Vision Based Lane detection for Active Security in Intelligent Vehicle

Abstract

This paper presents a monochromic camera lane can approximate normal road structures, it still cannot detection algorithm based on Snake for improving active describe some cases, i.e. an "S" turn. A deformable template security in intelligent vehicle. The algorithm allows the method was proposed by optimizing a likelihood function elimination of several common assumptions: flat road, based on this model. However, this algorithm cannot constant pitch angle or absence of roll angle, which are the guarantee a global optimum and the accuracy, without limitations of many former algorithms. Moreover, compared . . .with the former methods, the snake based lane model is able to requirig hu c ationa mresourcs. describe any arbitrary shape by a set of control points. We use The approach based on morphological filterig has been an improved Hough transform method to wipe off suggested [8] [9]. This technique used the morphological disturbances caused by shadow or edge of other things. Then watershed" transformation to locate the lane edges in the we get the accurate vanishing points using a method of mutual intensity gradient magnitude image, although this technique comparison to desert some bad ones. The method of update has the advantage of not requiring any threshold for the and tracking is also referenced at last. The results of gradient magnitudes, it has the disadvantage ofnot imposing experiments, which have been done in urban road, show that any global constraints on the lane edge shapes. the algorithm is adapted to many road conditions. Even in aIn this paper, we present an algorithm that is able to detect complex driving environment, it also has a good performance. and follow the roads quickly and robustly. In the first step,

Autonomous cross-floor navigation of a stair-climbing mobile robot using wireless and vision sensor

“…A vanishing point detection method which searches the intersection points at infinity using imaginary left and right lines was applied in this study. Define two linear equations for the two side lines as follows: (1) where (X2,Y2) and (X4,Y4) are end point coordinate values, mr and mR are slopes. The intersection point (xc,Yc) in eq.…”

“…For example, some researchers applied Canny Detection [1] and Hough Transform [2,3] algorithms to perform lane detection and tried some visual methods [4] to alleviate the influences of luminance change. Due to their effectiveness in recent years these methods are further applied to robot indoor navigation [5].…”

Multi-Sensor Based Perception Network for Vehicle Driving Assistance

“…That's why autonomous vehicle driving is still hard to achieve [12]. In DARPA competition held in USA, only a few autonomous vehicles can reach the destination in the end [4].…”

Lane Detection and Tracking Using a Parallel-snake Approach

“…Lanes have been modeled as being straight in [4,5,8,13] to reduce the computational complexity; as piecewise lines in [18] to achieve better accuracy; and as parabolas [17,25,29] and B-splines [27,28] to achieve greater flexibility. Unfortunately, few of these works have taken into account geometric properties, such as the parallelism between lane boundaries.…”