基于七宫格的车辆分割算法研究

发布时间：2018-06-06 08:03

本文选题：凹性分析 + 七宫格　；参考：《辽宁工程技术大学》2014年硕士论文

【摘要】：道路交通场景中,车辆间的距离过近或相互遮挡容易造成识别上的粘连。为此,本文主要对如何有效分割粘连车辆的问题展开了研究。将相邻的横竖2×2个像素点归为一个子块,并通过这四个像素点的值求出该子块信息值,以这种方式对图像进行分块处理并通过背景差分法得到前景图像。对前景图像进行阴影及噪声点的去除操作,计算出运动区域。通过粘连车辆判断准则检测出粘连车辆区域。使用“七宫格”扫描车辆的轮廓,检测出粘连车辆间的凹陷区域。通过匹配对应的凹陷区域,找到车辆间的粘连区域。采用边缘检测算法检测出粘连区域内的边缘信息,将信息量最多的边缘作为疑似分割线。将疑似分割线的端点和粘连区域两个侧边的中心点分别相连,形成的线段就是粘连车辆的分割线,并通过分割线去分割粘连车辆。最后,将车辆跟踪等算法与粘连车辆分割方法结合运用,实现对粘连车辆分割准确性的评估。
[Abstract]:In road traffic scene, the distance between vehicles is too close or mutual occlusion is easy to cause recognition adhesion. Therefore, this paper mainly studies how to partition the adhesion vehicle effectively. The adjacent 2 脳 2 pixels are classified into one sub-block, and the information value of the sub-block is obtained by the values of the four pixels. The image is divided into blocks in this way and the foreground image is obtained by the background difference method. The shadow and noise points are removed from the foreground image, and the moving region is calculated. The adhesive vehicle area is detected by the criterion of adhesion vehicle judgment. Scan the contour of the vehicle using the "seven grid" to detect the concave area between the vehicles. By matching the corresponding depression area, the adhesion area between the vehicles is found. Edge detection algorithm is used to detect the edge information in the adhesive area, and the edge with the most information is regarded as the suspected split-line. The ends of the suspected Secant and the central points of the two sides of the adhesive region are connected respectively. The line segment formed is the split-line of the adhesion vehicle and the vehicle is separated by the Secant Line. Finally, the vehicle tracking algorithm and the adhesion vehicle segmentation method are used to evaluate the accuracy of the adhesion vehicle segmentation.
【学位授予单位】：辽宁工程技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;TP391.41

【参考文献】