• 예술인문사회 융합 멀티미디어 논문지
• /
• 제7권3호
• /
• pp.339-349
• /
• 2017

본 논문에서는 주행 중인 차량의 차선 인식을 위해 4단계로 구성된 알고리즘을 제안한다. 첫 번째 단계에서는 관심영역 추출한다. 두 번째 단계에서는 신호 잡음을 제기하기 위해 중간 값 필터를 이용한다. 세 번째 단계에서는 입력되는 이미지의 배경과 전경의 두 클래스로 구분하기 위한 이진화 알고리즘을 수행한다. 마지막 단계에서는 이진화 과정 후에 남아 있는 노이즈나 불완전한 에지 등을 제거하여 선명한 차선을 얻기 위해 이미지 침식 알고리즘을 이용한다. 하지만 이러한 차선 인식 앍고리즘은 높은 계산량을 요구하여 실시간 처리가 어려운 실정이다. 따라서 본 논문에서는 멀티코어 아키텍처를 이용하여 실시간 차선이탈 감지 알고리즘을 병렬구현 한다. 또한, 차선이탈 감지 알고리즘을 위한 최적의 멀티코어 아키텍처의 구조를 탐색하기 위해 총 8가지의 서로 다른 프로세싱 엘리먼트 구조를 이용하여 실험하였고, 모의실험 결과 40×40의 프로세싱 엘리먼트 구조에서 최적의 성능, 에너지 효율 및 면적 효율을 보였다.

• 한국지능시스템학회논문지
• /
• 제18권1호
• /
• pp.133-138
• /
• 2008

In this paper, we propose a novel anchor shot detection system, named to MASD (Multi-phase Anchor Shot Detection), which is a core step of the preprocessing process for the news video analysis. The proposed system is composed of four modules and operates sequentially: 1) skin color detection module for reducing the candidate face regions; 2) face detection module for finding the key-frames with a facial data; 3) vector representation module for the key-frame images using a non-negative matrix factorization; 4) one class SVM module for determining the anchor shots using a support vector data description. Besides the qualitative analysis, our experiments validate that the proposed system shows not only the comparable accuracy to the recently developed methods, but also more faster detection rate than those of others.

• 한국정보전자통신기술학회논문지
• /
• 제16권6호
• /
• pp.392-399
• /
• 2023

본 논문에서는 김치 제조 공정 중 배추 심 제거 공정의 로봇 자동화를 위한 배추 심 영역 및 깊이를 판별하는 딥러닝 모델을 제안하는 것이다. 또한 계측된 배추의 심 깊이를 예측하는 것이 아닌 discrete 클래스로 변환하여 영역 검출 및 분류를 동시에 하는 모델을 제시하였다. 딥러닝 모델 학습 및 검증을 위하여 전처리 과정을 거지치 않고 수확된 배추 522 포기에 대한 RGB 영상을 획득하였다. 획득한 영상으로부터 심 영역 및 깊이 라벨링 그리고 데이터 증강 기법을 적용하였다. 제안하는 YOLO-v4 딥러닝 모델 기반 배추 심 영역 검출 및 분류 모델의 성능을 평가하기 위하여 mAP, IoU, accuracy, sensitivity, specificity 그리고 F1-score로 선정하였다. 그 결과 배추 심 영역 검출은 mAP 그리고 IoU 값이 각각 0.97 그리고 0.91로 나타났으며, 심 깊이 분류의 경우 accuracy 그리고 F1-score 값이 각각 96.2% 그리고 95.5%로 나타났다. 본 연구 결과를 통하여 배추의 심 영역 검출 및 깊이 정보 분류가 가능하며, 추후 배추 심 제거 공정의 로봇-자동화 시스템 개발에 활용될 수 있는 가능성을 확인하였다.

• 전기학회논문지
• /
• 제56권7호
• /
• pp.1321-1323
• /
• 2007

A new reference point location method based on an edge map is proposed, where an orientation map is defined and used to find the edge map. Experimental results show that the proposed method can effectively detect the core point in poor quality and arch-type fingerprint images and produces better results in terms of the detection rate and accuracy than the sine map-based method.

• 한국지능시스템학회:학술대회논문집
• /
• 한국지능시스템학회 2007년도 추계학술대회 학술발표 논문집
• /
• pp.217-220
• /
• 2007

In this paper, we present a new anchor shot detection system which is a core step of the preprocessing process for the news video analysis. The proposed system is composed of four modules and operates sequentially: 1) skin color detection module for reducing the candidate face regions; 2) face detection module for finding the key-frames with a facial data; 3) vector representation module for the key-frame images using a non-negative matrix factorization; 4) anchor shot detection module using a support vector data description. According to our computer experiments, the proposed system shows not only the comparable accuracy to the recent other results, but also more faster detection rate than others.

