• 대한전자공학회논문지SD
• /
• 제44권3호
• /
• pp.12-19
• /
• 2007

본 논문에서는 피사체 인식 및 자동 선택이 가능한 모바일 폰 용 자동초점 알고리즘을 제안하였다. 제안된 알고리즘은 피사체 인식 단계와 목적물 자동 선택 단계로 구성된다. 피사체 인식 단계에서는 화면 전체에 배치된 다중 소형 측거점과 초점값 연산자를 사용하여 복수개의 피사체를 모바일 폰에 내장된 이미지 센서만을 사용하여 인식함으로써 기존의 적외선이나 초음파와 같은 외부 장치를 사용하는 방식과 달리 모바일 폰에서의 피사체 인식을 가능케 하고자 하였다 목적물 자동 선택 단계에서는 퍼지 멤버십 변수와 퍼지 추론을 통해 사용자가 촬영하고자 하는 목적물을 자동 선택하는데 이는 사진 기술이 없는 사용자라도 선명한 화질의 디지털 이미지를 획득할 수 있도록 하기 위함이다. 제안된 알고리즘은 프로그램 언어로 구현되었으며, 초점 거리 제어가 가능한 CCD 카메라와 PC를 사용하여 실시간으로 이미지를 분석, 검증하였다.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2000년도 춘계학술대회논문집
• /
• pp.894-899
• /
• 2000

Optimization of 26 design variables selected from suspension characteristics for Korean High Speed Train (KHST) is performed according to the minimization of 58 responses which represent running safety and ride comfort for KHST and analyzed by using the each response surface model from stochastic design experiments. Sensitivity of design variables is also analyzed through the response surface model which ineffective design prameters to the performance index are screened by using stepwise regression method. The response surface models are used for optimizing design variables through simplex algorism. Values of performance index simulated by optimized design parameters are totally lower than those by initial design parameters. It shows that this method is effective for optimizing multi-design variables to multi-object function.

• 조명전기설비학회논문지
• /
• 제18권1호
• /
• pp.137-145
• /
• 2004

본 논문은 평행식 스테레오 영상감시 시스템을 이용하여 3차원 공간에서 다중 이동물체를 검출하고, 카메라에서 이동 물체까지의 거리를 측정하는 알고리즘을 제안하였다. 스테레오 영상감시 시스템의 좌, 우측 영상을 입력을 받아 적응형 임계값과 화소귀납 알고리즘(PRA, pixel recursive algorithm)을 사용하여 다중 이동물체 영역을 추출하였다. 그리고 각각의 물체영역을 윈도우 마스크로 검출하고, 다중 이동물체 각각의 위치좌표와 스테레오 시차를 구하였다. 이 시차와 스테레오 비전 시스템의 특성 및 삼각 함수를 이용하여 다중 이동물체의 거리를 측정하였다. 실험결과 거리측정 오차도 7.28％이내에 존재하였으며, 따라서 제안한 알고리즘을 이용하여 시스템을 구현할 경우 스테레오 방범 시스템, 자율 이동로봇 및 스테레오 원격제어 시스템 등에 응용될 수 있을 것이다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2005년도 춘계학술대회 논문집
• /
• pp.163-168
• /
• 2005

It is one of the most important things to determinate optimized cutting conditions which satisfy productivity and cost simultaneously in production and CAPP systems. These days many researchers have figured out the optimizing way for solutions of multi-object function to find the approach methods using algorithm such as genetic algorithm or tabu search, etc., instead of mathematical methods. The main creation of objective function is proposed by empirical method but which is difficult to set it up and to analysis. In this paper, an optimization method of cutting condition is shown using the ANN and GA for the multi-objective function in high speed machining.

• Journal of Mechanical Science and Technology
• /
• 제20권3호
• /
• pp.366-375
• /
• 2006

In this paper, it is intended to introduce a method to solve multi-objective optimization problems and to evaluate its performance. In order to verify the performance of this method it is applied for a vertical roller mill for Portland cement. A design process is defined with the compromise decision support problem concept and a design process consists of two steps: the design of experiments and mathematical programming. In this process, a designer decides an object that the objective function is going to pursuit and a non-linear optimization is performed composing objective constraints with practical constraints. In this method, response surfaces are used to model objectives (stress, deflection and weight) and the optimization is performed for each of the objectives while handling the remaining ones as constraints. The response surfaces are constructed using orthogonal polynomials, and orthogonal array as design of experiment, with analysis of variance for variable selection. In addition, it establishes the relative influence of the design variables in the objectives variability. The constrained optimization problems are solved using sequential quadratic programming. From the results, it is found that the method in this paper is a very effective and powerful for the multi-objective optimization of various practical design problems. It provides, moreover, a reference of design to judge the amount of excess or shortage from the final object.

