• 대한임베디드공학회논문지
• /
• 제13권4호
• /
• pp.195-204
• /
• 2018

Lane keeping assist and departure prevention system are the key functions of ADAS. In this paper, we propose lane detection method which uses Gaussian function based RANSAC. The proposed method consists mainly of IPM (inverse perspective mapping), Canny edge detector, and Gaussian function based RANSAC (Random Sample Consensus). The RANSAC uses Gaussian function to extract the parameters of straight or curved lane. The proposed RANSAC is different from the conventional one, in the following two aspects. One is the selection of sample with different probability depending on the distance between sample and camera. Another is the inlier sample score that assigns higher weights to samples near to camera. Through simulations, we show that the proposed method can achieve good performance in various of environments.

• 소성∙가공
• /
• 제13권1호
• /
• pp.90-101
• /
• 2004

An inverse mosaic method has been proposed to generate an initial blank shape from the final product shape. Differently from the geometric mapping method, the method can handle triangular patches. However, the generated blank shape is strongly dependent on the order of determination of nodes. In order to compensate the dependency error smoothing technique has been also developed. Although the accuracy has been improved greatly compared with the geometrical mapping method, the method has limitation, due to the no incorporation of plasticity theory. Even though the accuracy of the radius vector method is already proved. the method requires initial guess to start the method. In order to compromise the limitation of the present method and the radius vector method, the method has been connected to the radius vector method. The efficiency of the present optimal blank design method has been verified with some chosen examples.

• East Asian mathematical journal
• /
• 제26권1호
• /
• pp.25-38
• /
• 2010

In this paper, we introduce a hybrid iterative method for finding a common element of the set of solutions of a mixed equilibrium problem, the set of common mixed points of finitely many nonexpansive mappings and the set of solutions of the variational inequality for an inverse strongly monotone mapping in a Hilbert space. We show that the iterative sequences converge strongly to a common element of the three sets. The results extended and improved the corresponding results of L.-C.Ceng and J.-C.Yao.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 ITC-CSCC -1
• /
• pp.23-26
• /
• 2000

An improved inverse perspective mapping (IIPM) is proposed so as to reduce computational expense of recovery of 3D road surface. An experimental system based on IIPM is developed to detect lane parameters for a driver assistant system. A re-organized image is obtained quickly and exactly by IIPM. Efficient preprocessing techniques are used to enhance the information of lane and obstacles. Lane in the preprocessed. image is located with region identification. Lane parameters are estimated effectively. An algorithm to adaptively modify the parameters of IIPM is given. Properties of obstacle on 3D road surface are discussed and used to detect obstacles in the current lane and neighboring lanes. Experimental results show that the new method can extract lane state information effectively.

• 대한기계학회논문집A
• /
• 제28권3호
• /
• pp.230-237
• /
• 2004

Lane Sensing techniques based on vision sensors are regarded promising because they require little infrastructure on the highway except clear lane markers. However, they require more intelligent processing algorithms in vehicles to generate the previewed roadway from the vision images. In this paper, a lane sensing algorithm using vision sensors is developed to improve the sensing robustness. The parallel stereo-camera is utilized to regenerate the 3-dimensional road geometry. The lane geometry models are derived such that their parameters represent the road curvature, lateral offset and heading angle, respectively. The parameters of the lane geometry models are estimated by the Kalman filter and utilized to reconstruct the lane geometry in the global coordinate. The inverse perspective mapping from the image plane to the global coordinate considers roll and pitch motions of a vehicle so that the mapping error is minimized during acceleration, braking or steering. The proposed sensing system has been built and implemented on a 1/10-scale model car.

• 대한수학회보
• /
• 제48권5호
• /
• pp.991-1002
• /
• 2011

An iterative algorithm is provided to find a common element of the set of fixed points of a nonexpansive mapping and the set of solutions of some variational inequality in a Hilbert space. Using this result, we consider a strong convergence result for finding a common fixed point of a nonexpansive mapping and a strictly pseudocontractive mapping. Our results include the previous results as special cases and can be viewed as an improvement and refinement of the previously known results.

