計算機視覺在汽車碰撞預(yù)警系統(tǒng)中的應(yīng)用

打開文本圖片集

TheApplication of Computer Vision in AutomotiveCollision WarningSystems

Zhao Shengtian，Zhang Qiang (Zhengzhou Shengda University，Zhengzhou ，China）

【Abstract】Aiming at the detection limitations of traditional automotive collsion early warning systems in complex scenarios，this study proposes a visual perceptionalgorithmbased onthe improved YOLOv5-DeepSort fusion architecture. The test verificationwascariedoutfrom three aspects：basic detection performance，environmental adaptabilityand reliabilityof human-computerinteraction.Thetestresultsshowthatthevisualalgorithmbasedonmulti-sensor fusion exhibitsa strong environmental perceptionabilityinconventionaldriving scenarios.Itcanefectively identifythemotion characteristics of typical traic targetsand achieve risk earlywarningina timely manner.

【Key words】computer vision；automobile colision warning system；object detection;YOLOv5；multi-sensor fusion

0 引言

歷經(jīng)多年的發(fā)展與迭代，汽車碰撞預(yù)警系統(tǒng)已能有效降低事故發(fā)生概率。（剩余5090字）