• 토지주택연구
• /
• 제14권3호
• /
• pp.145-155
• /
• 2023

최근 건설산업 분야에 BIM 기술의 활용이 보편화되면서 3D 모델과 실제 시공 부위의 오류 확인 등을 위해 다양한 객체 탐지 알고리즘들이 활용되고 있다. 객체 탐지 기술은 건축물, 교량, 터널 등 건설시설물의 종류에 따라 객체 특성이 상이하므로 객체 탐지 기술도 적절한 방법을 사용할 필요가 있다. 또한 객체 탐지를 위해서는 초기 객체 이미지가 있어야 하며 이를 위해서도 드론, 스마트폰 등 다양한 방법으로 이미지 취득이 가능하다. 본 연구에서는 철도와 도로 시설의 터널 부위에 대하여 초기 이미지 구축을 위해 터널 내부 촬영에 최적화된 360° 카메라를 이용하여 이미지를 촬영하고, 촬영된 이미지로부터 실제 객체를 탐지하기 위한 객체 탐지 방법론으로 YOLO 알고리즘, SSD 알고리즘 및 R-CNN 알고리즘을 적용하여 방법론별 객체 탐지의 정확도를 비교 분석한다. 분석 결과 Faster R-CNN 알고리즘이 SSD, YOLO v5 알고리즘에 비해 높은 인식률 및 mAP 값을 가졌으며 인식률들의 최소·최대 값의 차이가 작아 균등한 검측 능력을 나타냈다. 이러한 연구는 철도와 도로 시설공사에 BIM 적용이 확산되고 있는 점을 고려하면 360° 카메라의 활용 방법 확대와 유지보수를 위한 터널 시설 부위의 객체 탐지 방법론 적용에 활용될 수 있다.

• 정보처리학회논문지B
• /
• 제17B권6호
• /
• pp.445-458
• /
• 2010

LCD 모니터의 백라이트로 CCFL 형광체를 많이 사용하고 있으나 그 불량여부는 육안에 의존하고 있다. 육안 검사를 함으로써 부품에 대한 일관성 있는 검사가 결여되고, 노동집약적인 검사로 인해 산업적 재해가 발생할 수 있다. 따라서, CCFL 불량유무를 자동으로 판별하기 위해서 물리적 촬영 환경과 영상처리 알고리즘은 중요하다. 본 논문에서는 CCFL 형광체를 자동으로 검사하기 위한 촬영환경 중 다섯 가지 조건과 세 가지조건 중 두 조건모두에서 사용되는 측면 촬영환경에서 획득한 영상을 이용하여 불량을 판별하기 위한 알고리즘을 제시하였다. 불량을 포함한 CCFL 형광체와 정상시료를 사용하여 영상 획득 및 실험을 수행하였고, 그 결과 제안한 촬영환경과 알고리즘은 과검율 4.65 %와 유출률 5.37 %의 성능을 보인다.

• 한국멀티미디어학회논문지
• /
• 제22권12호
• /
• pp.1447-1456
• /
• 2019

As artificial intelligence technology is actively used in image processing, it is possible to generate high-quality fake images based on deep learning. Fake images generated using GAN(Generative Adversarial Network), one of unsupervised learning algorithms, have reached levels that are hard to discriminate from the naked eye. Detecting these fake images is required as they can be abused for crimes such as illegal content production, identity fraud and defamation. In this paper, we develop a deep-learning model based on CNN(Convolutional Neural Network) for the detection of StyleGAN fake images. StyleGAN is one of GAN algorithms and has an excellent performance in generating face images. We experiment with 48 number of experimental scenarios developed by combining parameters of the proposed model. We train and test each scenario with 300,000 number of real and fake face images in order to present a model parameter that improves performance in the detection of fake faces.

• Smart Structures and Systems
• /
• 제26권1호
• /
• pp.23-33
• /
• 2020

Majority of the damage in engineering structures is nonlinear. Damage sensitive features (DSFs) extracted by traditional methods from linear time series models cannot effectively handle nonlinearity induced by structural damage. A new DSF is proposed based on vector space cosine similarity (VSCS), which combines K-means cluster analysis and Bayesian discrimination to detect nonlinear structural damage. A reference autoregressive moving average (ARMA) model is built based on measured acceleration data. This study first considers an existing DSF, residual standard deviation (RSD). The DSF is further advanced using the VSCS, and then the advanced VSCS is classified using K-means cluster analysis and Bayes discriminant analysis, respectively. The performance of the proposed approach is then verified using experimental data from a three-story shear building structure, and compared with the results of existing RSD. It is demonstrated that combining the linear ARMA model and the advanced VSCS, with cluster analysis and Bayes discriminant analysis, respectively, is an effective approach for detection of nonlinear damage. This approach improves the reliability and accuracy of the nonlinear damage detection using the linear model and significantly reduces the computational cost. The results indicate that the proposed approach is potential to be a promising damage detection technique.

