激光杂志
激光雜誌
격광잡지
LASER JOURNAL
2015年
2期
35-38
,共4页
车道检测%支持向量机%特征提取%图像去噪
車道檢測%支持嚮量機%特徵提取%圖像去譟
차도검측%지지향량궤%특정제취%도상거조
lane detection%support vector machine%feature extraction%image denoising
为了提高智能车辆导航系统中的车道线检测正确率,提出一种基于不变矩特征的车道线图像检测算法。首先对原始车道线图像进行预处理,提取感兴趣的区域,得到车道线图像的不变矩特征,然后采用支持向量机对车道线图像不变矩特征向量进行建模,得到车道线的检测结果,最后采用仿真实验对算法的性能进行分析。仿真结果表明,相对于其它车道线检测算法,本文算法不仅提高了车道线检测的正确率,而且获得更高的车道线检测速度,完全能够满足车辆安全驾驶实时性的要求。
為瞭提高智能車輛導航繫統中的車道線檢測正確率,提齣一種基于不變矩特徵的車道線圖像檢測算法。首先對原始車道線圖像進行預處理,提取感興趣的區域,得到車道線圖像的不變矩特徵,然後採用支持嚮量機對車道線圖像不變矩特徵嚮量進行建模,得到車道線的檢測結果,最後採用倣真實驗對算法的性能進行分析。倣真結果錶明,相對于其它車道線檢測算法,本文算法不僅提高瞭車道線檢測的正確率,而且穫得更高的車道線檢測速度,完全能夠滿足車輛安全駕駛實時性的要求。
위료제고지능차량도항계통중적차도선검측정학솔,제출일충기우불변구특정적차도선도상검측산법。수선대원시차도선도상진행예처리,제취감흥취적구역,득도차도선도상적불변구특정,연후채용지지향량궤대차도선도상불변구특정향량진행건모,득도차도선적검측결과,최후채용방진실험대산법적성능진행분석。방진결과표명,상대우기타차도선검측산법,본문산법불부제고료차도선검측적정학솔,이차획득경고적차도선검측속도,완전능구만족차량안전가사실시성적요구。
In order to improve the lane detection accuracy in intelligent vehicle navigation system, this paper pres﹣ents a lane detection algorithm based on moment invariant features. Firstly the original lane image are pre-processed and extract the interest region which invariant feature of lane image are extracted, and then the support vector machine is used to build detection model for the lane image by use feature vectors and obtain detection results of the lane, final﹣ly the simulation experiment is used to analysis the performance of algorithm. The simulation results show that com﹣pared with other lane detection algorithms, the proposed algorithm not only improves the detection correct rate of lane and gets higher detection speed, and it can completely meet the real-time requirement of vehicle safety driving.