• International Journal of Precision Engineering and Manufacturing
• /
• 제4권2호
• /
• pp.64-70
• /
• 2003

A new method using a transfer function is introduced in the present paper for analyzing the motion errors of hydrostatic bearing tables. The relationship between film reaction force in a single-side hydrostatic pad and the form error of guide rail is derived at various spatial frequencies by finite element analysis, and it is expressed as a transfer function. This transfer function clarifies so called 'the averaging effect of an oil film' quantitively. It is found that the amplitude of film force is reduced as the spatial frequency increases or the relative width of the pocket is reduced. The motion errors of a multi pad type table are estimated using a transfer function, the form errors of a guide rail and the geometric relationship between the pads. The method is named as the Transfer Function Method (TFM). The motion errors calculated by the TFM show good agreement with the motion errors calculated by the Multi Pad Method considering the entire table as an analysis object. From the results, it is confirmed that the proposed TFM is very effective to analyze the motion errors of hydrostatic tables.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제6권4호
• /
• pp.1063-1075
• /
• 2012

To locate an object accurately in the wireless sensor networks, the distance measure based on time-delay plays an important role. In this paper, we propose a maximum likelihood (ML) time-delay estimation algorithm in multi-path wireless propagation channel. We get the joint probability density function after sampling the frequency domain response of the multi-path channel, which could be obtained by the vector network analyzer. Based on the ML criterion, the time-delay values of different paths are estimated. Considering the ML function is non-linear with respect to the multi-path time-delays, we first obtain the coarse values of different paths using the subspace fitting algorithm, then take them as an initial point, and finally get the ML time-delay estimation values with the pattern searching optimization method. The simulation results show that although the ML estimation variance could not reach the Cramer-Rao lower bounds (CRLB), its performance is superior to that of subspace fitting algorithm, and could be seen as a fine algorithm.

• Journal of Advanced Marine Engineering and Technology
• /
• 제11권1호
• /
• pp.39-52
• /
• 1987

Recently, the design problem of heat insulation have been reappraised in the aspect of energy saving due to the rising trend of energy cost. For example, that design problem is increasingly requsted in the fields of accommodation air conditioning systems, hot water supply systems, cargo handling systems, district heating or cooling systems. The rational design of heat insulation of piping systems can not only improve the overall efficiency of energy transfer but also give energy saving. In this paper, the heat insulation problem of district heating systems is therefore modeled as the multi-stage decision processes, suitable for dynamic programming technique. And take the object function as the sum of heat insulation material cost involved construction cost and heat loss cost, and propose the design method to minimize the object function for overall piping systems by dynamic programing. Effectiveness of design method presented here is proved by a computer simulation.

• 한국컴퓨터정보학회논문지
• /
• 제16권6호
• /
• pp.89-97
• /
• 2011

현재의 객체 추적과 검색의 과정을 보면 고정된 단일 카메라를 통해 입력받은 영상에서 객체를 추출하여 추적하고, 추적된 객체의 구체적인 정보를 알기 위해 줌(Zoom) 기능으로 객체를 인식하는 과정이었다. 본 논문은 다중카메라를 이용하여 객체를 추적하고, 인식하는 것으로써 추적된 객체에 대해 검색인식가능한 영역에 대한 정보를 증강현실로 나타내는 시스템을 제안한다. 제안 시스템의 실험결과를 보면 연산에 포함된 픽셀 수가 현저히 줄어들고, 객체의 인식률이 향상되고, 정보 확인 시간도 단축되었다. 그리고 기존 방법과 비교하여 객체의 움직임을 검출하는 정확성은 물론 움직임 검출에 소요되는 시간도 단축되어 개선된 성능을 보였다.

• 한국멀티미디어학회논문지
• /
• 제24권3호
• /
• pp.416-422
• /
• 2021

Due to the large differences in human activity within classes, the large similarity between classes, and the problems of visual angle and occlusion, it is difficult to extract features manually, and the detection rate of human behavior is low. In order to better solve these problems, an improved Faster R-CNN-based detection algorithm is proposed in this paper. It achieves multi-object recognition and localization through a second-order detection network, and replaces the original feature extraction module with Dense-Net, which can fuse multi-level feature information, increase network depth and avoid disappearance of network gradients. Meanwhile, the proposal merging strategy is improved with Soft-NMS, where an attenuation function is designed to replace the conventional NMS algorithm, thereby avoiding missed detection of adjacent or overlapping objects, and enhancing the network detection accuracy under multiple objects. During the experiment, the improved Faster R-CNN method in this article has 84.7% target detection result, which is improved compared to other methods, which proves that the target recognition method has significant advantages and potential.