摘要
为解决常用车底阴影检测方法在复杂光照及背景条件下检测结果不稳定的问题,提出一种基于聚类分析的车底阴影检测方法。使用改进的高斯混合模型聚类算法对交通图像中的目标,即路面、车道线、车辆、车底阴影进行聚类,利用高斯阴影模型的均值与方差自适应阈值分割图像,提取路面与车底阴影的交线,利用阴影的几何结构特征对检测到的阴影线进行两次合并,得到最终结果。实验结果表明,该方法能有效检测车底阴影,适应白天不同时段、光强变化,在复杂投影的干扰下能实现准确检测。
To solve the problem of the instability of most shadow detection methods under the conditions of complex background and light,a vehicle shadow detection method based on clustering analysis was proposed.The Gauss mixture model was applied to cluster the objects in images such as the road,lane lines,shadows and vehicles,shadow threshold was calculated according to the mean and variance of the Gaussian shadow model,thus both the shadow image and the intersecting lines of the vehicle shadow and the road were generated.Finally the initial position of the vehicle was located by merging the lines twice with the use of the geometric structure of the vehicle.The experimental results show that the method can not only detect the shadow underneath a vehicle effectively,but also adapt to the change of different time of the day and the intensity of light.For the interference of com-plex proj ection,the shadow of preceding vehicles can be detected.
出处
《计算机工程与设计》
北大核心
2015年第5期1311-1316,共6页
Computer Engineering and Design
基金
国家自然科学基金项目(91220301)
南京理工大学重点实验室基金项目(30920130122005/6)
