• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.05a
• /
• pp.796-799
• /
• 2008

최근 유비쿼터스 컴퓨팅과 유비쿼터스 네트워크를 활용하여 새로운 서비스들을 개발하려는 노력이 진행 중이며, 이와 관련된 기술의 중요성도 급증하고 있다. 특히 감시정찰 센서네트워크의 핵심 구성요소인 저가의 경량 센서노드에서 측정한 미가공 데이터(raw data)를 사용하여 침입 물체의 실시간 탐지, 식별, 추적 및 예측하기 위한 디지털 신호처리 기술은 주요 기술 중 하나이다. 본 논문에서는 감시정찰 센서네트워크의 핵심 구성요소인 소리센서 노드에서 측정한 소리 미가공 데이터를 사용하여 차량을 탐지할 수 있는 소리센서 디지털 신호처리 알고리즘을 설계 및 구현 한다. 알고리즘의 주 목표는 감시정찰용 센서노드의 탐지 신뢰성을 높이기 위한 높은 침입물체 탐지 성공률(success rate)과 낮은 허위신고(false alarm) 횟수를 가지는 것이다. 성능평가 결과에 의하면 제안한 AMIDA 알고리즘은 90% 이상의 탐지 성공률과 2 회 이하의 허위신고 횟수를 가지는 것을 확인할 수 있었다.

• Journal of the Korea Society for Simulation
• /
• v.23 no.4
• /
• pp.41-49
• /
• 2014

Detection and processing techniques for small objects on seabed by the active synthetic aperture sonar can be increased the detection performance because it can be used by short sensor array in small unmanned underwater systems that are spatially constrained. But the limited conditions on constant speed and straight movement of the platform cause a large error in the number of external environmental factors and exact phase synthesis process. In this study, analyzed the applicability of active synthetic aperture processing that is mounted on such a system, and compared detection resolution change in accordance with the phase difference mismatch caused by the along track disturbance. Various simulations were performed as a coherently focus processing model by adding along track disturbance mismatched parameter on the configuring simulator. As the result, detection performance of active synthetic processing for small objects on seabed was found a number of changes by the phase difference mismatch errors according to track disturbances and S/N ratio variations.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.39 no.3
• /
• pp.399-407
• /
• 2019

In Korea, the construction industry has been known to have the highest risk of safety accidents compared to other industries. Therefore, in order to improve safety in the construction industry, several researches have been carried out from the past. This study aims at improving safety of labors in construction site by constructing an effective automatic safety helmet detection system using object detection algorithm based on image data of construction field. Deep learning was conducted using Region-based Fully Convolutional Network (R-FCN) which is one of the object detection algorithms based on Convolutional Neural Network (CNN) with Transfer Learning technique. Learning was conducted with 1089 images including human and safety helmet collected from ImageNet and the mean Average Precision (mAP) of the human and the safety helmet was measured as 0.86 and 0.83, respectively.

• Journal of Broadcast Engineering
• /
• v.26 no.2
• /
• pp.155-166
• /
• 2021

Most object detection algorithms are studied based on RGB images. Because the RGB cameras are capturing images based on light, however, the object detection performance is poor when the light condition is not good, e.g., at night or foggy days. On the other hand, high-quality infrared(IR) images regardless of weather condition and light can be acquired because IR images are captured by an IR sensor that makes images with heat information. In this paper, we performed the object detection algorithm based on the compression ratio in RGB and IR images to show the detection capabilities. We selected RGB and IR images that were taken at night from the Free FLIR Thermal dataset for the ADAS(Advanced Driver Assistance Systems) research. We used the pre-trained object detection network for RGB images and a fine-tuned network that is tuned based on night RGB and IR images. Experimental results show that higher object detection performance can be acquired using IR images than using RGB images in both networks.

• Proceedings of the KIEE Conference
• /
• 2007.07a
• /
• pp.1740-1741
• /
• 2007

본 논문은 이족 로봇에서의 효과적으로 동체를 탐지하는 방법에 대하여 논한다. 이족 로봇의 움직임은 모바일 로봇의 움직임과는 달리 종횡의 움직임이 동시에 나타나게 된다. 따라서 로봇의 비젼이 움직이는 상황에서 움직이는 물체를 탐지해야 한다. 따라서 본 논문에서는 로봇의 움직임을 분석하여 로봇의 움직임을 보정하여 보다 높은 성능의 동체 탐지 성능을 높였다. 제안된 방법을 실제의 로봇으로부터의 영상을 통하여 실험한 결과 우수한 탐지 성능을 얻을 수 있었다.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.25 no.4
• /
• pp.65-74
• /
• 2021

In order to efficiently inspect rapidly increasing old tunnels in many well-developed countries, many inspection methodologies have been proposed using imaging equipment and image processing. However, most of the existing methodologies evaluated their performance on a clean concrete surface with a limited area where other objects do not exist. Therefore, this paper proposes a 6-step framework for tunnel crack detection deep learning model development. The proposed method is mainly based on negative sample (non-crack object) training and Cascade Mask R-CNN. The proposed framework consists of six steps: searching for cracks in images captured from real tunnels, labeling cracks in pixel level, training a deep learning model, collecting non-crack objects, retraining the deep learning model with the collected non-crack objects, and constructing final training dataset. To implement the proposed framework, Cascade Mask R-CNN, an instance segmentation model, was trained with 1561 general crack images and 206 non-crack images. In order to examine the applicability of the trained model to the real-world tunnel crack detection, field testing is conducted on tunnel spans with a length of about 200m where electric wires and lights are prevalent. In the experimental result, the trained model showed 99% precision and 92% recall, which shows the excellent field applicability of the proposed framework.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.29 no.2
• /
• pp.139-146
• /
• 2023

