• Smart Structures and Systems
• /
• v.24 no.5
• /
• pp.669-681
• /
• 2019

Bridge collapses may deliver a huge impact on our society in a very negative way. Out of many reasons why bridges collapse, poor maintenance is becoming a main contributing factor to many recent collapses. Furthermore, the aging of bridges is able to make the situation much worse. In order to prevent this unwanted event, it is indispensable to conduct continuous bridge monitoring and timely maintenance. Visual inspection is the most widely used method, but it is heavily dependent on the experience of the inspectors. It is also time-consuming, labor-intensive, costly, disruptive, and even unsafe for the inspectors. In order to address its limitations, in recent years increasing interests have been paid to the use of unmanned aerial vehicles (UAVs), which is expected to make the inspection process safer, faster and more cost-effective. In addition, it can cover the area where it is too hard to reach by inspectors. However, this strategy is still in a primitive stage because there are many things to be addressed for real implementation. In this paper, a typical procedure of bridge inspection using UAVs consisting of three phases (i.e., pre-inspection, inspection, and post-inspection phases) and the detailed tasks by phase are described. Also, three major challenges, which are related to a UAV's flight, image data acquisition, and damage identification, respectively, are identified from a practical perspective (e.g., localization of a UAV under the bridge, high-quality image capture, etc.) and their possible solutions are discussed by examining recently developed or currently developing techniques such as the graph-based localization algorithm, and the image quality assessment and enhancement strategy. In particular, deep learning based algorithms such as R-CNN and Mask R-CNN for classifying, localizing and quantifying several damage types (e.g., cracks, corrosion, spalling, efflorescence, etc.) in an automatic manner are discussed. This strategy is based on a huge amount of image data obtained from unmanned inspection equipment consisting of the UAV and imaging devices (vision and IR cameras).

• Proceedings of the KIPE Conference
• /
• 2018.07a
• /
• pp.275-277
• /
• 2018

Recently, the importance of an early response has been emphasized due to the large fire. The most efficient method of extinguishing a large fire is early response to a small flame. To implement this solution, we propose a fire detection mechanism based on a deep learning artificial intelligence. In this study, a small amount of data sets is manipulated by an image augmentation technique using rotating, tilting, blurring, and distorting effects in order to increase the number of the data sets by 5 times, and we study the flame detection algorithm using faster R-CNN.

• International journal of advanced smart convergence
• /
• v.8 no.2
• /
• pp.116-125
• /
• 2019

The fire should extinguish as soon as possible because it causes economic loss and loses precious life. In this study, we propose a new atypical fire and smoke detection algorithm using deep learning and color histogram of fire and smoke. First, input frame images obtain from the ONVIF surveillance camera mounted in factory search motion candidate frame by motion detection algorithm and mean square error (MSE). Second deep learning (Faster R-CNN) is used to extract the fire and smoke candidate area of motion frame. Third, we apply a novel algorithm to detect the fire and smoke using color histogram algorithm with local area motion, similarity, and MSE. In this study, we developed a novel fire and smoke detection algorithm applied the local motion and color histogram method. Experimental results show that the surveillance camera with the proposed algorithm showed good fire and smoke detection results with very few false positives.

• The Journal of Korea Robotics Society
• /
• v.14 no.1
• /
• pp.14-21
• /
• 2019

In the ground environment, mobile robot research uses sensors such as GPS and optical cameras to localize surrounding landmarks and to estimate the position of the robot. However, an underwater environment restricts the use of sensors such as optical cameras and GPS. Also, unlike the ground environment, it is difficult to make a continuous observation of landmarks for location estimation. So, in underwater research, artificial markers are installed to generate a strong and lasting landmark. When artificial markers are acquired with an underwater sonar sensor, different types of noise are caused in the underwater sonar image. This noise is one of the factors that reduces object detection performance. This paper aims to improve object detection performance through distortion and rotation augmentation of training data. Object detection is detected using a Faster R-CNN.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.80-81
• /
• 2019

본 논문에서는 자연스러운 장면 영상에서 임의의 방향성을 가진 텍스트를 검출하기 위한 기법을 제안한다. 텍스트 검출을 위한 기본적인 프레임 워크는 Faster R-CNN[1]을 기반으로 한다. 먼저 RPN(Region Proposal Network)을 통해 다른 방향성을 가진 텍스트를 포함하는 bounding box를 생성한다. 이어서 RPN에서 생성한 각각의 bounding box에 대해 세 가지의 서로 다른 크기로 pooling된 특징지도를 추출하고 병합한다. 병합한 특징지도에서 텍스트와 텍스트가 아닌 대상에 대한 score, 정렬된 bounding box 좌표, 기울어진 bounding box 좌표를 모두 예측한다. 마지막으로 NMS(Non-Maximum Suppression)을 이용하여 검출 결과를 획득한다. COCO Text 2017 dataset[2]을 이용하여 학습 및 테스트를 진행하였으며 주관적으로 평가한 결과 기울어진 텍스트에 적합하게 회전된 영역을 얻을 수 있음을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.224-227
• /
• 2022