• 한국방사성폐기물학회:학술대회논문집
• /
• 한국방사성폐기물학회 2009년도 학술논문요약집
• /
• pp.84-85
• /
• 2009

New approaches for detecting, preventing and remedying environmental damage are important for protection of the environment. Procedures must be developed and implemented to reduce the amount of waste produced in chemical processes, to detect the presence and/or concentration of contaminants and decontaminate fouled environments. Contamination can be classified into three general types: airborne, surface and structural. The most dangerous type is airborne contamination, because of the opportunity for inhalation and ingestion. The second most dangerous type is surface contamination. Surface contamination can be transferred to workers by casual contact and if disturbed can easily be made airborne. The decontamination of the surface in the nuclear facilities has been widely studied with particular emphasis on small and large surfaces. The amount of wastes being produced during decommissioning of nuclear facilities is much higher than the total wastes cumulated during operation. And, the process of decommissioning has a strong possibility of personal's exposure and emission to environment of the radioactive contaminants, requiring through monitoring and estimation of radiation and radioactivity. So, it is important to monitor the radioactive contamination level of the nuclear facilities for the determination of the decontamination method, the establishment of the decommissioning planning, and the worker's safety. But it is very difficult to measure the surface contamination of the floor and wall in the highly contaminated facilities. In this study, the poly(styrene-ethyl acrylate) [poly(St-EA)] core-shell composite polymer for measurement of the radioactive contamination was synthesized by the method of emulsion polymerization. The morphology of the poly(St-EA) composite emulsion particle was core-shell structure, with polystyrene (PS)as the core and poly(ethyl acrylate) (PEA) as the shell. Core-shell polymers of styrene (St)/ethyl acrylate (EA) pair were prepared by sequential emulsion polymerization in the presence of sodium dodecyl sulfate (SOS) as an emulsifier using ammonium persulfate (APS) as an initiator. The polymer was made by impregnating organic scintillators, 2,5-diphenyloxazole (PPO) and 1,4-bis[5-phenyl-2-oxazol]benzene (POPOP). Related tests and analysis confirmed the success in synthesis of composite polymer. The products are characterized by IT-IR spectroscopy, TGA that were used, respectively, to show the structure, the thermal stability of the prepared polymer. Two-phase particles with a core-shell structure were obtained in experiments where the estimated glass transition temperature and the morphologies of emulsion particles. Radiation pollution level the detection about under using examined the beta rays. The morphology of the poly(St-EA) composite polymer synthesized by the method of emulsion polymerization was a core-shell structure, as shown in Fig. 1. Core-shell materials consist of a core structural domain covered by a shell domain. Clearly, the entire surface of PS core was covered by PEA. The inner region was a PS core and the outer region was a PEA shell. The particle size distribution showed similar in the range 350-360 nm.

• 한국지반환경공학회 논문집
• /
• 제16권8호
• /
• pp.21-35
• /
• 2015

본 연구에서는 국내 3개 필 댐의 ECRD(Earth Core Rock-fill Dams) 댐체의 코어존에 대한 건전성 평가를 목적으로 2차원 및 3차원 전기비저항 탐사를 수행하였다. 2차원 전기비저항 탐사 결과, 대부분의 점토재로 축조된 코어존은 $50{\sim}400ohm{\cdot}m$ 이하의 저비저항대로 나타났으며, 그 하부의 기반암은 $1,000ohm{\cdot}m$ 이상의 고비저항대로 나타났다. 또한 3차원 전기비저항 탐사 결과에 의하면 코어존의 연약대로 판단되는 $100ohm{\cdot}m$ 이하의 저비저항대의 공간적인 분포영역을 확인할 수 있었다. 아울러 코어존의 토질시료 특성을 파악하기 위하여 시추조사와 토질 실내시험을 수행하였다. 시추조사 시 굴착수에 의한 댐체 내의 구조적 훼손을 방지하기 위하여 무수보링 방법을 택하였다. 그 결과 저비저항대로 나타나는 코어존의 시료는 모두 통일분류상 CL에 해당되며, 일부 저비저항대에서는 함수비가 높은 포화상태의 시료가 관찰되었다. 이러한 결과를 통해 ECRD 댐체 코어존의 건전성 평가에 전기비저항 탐사와 무수보링의 적용성은 매우 효율적인 방법이라 판단된다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권10호
• /
• pp.5159-5178
• /
• 2018

Network Intrusion detection is a rapidly growing field of information security due to its importance for modern IT infrastructure. Many supervised and unsupervised learning techniques have been devised by researchers from discipline of machine learning and data mining to achieve reliable detection of anomalies. In this paper, a deep convolutional neural network (DCNN) based intrusion detection system (IDS) is proposed, implemented and analyzed. Deep CNN core of proposed IDS is fine-tuned using Randomized search over configuration space. Proposed system is trained and tested on NSLKDD training and testing datasets using GPU. Performance comparisons of proposed DCNN model are provided with other classifiers using well-known metrics including Receiver operating characteristics (RoC) curve, Area under RoC curve (AuC), accuracy, precision-recall curve and mean average precision (mAP). The experimental results of proposed DCNN based IDS shows promising results for real world application in anomaly detection systems.

• 동력기계공학회지
• /
• 제16권6호
• /
• pp.86-91
• /
• 2012

A multi purpose environmental monitoring system has been developed as a commercially available standard using the techniques which are FBG(Fiber Bragg Grating), Hetero-core spliced fiber optic sensor and etc, for the purposes of monitoring large scaled structures and preserving natural environments. The monitoring system has been tested and evaluated in a possible outdoor condition in view of the full scaled operation at actual sites to be monitored. Additionally, the developed systems in the previous works conveniently provided us with various options of sensor modules intended for monitoring such physical quantities as displacement, distortion, pressure, binary states, and liquid adhesion. In this paper, we extend the previous results to a water leakage detection problem and develop a sensing system as a result. By the experimental study, it is verified that multi-point leakage detection is possible using single line optical fiber.

• 로봇학회논문지
• /
• 제14권1호
• /
• pp.1-7
• /
• 2019

Recently, smart factories have attracted much attention as a result of the 4th Industrial Revolution. Existing factory automation technologies are generally designed for simple repetition without using vision sensors. Even small object assemblies are still dependent on manual work. To satisfy the needs for replacing the existing system with new technology such as bin picking and visual servoing, precision and real-time application should be core. Therefore in our work we focused on the core elements by using deep learning algorithm to detect and classify the target object for real-time and analyzing the object features. We chose YOLO CNN which is capable of real-time working and combining the two tasks as mentioned above though there are lots of good deep learning algorithms such as Mask R-CNN and Fast R-CNN. Then through the line and inside features extracted from target object, we can obtain final outline and estimate object posture.