• 한국해양공학회지
• /
• 제35권1호
• /
• pp.91-97
• /
• 2021

Marine accidents caused by ships have brought about economic and social losses as well as human casualties. Most of these accidents are caused by small and medium-sized ships and are due to their poor conditions and insufficient equipment compared with larger vessels. Measures are quickly needed to improve the conditions. This paper discusses a video-integrated collision prediction and fall detection system to support the safe navigation of small- and medium-sized ships. The system predicts the collision of ships and detects falls by crew members using the CCTV, displays the analyzed integrated information using automatic identification system (AIS) messages, and provides alerts for the risks identified. The design consists of an object recognition algorithm, interface module, integrated display module, collision prediction and fall detection module, and an alarm management module. For the basic research, we implemented a deep learning algorithm to recognize the ship and crew from images, and an interface module to manage messages from AIS. To verify the implemented algorithm, we conducted tests using 120 images. Object recognition performance is calculated as mAP by comparing the pre-defined object with the object recognized through the algorithms. As results, the object recognition performance of the ship and the crew were approximately 50.44 mAP and 46.76 mAP each. The interface module showed that messages from the installed AIS were accurately converted according to the international standard. Therefore, we implemented an object recognition algorithm and interface module in the designed collision prediction and fall detection system and validated their usability with testing.

• Computers and Concrete
• /
• 제26권5호
• /
• pp.411-420
• /
• 2020

Deep Convolution neural network (DCNN) has been widely used in the healthy maintenance of civil infrastructure. Using DCNN to improve crack detection performance has attracted many researchers' attention. In this paper, a light-weight spatial attention network module is proposed to strengthen the representation capability of ResNet and improve the crack detection performance. It utilizes attention mechanism to strengthen the interested objects in global receptive field of ResNet convolution layers. Global average spatial information over all channels are used to construct an attention scalar. The scalar is combined with adaptive weighted sigmoid function to activate the output of each channel's feature maps. Salient objects in feature maps are refined by the attention scalar. The proposed spatial attention module is stacked in ResNet50 to detect crack. Experiments results show that the proposed module can got significant performance improvement in crack detection.

• 대한임베디드공학회논문지
• /
• 제18권5호
• /
• pp.259-265
• /
• 2023

Pedestrian detection is used to effectively detect pedestrians in various situations based on deep learning. Pedestrian detection has difficulty detecting pedestrians due to problems such as camera performance, pedestrian description, height, and occlusion. Even in the same pedestrian, performance in detecting them can differ according to the height of the pedestrian. The height of general pedestrians encompasses various scales, such as those of infants, adolescents, and adults, so when the model is applied to one group, the extraction of data becomes inaccurate. Therefore, this study proposed a pedestrian detection method that fine-tunes the pedestrian area by Refining Layer and Feature Concatenation to consider various heights of pedestrians. Through this, the score and location value for the pedestrian area were finely adjusted. Experiments on four types of test data demonstrate that the proposed model achieves 2-5% higher average precision (AP) compared to Faster R-CNN and DRPN.

• Structural Engineering and Mechanics
• /
• 제69권4호
• /
• pp.407-416
• /
• 2019

This paper develops a wireless sensor for online fatigue crack detection and failure warning based on crack-induced nonlinear ultrasonic modulation. The wireless sensor consists of packaged piezoelectric (PZT) module, an excitation/sensing module, a data acquisition/processing module, a wireless communication module, and a power supply module. The packaged PZT and the excitation/sensing module generate ultrasonic waves on a structure and capture the response. Based on nonlinear ultrasonic modulation created by a crack, the data acquisition/processing module periodically performs fatigue crack diagnosis and provides failure warning if a component failure is imminent. The outcomes are transmitted to a base through the wireless communication module where two-levels duty cycling media access control (MAC) is implemented. The uniqueness of the paper lies in that 1) the proposed wireless sensor is developed specifically for online fatigue crack detection and failure warning, 2) failure warning as well as crack diagnosis are provided based on crack-induced nonlinear ultrasonic modulation, 3) event-driven operation of the sensor, considering rare extreme events such as earthquakes, is made possible with a power minimization strategy, and 4) the applicability of the wireless sensor to steel welded members is examined through field and laboratory tests. A fatigue crack on a steel welded specimen was successfully detected when the overall width of the crack was around $30{\mu}m$, and a failure warnings were provided when about 97.6% of the remaining useful fatigue lives were reached. Four wireless sensors were deployed on Yeongjong Grand Bridge in Souht Korea. The wireless sensor consumed 282.95 J for 3 weeks, and the processed results on the sensor were transmitted up to 20 m with over 90% success rate.

• 센서학회지
• /
• 제31권2호
• /
• pp.131-137
• /
• 2022

Crimes related to illegal cameras are steadily increasing and causing social problems. Owing to the development of camera technology, the miniaturization and high performance of illegal cameras have caused anxiety among many people. This study is for detecting hidden cameras effectively such that they could not be easily detected by human eyes. An image sensor-based module with 940 nm wavelength infrared detection technology was developed, and an image processing algorithm was developed to selectively detect illegal cameras. Based on the Android smartphone environment, image processing technology was applied to an image acquired from an infrared camera, and a detection sensor module that is less sensitive to ambient brightness noise was studied. Experiments and optimization studies were conducted according to the Gaussian blur size, adaptive threshold size, and detection distance. The performance of the infrared image-based illegal camera detection sensor module was excellent. This is expected to contribute to the prevention of crimes related to illegal cameras.

