The use of optical flow for road navigation

Abstract

Abstract-This paper describes procedures for obtaining a reliable and dense optical flow from image sequences taken by a television (TV) camera mounted on a car moving in usual outdoor scenarios. The optical flow can be computed from these image sequences by using several techniques. Differential techniques to compute the optical flow do not provide adequate results, because of a poor texture in images and the presence of shocks and vibrations experienced by the TV camera during image acquisition. By using correlation based techniques and by correcting the optical flows for shocks and vibrations, useful sequences of optical flows can be obtained. When the car is moving along a flat road and the optical axis of the TV camera is parallel to the ground, the motion field is expected to be almost quadratic and have a specific structure. As a consequence the egomotion can be estimated from this optical flow and information on the speed and the angular velocity of the moving vehicle are obtained. By analyzing the optical flow it is possible to recover also a coarse segmentation of the flow, in which objects moving with a different speed are identified. By combining information from intensity edges a better localization of motion boundaries are obtained. These results suggest that the optical flow can be successfully used by a vision system for assisting a driver in a vehicle moving in usual streets and motorways.

Vehicle Recognition using NMF in Urban Scene

“…Sun [2] 은 영상을 이용한 차량 인식 시스템은 기본적으로 후 보영역 검출단계와 후보영역에서의 특징 추출 후에 인식 성능을 높이기 위한 검증 단계로 구성된다고 설명한다. 후보 영역 검출을 위한 이전 연구에는 차 량 전면의 밝기 대칭성 [3] 을 이용한 방법, 컬러 정보 [4] 나 밝기 정보의 임계값 이상의 히스토그램을 이용 하여 차량과 배경을 분리하여 검출하는 방법, 코너 점 [6] 이나 수평, 수직 에지성분 [7] 을 이용한 방법, 배 경과 구분되는 차량의 움직임 [8] 이용한 방법, 거리정 보를 이용하는 스테레오 비전 [9,10] 등이 있다. 검증 단계의 연구로는 크게 이미지의 상관관계를 이용하 여 검증을 수행하는 템플릿 기반 방법 [11] 과 신경망 [12] 이나 SVM [13] [14] , 학습 단 계에서 클래스의 구분에 중점을 두는 LDA(Linear Discriminant Analysis) [15] , 통계적 독립성을 이용하 여 독립적인 성분만을 추출하는 ICA(Independent Component Analysis) [16] …”

Lower Tail Light Learning-based Forward Vehicle Detection System Irrelevant to the Vehicle Types

“…HG 단계에서는 차량이 가지고 있는 다양한 특성을 활용한다. 차량의 대칭 특성, 그림자 영역, 색 공간, 수평 및 수직 경계 정보 등을 이용하는 지식 기반 [4,8,9] , 스테레오 이미지 쌍을 이용한 스테레오 기반 [10] , 자동차와 배경 사이를 구분하는 공간적 특성을 활용한 모션 기반이 있다 [11] . 검증단계에서 는 미리 정의된 패턴을 이용하는 템플릿 기반과 자동차 외 형을 학습하는 형상 기반이 있다 [1] .…”

Fixation as a Mechanism for Stabilization of Short Image Sequences

“…In realistic situations however, the computation of optic flow has to cope with undesired motion of the camera due to shocks or vibrations of the vehicle or robot on which it is mounted. These perturbations typically manifest themselves as fast, rotational camera movements [8] that induce large local motions over very short time spans [11]. Consequently, the assumption of locally constant velocities is often violated.…”

“…Consequently, the assumption of locally constant velocities is often violated. A possible solution is to use optic flow algorithms that do not make this assumption [11], such as correlation-based matching techniques. Since the performance and reliability of these techniques on stable sequences, is typically much lower than those of a differential or phase-based approach [5], a better solution is to stabilize the image sequence first.…”