• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.07a
• /
• pp.195-197
• /
• 2010

본 논문은 기존 평행좌표를 이용하는 얼굴 영상 대신 극좌표계 변환을 이용한 얼굴 영상을 이용하여 회전에 강인한 얼굴인식 방법을 제안한다. 극좌표계 변환 방법은 얼굴의 중심부분의 한 점을 극으로 삼아 이 점을 기준으로 360도 각 방향으로 일정 길이만큼 얼굴 영상을 샘플링 하여 새로운 얼굴 영상을 제작하는 방법이다. 이 극좌표계 변환 방법을 이용해 재구성된 영상에 대해 회귀( regression )문제 해결을 위해 변형된 LDA인 LDAr(LDA for regression)을 이용하여 얼굴의 중심부분의한 점인 극을 중심으로 임의의 각도로 회전된 영상의 회전 정도를 추정하여 이를 정규화 시키는 방법을 통해 얼굴 인식의 인식률을 향상시키고자 한다. LDAr은 LDA의 기본개념인 각 클래스 간 떨어진 정도를 최대화하는 것이 목적으로 클래스간 분산과 클래스내 분산의 비율을 최대화 하는 방법을 응용하여 이를 회귀문제에 적용할 수 있게 변형을 한 것이다. 즉, LDAr은 목표값(target)의 차이가 큰 샘플들과 목표값의 차이가 작은 샘플들 간의 거리의 비율을 최대화 하는 것을 목적으로 하게 된다. 제안된 방법을 Yale데이터에 적용하여 임의의 각도로 회전시킨 영상에 대해 회전 각도를 정확히 찾아내는 것을 확인할 수 있었다.

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 2007.10a
• /
• pp.324-326
• /
• 2007

• Journal of the Korean Institute of Gas
• /
• v.21 no.5
• /
• pp.1-8
• /
• 2017

Currently gas safety management in the industrial field has been done by LDAR as contact method or methane leak detector as non-contact method. But LDAR method requires a lot of man-power and methane leak detector have the limitation of methane only. Therefore the Research on the OGI(optical gas image) has big attention by industry. This research was undertaken to see the effect of background temperature difference of gas cloud on the clarity of OGI. The background temperature control panel was constructed to cool down the background temperature. OGI was taken at the various methane gas ejection rate and the designed temperature difference. The experimental results showed that the OGI(when the temperature difference is $-6^{\circ}C$) is more clear thane the OGI(when the temperature difference is zero). To quantify the clarity difference, MATLAB's RGB analysis method was employed. The RGB value of the OGI at ${\Delta}T-6^{\circ}C$ was 20% lower than the OGI at ${\Delta}T0^{\circ}C$. The clarity difference by T difference can be explained by the total radiation law. When the background temperature of the gas is lower than the air temperature, the radiation energy coming into the OGI lens is increasing. As the energy is increasing, the OGI image becomes clear.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.11a
• /
• pp.86-88
• /
• 2010

본 논문에서는 회귀문제를 위한 비선형 특징 추출방법을 제안하고 분류문제에 적용한다. 이 방법은 이미 제안된 선형판별 분석법을 회귀문제에 적용한 회귀선형판별분석법(Linear Discriminant Analysis for regression:LDAr)을 비선형 문제에 대해 확장한 것이다. 본 논문에서는 이를 위해 커널함수를 이용하여 비선형 문제로 확장하였다. 기본적인 아이디어는 입력 특징 공간을 커널 함수를 이용하여 새로운 고차원의 특징 공간으로 확장을 한 후, 샘플 간의 거리가 큰 것과 작은 것의 비율을 최대화하는 것이다. 일반적으로 얼굴 인식과 같은 응용 분야에서 얼굴의 크기, 회전과 같은 것들은 회귀문제에 있어서 비선형적이며 복잡한 문제로 인식되고 있다. 본 논문에서는 회귀 문제에 대한 간단한 실험을 수행하였으며 회귀선형판별분석법(LDAr)을 이용한 결과보다 향상된 결과를 얻을 수 있었다.

• Journal of the Korean Institute of Gas
• /
• v.26 no.6
• /
• pp.45-51
• /
• 2022

Technologies for collecting and treating pollutants from point sources are steadily being developed, but Non-point sources, it is difficult to develop emission treatment technologies and effective emission coefficients. However, since non-point sources make up about 60% of domestic emissions, and first of all, the method of calculating emissions should be reasonable, and the workplace should develop emission reduction technologies based on this. This study suggest the effectiveness and improvement of the emission coefficient currently used for the petrochemical industry with high emissions. The emission characteristics of non-point sources emission were confirmed by analyzing the LDAR (Leak Detection And Repair) data of OO company located in Yeosu, Jeollanam-do over the past five years. As a result, there was no difference in discharge characteristics according to fluid phase, but it was confirmed that there was a difference in the size of the device and the characteristics of each manufacturer. In addition, it was confirmed that the emission coefficient applied in the petrochemical industry was larger than that of the refining industry, and improvement measures were suggested. Through these studies, it is expected that emission coefficients specialized in the petrochemical industry can be applied and that the workplace itself will contribute to the development of technologies that can drastically reduce them.

• Journal of Korean Society for Atmospheric Environment
• /
• v.24 no.2
• /
• pp.229-237
• /
• 2008

Hazardous air pollutants emitted from the oil refinery plant were surveyed from 1993 US and 2005 Korean TRI data. Toluene, xylene, methanol, MTBE and n-hexane, relatively large in amount of fugitive emission, are considered as candidates of newly designated HAPs in Korea. The sealed oil pump, one of equipments among fugitive sources in the crude oil distillation tower was examined for the estimation of amounts of annual HAPs emissions according to several calculation methods using registered emission factors. Emission rates showed to be decreased with following calculation factors; average emission factor>pegged emission factor>concentration emission factor>correlation equation. Annual emission amounts of benzene, toluene, xylene. ethyl benzene and nhexane from the distillation tower were estimated and amounts of these HAPs calculated with TVOC concentrations obtained from LDAR program and correlation equations showed only 6% of those from using concentration factors.

• Journal of the Korean Institute of Gas
• /
• v.7 no.2 s.19
• /
• pp.7-13
• /
• 2003

Abstract This paper presents new emission mechanism and emission estimation model in volatile organic compounds(VOCs) emission sources. Also classifies applicable emission reduction techniques and presents new economical evaluation method for each techniques. We ultimately developed VEER(VOCs Emission Estimation and Reduction) software, which is backed by above mentioned model, emission source DB, Chemical properties DB, meteorological DB, and emission factor DB. With VEER, users in enterprise, central government and local self-governing body can get reliable emission results easily, and choose suitable emission reduction techniques.