객체 분류는 입력으로 주어진 이미지에 포함된 객체의 종류를 판단하는 기술이다. 대표적인 딥러닝 기반의 객체 분류 방법으로서 Faster R-CNN[2], YOLO[3] 등의 모델이 개발되었으나, 여전히 성능 향상의 여지가 있다. 본 연구에서는 각도 마진 손실 함수를 기존의 몇 가지 객채 분류 모델에 적용하여 성능 향상을 유도한다. 각도 마진 손실 함수는 얼굴 인식 모델인 SphereFace [4]에서 제안한 방법으로, 얼굴 인식과 같이 단일 도메인의 데이터셋을 분류하는 문제를 풀기 위해 제안되었다. 이는 기존 소프트맥스 함수에서 클래스 결정 경계선에 마진을 주는 방식으로 클래스 간의 구분 능력을 향상시킨다. 본 논문은 각도 마진 손실 함수를 CIFAR10, CIFAR100 데이터셋의 분류 문제에 적용하였으며 ResNet, EfficientNet, MobileNet 등의 백본 네트워크로 실험하여 평균적으로 mAP 성능이 향상되는 것을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.146-149
• /
• 2022

기존 RDO(Rate Distortion Optimization) 기반 압축 방식은 압축 성능에 초점을 두기 때문에 영상 내 인지 특성이 무시될 수 있다. 따라서 RoI(Region of Interest)을 기반으로 압축률을 조절하는 연구가 고안[1, 2, 3, 4] 되었으며, HVS(Human Visual System) 관점에서 영상 내 중요한 부분에 대해 더 높은 품질로 영상을 압축하는 연구가 대부분이다. 최근 인공지능 기술이 발전함에 따라 지능형 영상 분석에 대한 수요가 증가하고 있으며, 이에 따라 머신 비전을 위한 영상 부호화 및 효율적인 전송에 대한 필요성이 대두되고 있다. 본 논문에서는 VVC(Versatile Video Coding)의 dQP(delta Quantization Parameter)를 활용하여 RoI(Region of Interest) 기반압축 방법을 제안하고, 두가지의 RoI 추출 방식을 소개한다. Detectron2 Faster R-CNN X101-FPN [5]의 첫번째 탐지기를 통해 후보 영역 기반 RoI 을 추출하고, 두번째 탐지기를 통해 객체 기반 RoI 을 추출하여, 영상 내 객체 부분과 비객체 부분으로 나누어 서로 다른 압축률로 압축을 수행하였으며, 이에 따른 성능을 비교하고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.740-742
• /
• 2023

객체 인식 연구에 있어서 딥러닝 기반의 사람 인식에 있어서 많은 연구들이 공개되고 있다. 특히 화재사고에 있어서 연기로 인해 가시성이 떨어져 인명구조에 어려움이 발생한다. 이에 열화상 카메라와 딥러닝을 통해 사람을 인식하는 기술이 연구되고 있다. 기존 연구에서는 열화상 카메라와 YOLO 딥러닝을 통해 사람을 인식하는데 95%의 성능을 보였지만, YOLO는 그리드 셀에서 하나의 분류만하기 때문에 물체에 가려진 사람을 판별하는데 정확도가 낮았다. 본 논문에서는 이와 같은 한계를 극복하기 위해 기존 Faster R-CNN 알고리즘을 사용한다. 신체부위 Guided Attention mechanism을 사용하여 가중치를 준 Feature Map을 RPN에 적용시켜 학습모델을 구현한다면 더 높은 정확도를 얻을 수 있다. 향후 본 논문에서 제안하는 기법은 많은 실험과 다양한 데이터 셋을 통해 실질적인 검증을 할 예정이다.

• The Bulletin of The Korean Astronomical Society
• /
• v.46 no.2
• /
• pp.46.2-46.2
• /
• 2021

We present solar event auto detection using deep-learning-based object detection algorithms and DeepSDO event dataset. DeepSDO event dataset is a new detection dataset with bounding boxed as ground-truth for three solar event (coronal holes, sunspots and prominences) features using Solar Dynamics Observatory data. To access the reliability of DeepSDO event dataset, we compared to HEK data. We train two representative object detection models, the Single Shot MultiBox Detector (SSD) and the Faster Region-based Convolutional Neural Network (R-CNN) with DeepSDO event dataset. We compared the performance of the two models for three solar events and this study demonstrates that deep learning-based object detection can successfully detect multiple types of solar events. In addition, we provide DeepSDO event dataset for further achievements event detection in solar physics.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.5
• /
• 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.