring different lighting condition in flower image. The texture features are extracted by Segmentation-based Fractal Texture Analysis (SFTA). The combination features which are consisted of 10 color features and 48 texture features are used as input in k-Nearest Neighbor (kNN) classifier method with cosine distance. The flower classification achieves the best result with accuracy 73.63%.

To group the agents from various areas for the solution modeling of large-scale, sophisticated systems or issues, we has developed a distributed modeling methodology and its networked supporting platform [1]. The upcoming problems are, however, how to coordinate (organize, supervise, evaluate) such distributed modeling agents, and how to aggregate a number of modeling templates for the best solution(s). The Soft-Agents system is designed to perform coordinating user-agents teams. Such coordination enables judgment on working characteristics and modeling quality of each team and individual separately. To work out the best solution(s), the individual template is aggregated by using Analytic-Hierarchy-Process and multiple templates are aggregated by the Ordered-Weighted-Geometric algorithm.

To effectively diagnose the deterministic faults of a LPRE ground-testing bed, the fault diagnosis method based on PCA and SOM is proposed. The dimension reduction process of PCA not only reduces data size, but also reduces noise influence. It also implements a visualization of fault status identification and fault variable orientation by SOM. Simulation and real fault data results indicate that the efficiency and identification ratio of PCA-SOM method are better than SOM method.

Nowadays, the traditional parameters recognition method cannot match the requirements of the increasing new modulation radar signals. In order to solve this problem, in this paper, it proposes the improved radar signal recognition algorithm based on fractal theory and Neural Network theory. Taking the advantage of the characteristics of relevant dimension which will be able to measure the relevant complex degree of the radial source signals, we extract the relevant point as the input of neutral network in order to recognize and classify the signals. Simulation results show that, this algorithm has a distinguish effect on classification under low SNR, which is suitable for the feature extraction and recognition of various styles of radar signals.