Maritime object detection systems, which detects small maritime obstacles such as fish farm buoys and visualizes distance and direction, is equipped with a 3-axis gimbal to compensate for errors caused by hull motion, but there is a limit to distance error corrections necessitated by the vertical movement of the camera and the maritime object due to wave motions. Therefore, in this study, the distance error of maritime object detection systems caused by the movement of the water surface according to the external environment is analyzed and corrected using average filter and moving average filter. Random numbers following a Gaussian standard normal distribution were added to or subtracted from the image coordinates to reproduce the rise or fall of the buoy under irregular waves. The distance calculated according to the change of image coordinates, the predicted distance through the average filter and the moving average filter, and the actual distance measured by laser distance meter were compared. In phases 1 and 2, the error rate increased to a maximum of 98.5% due to the changes of image coordinates due to irregular waves, but the error rate decreased to 16.3% with the moving average filter. This error correction capability was better than with the average filter, but there was a limit due to failure to respond to the distance change. Therefore, it is considered that use of the moving average filter to correct the distance error of the maritime object detection system will enhance responses to the real-time distance change and greatly improve the error rate.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2001.06a
• /
• pp.161-164
• /
• 2001

본 논문에서는 배경과 색 정보를 이용하여 얼굴을 추출하는 알고리즘을 제안한다. 영상에서의 얼굴 추출에 관한 방법에는 칼라 영상을 가정한 방법, 농담 영상을 가정한 방법, 얼굴의 회전에 덜 민감한 방법, 복잡한 배경에서의 얼굴 추출 방법 등이 연구되어 있다. 본 논문에서는 배경생성을 통해 물체를 구분하고 칼라 정보(HSI 칼라 모델)를 이용하여 얼굴을 추출한다. 배경생성은 각 픽셀 위치에서의 밝기 값을 장시간 평균하거나 혹은 장시간 누적된 밝기 값들 중 최빈 값을 사용하는데 이 방법은 영상 내 물체의 이동이 정체가 별로 없이 원활한 곳에서는 질 좋은 배경을 생성 할 수 있다. 하지 만 배경의 밝기 값을 누적하는 과정에서 물체의 정지상황이 장시간 반영될 경우 배경 영상의 질이 낮아지는 난점이 있다. 따라서, 배경생성 과정에 하이레벨 정보인 물체의 탐지 결과를 이용하여 움직임이 없는 부분에 대해서만 배경생성에 반영함으로써 좀 더 나은 배경을 생성할 수 있다. 이렇게 생성된 배경을 이용해서 입력 영상과의 배경차이를 하게되면 영상 내에서 배경이 아닌 모든 물체를 추출할 수 있다. 물체를 추출 한 후 얼굴 색깔과 유사한 칼라 영역을 분리하고 추출된 물체의 윗 부분에 얼갈이 위치한다는 가정 하에 일괄을 추출한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.11a
• /
• pp.869-872
• /
• 2017

참조 표현이란 영상 내의 특정 물체를 가리키는 자연어 문장을 의미한다. 그리고 이러한 자연어 참조 표현을 기초로, 한 영상에서 실제로 대상 물체의 영역을 찾아내는 일을 참조 표현 이해라고 한다. 본 논문은 참조 표현 이해를 위한 새로운 심층 신경망 모델과 학습 방법을 제안한다. 본 논문에서 제안하는 모델은 효과적인 참조 표현 이래를 위해, 참조 표현에서 언급하는 대상 물체와 보조 물체를 모두 고려할 뿐만 아니라, 두 물체간의 관계정보도 활용한다. 또한, 본 논문에서 제안하는 모델은 이러한 다양한 맥락 정보들을 참조 표현 의존적인 방식으로 가중 결합함으로써, 참조 표현에 부합하는 대상 물체 영역을 보다 정확히 탐지해낼 수 있도록 설계하였다. 본 논문에서는 대규모 참조 표현 데이터 집합인 Google RefExp를 이용한 성능 비교 실험들을 통해, 제안하는 모델의 우수성을 확인하였다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.10a
• /
• pp.999-1001
• /
• 2015

CW Doppler radar is capable of giving the relative velocity of an object using the Doppler effect. When detecting more than an object, frequency domain analysis is needed using CW Doppler radar and FFT. Even though the number of objects and velocities can be obtained within the frequency domain, there is a disadvantage that it is difficult to assess information on approach or deviation of an object. When detecting more than an object using FFT, this study suggests an algorithm for efficiently assessing information about approach or deviation of objects within the frequency domain. The proposed algorithm divides sections into real and imaginary numbers in the frequency domain, and then determines deviation if the total sum of the amplitudes of each frequency is on the left side and approach if the total sum of the amplitudes is on the right side.