• 방송공학회논문지
• /
• 제22권4호
• /
• pp.448-461
• /
• 2017

본 논문에서는 교통관제용 CCTV로부터 촬영된 영상에서 교통 상황 분석을 위해 차량의 통행량 및 통행 속도를 획득하는 방법을 제안한다. 제안하는 방법은 촬영된 영상에 역 투영 사상(IPM, Inverse Perspective Mapping) 방법을 이용하여 카메라 각도에 따른 시각적 관점에서 기인한 왜곡을 제거하고, 차선 검출 알고리즘을 통해 1) 차량 통행량, 2) 차량 통행 속도를 측정할 소정 영역을 획득한다. 소정 영역에 대하여 광류(Optical flow) 기반 모션 추정을 이용하여 차량 통행량 및 통행 속도를 획득한다. 본 논문에서 제안한 방법을 지역별 다양한 CCTV 영상인 총 106,993 프레임, 약 세 시간 길이의 영상에 적용하여 88.94%의 검출 성능을 얻을 수 있었다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제4권4호
• /
• pp.201-208
• /
• 2015

본 논문은 차량에 설치된 블랙박스 영상으로부터 도로노면에 표시된 방향지시기호를 효율적으로 검출하는 방안을 제안한다. 제안한 연구에서는 원근 효과를 가진 입력영상에서 역원근변환 방법을 통해 원근 효과를 제거한 실세계 좌표로 매핑 한 평면 영상에서 BOF 특징정보 기반의 신경망 인식기를 통해 검출한다. 입력영상에서 역원근변환과 특징정보의 검출 및 인식은 높은 계산량 때문에 실시간 처리가 어려운 점이 있다. 이를 보완하기 위해 제안한 방안에서는 입력영역의 도로노면 방향지시기호 영역의 특징을 분석하여 도로노면 기호가 포함된 후보 ROI영역을 정의하고 후보 ROI영역의 Gray 색상에서 역원근변환을 수행한다. 그리고 각 도로기호 영역들을 실시간 검출 및 인식하기 위해 인식코자 하는 영역 극소 특징벡터를 추출하고 이를 근소화시킨 클래스로 군집화하여 BOF를 생성한 후 이를 활용한 신경망을 통해 검출한다. 제안한 방안을 도로노면 방향지시기호 검출 연구에 적용한 결과, 약 89% 이상 비교적 정확한 검출률을 제시하였으며, 다양한 도로 환경에서도 높은 검출률을 제시하였다. 따라서 제안한 방안을 안전운전지원시스템을 위한 보다 정확한 도로정보 제공시스템에 적용 가능함을 보인다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1992년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 19-21 Oct. 1992
• /
• pp.479-484
• /
• 1992

The resolved motion rate control (RMRC) is converting to Joint space trajectory from given Cartesian space trajectory. The RMRC requires the inverse of Jacobian matrix. Since the Jacobian matrix of the redundant robot is generally not square, the pseudo-inverse must be introduced. However the pseudo-inverse is not easy to be implemented on a digital computer in real time as well as mathematically complex. In this paper, a simple fuzzy resolved motion rate control (FRMRC) that can replace the RMRC using pseudo-inverse of Jacobian is proposed. The proposed FRMRC with appropriate fuzzy rules, membership functions and reasoning method can solve the mapping problem between the spaces without complexity. The mapped Joint space trajectory is sufficiently accurate so that it can be directly used to control redundant manipulators. Simulation results verify the efficiency of the proposed idea.

• 대한전기학회논문지
• /
• 제39권9호
• /
• pp.985-996
• /
• 1990

An inverse kinematic solution of a robot manipulator using multilayer perceptrons is proposed. Neural networks allow the solution of some complex nonlinear equations such as the inverse kinematics of a robot manipulator without the need for its model. However, the back-propagation (BP) learning rule for multilayer perceptrons has the major limitation of being too slow in learning to be practical. In this paper, a new algorithm named Dynamically Reconfiguring BP is proposed to improve its learning speed. It uses a modified version of Kohonen's Self-Organizing Feature Map (SOFM) to partition the input space and for each input point, select a subset of the hidden processing elements or neurons. A subset of the original network results from these selected neuron which learns the desired mapping for this small input region. It is this selective property that accelerates convergence as well as enhances resolution. This network was used to learn the parity function and further, to solve the inverse kinematic problem of a robot manipulator. The results demonstrate faster learning than the BP network.