基于信息融合的前方车辆检测技术

Forward Vehicle Detection Technology Based on Information Fusion

为了避免出现交通事故,使车辆检测效率与准确率得到提高,提出基于车辆阴影、纹理、边缘等信息融合的前方车辆检测技术。基于机器视觉使用多特征结合方法对目标车辆进行确定,通过激光扫描得到车辆距离与宽度,对目标特征提取创建车辆验证函数。之后,根据车辆验证函数创建分类器,验证视觉图像跟踪车辆,之后融合视觉传感器和激光扫描信息。通过验证表示,能够对前方车辆精准检测,能够避免汽车碰撞。

In order to avoid traffic accidents and improve the efficiency and accuracy of vehicle detection, a front vehicle detection technology based on vehicle shadow, texture, edge and other information fusion is proposed. The vehicle feature is created and verified by using the laser scanning function. Then, a classifier is created according to the vehicle verification function to verify the visual image, track the vehicle, and then fuse the visual sensor and laser scanning information. The verification shows that it can accurately detect the vehicle in front and avoid car collision.