• 조명전기설비학회논문지
• /
• 제25권8호
• /
• pp.99-105
• /
• 2011

This paper is about the design of 3-axis magnetic sensor module which detects parking and moving vehicle. For the sensor module, MR Sensor from Honeywell of which maximum measurement range is ${\pm}2$[G] is used. It also consisted of amplifier and sensor filter and fabricated $30{\times}50$[mm] PCB. Fabricated sensor module produced helmholtz coil of which the length is 1.2[m] of 3-axis to know the performance. It installed sensor module at the center and measured the detected magnetic field. In result, 3-axis were detected as 0.2~0.3[mG] and the drift of the fluctuation of magnetic field was stabilized at 0.03[mG] unit. For the performance evaluation of the vehicle detection, after the entry and parking of the vehicle, variation of magnetic field was measured as 0.323~0.695[G] which the average 0.5[G] of the earth magnetic field was the center and the range of variation was confirmed as 0.37[G]. Therefore, the designed magnetic sensor can be used as the vehicle detection sensor module.

• 대한원격탐사학회지
• /
• 제38권5_2호
• /
• pp.793-810
• /
• 2022

실시간 범람 모니터링을 위해 인공위성 SAR영상을 활용하는 수체탐지에 대한 필요성이 대두되었다. 주야와 기상에 상관없이 주기적으로 촬영 가능한 인공위성 SAR 영상은 육지와 물의 영상학적 특징이 달라 수체탐지에 적합하나, 스페클 노이즈와 영상별 상이한 밝기 값 등의 한계를 내포하여 다양한 시기에 촬영된 영상에 일괄적으로 적용 가능한 수체탐지 알고리즘 개발이 쉽지 않다. 이를 위해 본 연구에서는 Convolutional Neural Networks (CNN)기반 모델인 U-Net 아키텍처에 레이어의 조합인 모듈을 추가하여 별도의 전처리 없이 수체탐지의 정확도 향상 방법을 제시하였다. 풀링 레이어의 조합을 활용하여 형태학적 연산처리 효과를 제공하는 Morphology Module과 전통적인 경계탐지 알고리즘의 가중치를 대입한 컨볼루션 레이어를 사용하여 경계 학습을 강화시키는 Edge-enhanced Module의 다양한 버전을 테스트하여, 최적의 모듈 구성을 도출하였다. 최적의 모듈 버전으로 판단된 min-pooling과 max-pooling이 연속으로 이어진 레이어와 min-pooling로 구성된 Morphology 모듈과 샤를(Scharr) 필터를 적용한 Edge-enhanced 모듈의 산출물을 U-Net 모델의 conv 9에 입력자료로 추가하였을 때, 정량적으로 9.81%의 F1-score 향상을 보여주었으며, 기존의 U-Net 모델이 탐지하지 못한 작은 수체와 경계선을 보다 세밀하게 탐지할 수 있는 성능을 정성적 평가를 통해 확인하였다.

• 한국도서관정보학회지
• /
• 제46권1호
• /
• pp.321-344
• /
• 2015

이 논문은 여러가지 자료를 결합해 어떤 주제에 관한 의견이 실려있는 블로그를 찾는 융합적 의견탐지방법을 소개한다. 주제에 관한 의견이 담긴 블로그를 찾기위해 이 연구는 기존의 IR 방법으로 주제에 관한 블로그를 검색한 후 여러가지 의견탐지 방법을 합산한 의견점수로 검색결과의 순위를 조정하는 방법을 쓴다. 의견탐지 모듈의 주요 구성 요소는 의견이 실려있는 블로그에 자주 나오는 단어들을 활용한 고빈도 모듈, 강한 감정을 표현하는 희귀 한 용어들을 (e.g., "sooo good") 활용한 저빈도 모듈, "I"와 "you"에 묶인 n-gram을 (e.g., I believe, You will love) 활용한 IU모듈, 윌슨의 주관 용어 목록을 바탕으로 한 윌슨의 어휘모듈, 그리고 소수의 의견 약어를 (e.g., imho) 이용한 의견 약어 모듈들 이다. 본 연구의 결과는 여러 가지 방법을 융합하는 것이 의견 검출 성능을 향상시키는데 효과적이 다는 것을 보여주었다.

• 한국전자통신학회논문지
• /
• 제12권1호
• /
• pp.109-114
• /
• 2017

본 연구에서는 mA급 유효성분(저항성) 누설전류를 감지할 수 있는 모듈을 개발하였다. 감지된 누설전류를 기술표준규격에서 규정하는 0.03초 이내에 차단을 할 수 있도록 16 bit 신호처리 프로세서인 MSP430 프로세서를 사용하여 모듈을 구현하였다. 개발된 모듈은 스마트그리드의 스마트분전반에서 감전예방을 위한 모듈로 적용 가능할 것으로 기대된다.

• 한국통신학회논문지
• /
• 제13권2호
• /
• pp.171-181
• /
• 1988

본 논문에서는 packet-switched network에서 음성을 서비스하는데 있어서 고려해야 할 여러가지 점들을 살펴보고, 실제로 음성과 데이터를 동시에 서비스하는 packet voice/data terminal을 구현하였으며 그 성능 분석을 시도하였다. PVDT의 software는 OSI 7 layer architecture에 맞추어 설계하였으며 음성과 데이터를 link level부터 구별하여 서비스하였다. 또한 음성 packet의 전송 delay를 작게 하기 위해 데이터보다 음성을 우선적으로 서비스하도록 하였으며 간략화된 protocol로 재전송에 의한 overhead를 없앴다. PVDT의 hardware의 구성은 기능별로 master control module, speech processing module, speech activity detection module, telelphone interface module, input/output inteface module로 나누어진다. Packet음성통신망에 대한 해석으로는 음성 packet의 전송 delay의 variance에 의한 영향을 줄이기 위한 최적 재생지연시간을 전송 delay의 분포를 통해 계산하였다.