• 제어로봇시스템학회논문지
• /
• 제18권1호
• /
• pp.35-41
• /
• 2012

This paper proposes a moving object detection algorithm for active camera system that can be applied to mobile robot and intelligent surveillance system. Most of moving object detection algorithms based on a stationary camera system. These algorithms used fixed surveillance system that does not consider the motion of the background or robot tracking system that track pre-learned object. Unlike the stationary camera system, the active camera system has a problem that is difficult to extract the moving object due to the error occurred by the movement of camera. In order to overcome this problem, the motion of the camera was compensated by using SURF and Pseudo Perspective model, and then the moving object is extracted efficiently using stochastic Label Cluster transport model. This method is possible to detect moving object because that minimizes effect of the background movement. Our approach proves robust and effective in terms of moving object detection in active camera system.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2013년도 추계학술대회
• /
• pp.165-168
• /
• 2013

Image distortions, such as quantization errors, can have a severe negative impact on the performance of computer vision algorithms, and, more specifically, on object detection algorithms. State-of-the-art implementations of the JPEG-2000 image coder commonly allocate the available bits to minimize the Mean-Squared-Error (MSE) distortion between the original image and the resulting compressed image. However, considering that some state-of-the-art object detection methods use the gradient information as the main image feature, an improved object detection performance is expected for JPEG-2000 image coders that allocate the available bits to minimize the distortions on the gradient content. Accordingly, in this work, the Gradient Mean-Squared-Error (GMSE) based JPEG-2000 coder presents an improved object detection performance over the MSE based JPEG-2000 image coder when the object of interest is located at the same spatial location of the image regions with the strongest gradients and also for high bit-rates. For low bit-rates (e.g. 0.07bpp), the GMSE based JPEG-2000 image coder becomes overly selective in choosing the gradients to preserve, and, as a result, there is a greater chance of mismatch between the spatial locations of the gradients that the coder is trying to preserve and the spatial locations of the objects of interest.

• 로봇학회논문지
• /
• 제9권2호
• /
• pp.111-116
• /
• 2014

Human-robot co-operation becomes increasingly frequent due to the widespread use of service robots. However, during such co-operation, robots have a high chance of colliding with humans, which may result in serious injury. Thus, many solutions were proposed to ensure collision safety, and among them, collision detection algorithms are regarded as one of the most practical solutions. They allow a robot to quickly detect a collision so that the robot can perform a proper reaction to minimize the impact. However, conventional collision detection algorithms required the precise model of a robot, which is difficult to obtain and is subjected to change. Also, expensive sensors, such as torque sensors, are often required. In this study, we propose a novel collision detection algorithm which only requires motor encoders. It detects collisions by monitoring the high-pass filtered version of the velocity error. The proposed algorithm can be easily implemented to any robots, and its performance was verified through various tests.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2017년도 추계학술발표대회
• /
• pp.891-894
• /
• 2017

Lane detection algorithms became a key factor of advance driver assistance system (ADAS), since the rapidly increasing of high-technology in vehicles. However, one common problem of these algorithms is their performance's instability under various illumination conditions. We recognize a feasible complementation between image processing and color science to address the problem of lane marks detection on the road with different lighting conditions. We proposed a novel lane detection algorithm using the attributes of a uniform color space such as $CIEL^*C^*h$ with the implementation of image processing techniques, that lead to positive results. We applied at the final stage Clustering to make more accurate our lane mark estimation. The experimental results show the effectiveness of our method with detection rate of 91.80%. Moreover, the algorithm performs satisfactory with changes in illumination due to our process with lightness ($L^*$) and the color's property on $CIEL^*C^*h$.

• Journal of Communications and Networks
• /
• 제5권1호
• /
• pp.33-42
• /
• 2003

In this paper a two-stage maximum-likelihood (ML) detection structure for group detection in DS/CDMA systems is presented. The first stage of the receiver is a linear filter, aimed at suppressing the effect of the unwanted (i.e., out-of-grout) users' signals, while the second stage is a non-linear block, implementing a ML detection rule on the set of desired users signals. As to the linear stage, we consider both the decorrelating and the minimum mean square error approaches. Interestingly, the proposed detection structure turns out to be a generalization of Varanasi's group detector, to which it reduces when the system is synchronous, the signatures are linerly independent and the first stage of the receiver is a decorrelator. The issue of blind adaptive receiver implementation is also considered, and implementations of the proposed receiver based on the LMS algorithm, the RLS algorithm and subspace-tracking algorithms are presented. These adaptive receivers do not rely on any knowledge on the out-of group users' signals, and are thus particularly suited for rejection of out-of-cell interference in the base station. Simulation results confirm that the proposed structure achieves very satisfactory performance in comparison with previously derived receivers, as well as that the proposed blind adaptive algorithms achieve satisfactory performance.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2017년도 추계학술대회
• /
• pp.687-690
• /
• 2017

에지 검출은 차선 인식, 물체 및 패턴 검출 등의 성능을 결정하는 중요한 단계이며, 현재까지도 이를 위한 많은 연구가 이루어지고 있다. 지금까지 널리 알려져 있는 에지 검출 알고리즘은 Sobel, Prewitt, Roberts, Canny 에지 검출 알고리즘 등이 있으며, 이러한 알고리즘들은 밝기값의 변화가 완만한 영상을 처리할 때, 에지가 아닌 영역으로 판단할 경우가 많다. 따라서 본 논문에서는 마스크 영영에서 침식, 팽창, 열기, 닫기 등을 활용하는 그레이 레벨 모폴로지를 이용한 에지검출 알고리즘을 제안하였다.