Search

Aiming at the target detection technique is time-consuming, noise and other issues, improved the application of FSVM in target detection, a piecewise fuzzy membership function is proposed based on LDA algorithm. The algorithm projected the original high-dimensional samples onto a one-dimensional space by LDA ,the original sample’s fuzzy weights is segment calculated based on the distribution of the projection point, reduce the impact of noise and outliers on classification results. In the simulation experiments, this method can effectively reduce the impact of unbalanced data to FSVM, improve the classification accuracy.

In this paper, on the basis of the classical Retinex enhancement algorithm, we propose a MSR (Multi-Scale Retinex) algorithm based on HSI color space for foggy images enhancement. This algorithm can self-adaptively choose the Scale of the filter function, and it can automatically merge the enhancement effect of the multi-scales. The algorithm can highlight the detail characteristics of target while maintaining the background smoothly. In order to improve the brightness and contrast of the image, we use the Weber's law and the characteristics of human visual system to select the scale parameters, which can better realized the foggy images enhancement.

According to the characteristics of the incipient fire flame, combined with the color characteristic, static characteristic and dynamic characteristic, design identification model of Dynamic Fuzzy Logic (DFL), to determine the early fire quickly. First eliminate the least likely fire target, through the flame color characteristic. Then distinguish the fire target and interference, through the static characteristics, flame sharp and flame circularity. At last recognize the fire and secure flame, through the dynamic characteristics, flicker frequency and change rate of area. Through the DFL Model, identify whether it is the incipient fire. The results of the experiments show that this method can identify the early fire accurately, with stronger anti-jamming.

Interactive visual simulation is an important tool in virtual maintenance of aviation equipment. This paper uses the virtual maintenance training of aircraft landing gear as the study object, applying parametric feature modeling technique in UG and in accordance with design requirements the 3D model of aircraft landing gear was established, meanwhile, using the cubic spline interpolation to solve the difficult problem of UG modeling. This paper focuses on using interactive three-dimension visual simulation methods and based on the agent model and its auxiliary model to realize assembly/disassembly process of landing gear. We propose a real-time method is using texture mapping and light render techniques to producing realistic animation. Based on the evaluation were carried out, since the animation is intuitionist, lifelike and didactic in character, it can be used to support the training of maintenance man and reflect the process accurately.

The existing background modelings are mostly color vision characteristics modeling based on single pixel, which are easily influenced by light shadow, weather and noise, and can easily cause foreground apertures and false alarm discrete noise. This paper presents the background modeling based on multiscale Gauss parameters against deficiencies. the experimental results shows that it can efficiently solve the problem of cavity and false alarm discrete noise.

The paper presents a motor bearing fault diagnosis method based on MSICA (Multi-scale Independent Principal Component Analysis) and LSSVM (Least Squares Support Vector Machine). MSICA is introduced into motor fault diagnosis. First, wavelet decomposition is used, and then ICA models are built by wavelet coefficients in each scale, which detect fault, and finally LSSVM is used to classify fault type. Conclusions are obtained from the analysis of the experimental data provided by Case Western Reserve University’s Bearing Data Website. And it indicates that the proposed method is simple and effective.