Vision-based Lane Detection using Hough Transform1. Introduction2. Related work3. Approach(1) Find the road background and subtract it from the original image(2) Hough Transform(3) Search in the Hough space for the long straight lines(4) Decide the lane markings and mid-line of each lane4. Results5. SummaryReferenceAppendixECE 533 Course Project(Digital Image Processing)Vision-based Lane Detection using Hough Transform By Zhaozheng YinInstructor: Prof.Yu Hen HuDec.12 2003PrefacePurpose The purpose of the project component of this course is to demonstrate your ability toapply the knowledge and technique learned during this course. Types of Projects Applications -- Applications of image processing to specific research area. In anapplication project, it should contain the following: (a) explanation of the nature ofthe application and why image processing is needed, (b) image processing techniquesthat can be applied to the problem on hand, (c) preliminary results. You should haveresults ready and report it. Just propose to apply image processing without any resultis not acceptable. Final report content Introduction Related work Approach Results Summary Reference AppendixVision-based Lane Detection using Hough Transform1. IntroductionLane detection is an important enabling or enhancing technology in a number ofintelligent vehicle applications, including lane excursion detection and warning,intelligent cruise control and autonomous driving.Various lane detection methods have been proposed. They are classified intoinfrastructure-based and vision-based approaches. While the infrastructure-basedapproaches achieve highly robustness, construction cost to lay leaky coaxial cables orto embed magnetic markers on the road surface is high. Vision based approaches withcamera on a vehicle have advantages to use existing lane markings in the roadenvironment and to sense a road curvature in front view.Vision-based location of lane boundaries can be divided into two tasks: lanedetection and lane tracking. Lane detection is the problem of locating lane boundarieswithout prior knowledge of the road geometry. Lane tracking is the problem oftracking the lane edges from frame to frame given an existing model of roadgeometry. Lane tracking is an easier problem than lane detection, as prior knowledgeof the road geometry permits lane tracking algorithms to put fairly strong constraintson the likely location and orientation of the lane edges in a new image. Lane detectionalgorithms, on the other hand, have to locate the lane edges without a strong model ofthe road geometry, and do so in situations where there may be a great deal of clutter inthe image. This clutter can be due to shadows, puddles, oil stains, tire skid marks, etc.This poses a challenge for edge-based lane detection schemes, as it is often impossibleto select a gradient magnitude threshold which doesn’t either remove edges of interestcorresponding to road markings and edges or include edges corresponding toirrelevant clutter.Detection of long thick lines, such as highway lane markings from input images,is usually performed by local edge extraction followed by straight line approximation.In this conventional method many edge elements other than lane makings are detectedwhen the threshold of edge magnitude is low, or, in the opposite case, edge elementsexpected to be detected are fragmented when it is high. This makes it difficult to tracethe edge elements and to fit the approximation lines on them.2. Related workIn recent years, lane detection has been broadly studied and many state-of-the-artsystems for detecting and tracking lane/pavement boundaries have shown up. Most lane detection algorithms are edge-based. They relied on thresholding theimage intensity to detect potential lane edges, followed by a perceptual grouping ofthe edge points to detect the lane markers of interest. In many road scenes it isn’tpossible to select a threshold which eliminates noise edges without also eliminatingmany of the lane edge points of interest. University of Michigan’s AI lab uses a test bed vehicle as a platform for datacollection and testing. The approach used by Karl kluge looks for a deformabletemplate model of lane structure to locate lane boundaries without thresholding theintensity gradient information. Chris Kreucher introduced a algorithm based thefrequency domain features that capture relevant information concerning the strengthand orientation of spatial edges. Ohio State University’s Center for Intelligent Transportation Research has alsodeveloped two systems: one performs curve identification in the image plane and useheuristic optimization techniques to construct the road geometry, the other used ageneral image segmentation algorithm to isolate lane marker segments and a votingscheme to unite individual features into lane boundary contours.Toyota Center R&D Labs describes a lane detection method using a real-timevoting processor for a driver assist system. For robust lane detection in variousenvironments, they proposed the method based on complete search in a parameterspace.3. ApproachLane detection is a complicated problem under different light/weather conditions.In this class project we analysis the easy case first: the images are captured from thecrossover above the road, assume the lanes to be detected are straight, at daytime andwith good weather condition. The lane markings can be solid or dash lines. Other thandetecting the lane markers, the mid-line of each lane is also calculated to identify theposition of the vehicle with respect to lane makings, which is useful for autonomousdriving. Fig.1 The lane detection algorithmFig. 1 is the flowchart of the lane detection algorithm, which is based on edgedetection and Hough Transform. First the RGB road image is read in and convertedinto the grayscale image. Then we use the global histogram to find the roadbackground gray and subtract it from the grayscale image to get img1. Edge operationis executed on img1 and lane marking features are preserved in img2. The keytechnology here is using Hough Transform to convert the pixels in img2 from theimage coordinate ),( yx to the
View Full Document
Unlocking...