摘要
针对道路岔口事故多发的交通环境,提出了一种基于逆投影变换的道路斑马线识别方法.该方法可用于多种视觉采集处理设备如智能眼镜,智能车辆等.方法首先对逆投影变换方法进行改良,使用改良后的逆投影变换将图像进行俯视矫正;之后,对变换后的图片进行形态学变换以消除噪声干扰,再应用双极系数法对斑马线进行识别;最后,采用连通域判决方法消除图像中其他景物对识别造成的影响.实验结果表明,斑马线识别的准确率和实时性得到了提升,改良后的逆投影变换效果突出.
In view of the fact that most traffic accidents occur at the intersections,a method of zebra crossing recognition based on improved IPM is put forward. This method can be used in many visual processing equipment including intelligent glasses and intelligent vehicles. The pictures are first pre-processed. Then an improved inverse perspective mapping is carried out to reduce the per- speetive deformation. After the morphological transform is used to eliminate the noise,bipolarity is applied to recognize the zebra crossing. Finally, connected domain method is used to minimize the noise from the background. The result shows that the improved IPM plays an important role in ze- bra crossing recognition. The high accuracy zebra crossing recognition enhances the safety of pedestrians and cars at the intersections.
出处
《北方工业大学学报》
2013年第3期31-35,66,共6页
Journal of North China University of Technology
关键词
图像处理
逆投影变换
斑马线识别
机器视觉
image processing
inverse perspective mapping
zebra crossing recognition
computer vision
