基于形状角与分块图像的车辆轮廓分割算法

Vehicle contour segmentation algorithm based on shape angle and block image

为了减小对车辆进行外形超限测量时其他车辆对图像的影响,精确地将目标车辆的轮廓信息分割并测量,提出一种基于形状角、结合分块图像与边缘提取的分割算法。对获取图像进行帧差法削弱背景边缘信息的影响,再用Canny算子边缘提取,将得到的轮廓信息进行基于形状角的识别与提取。提出了一种对图像进行分块分类的图像分割法。该算法有效解决了测量中背景对前景的影响。实验结果表明该算法具有较高的鲁棒性与准确性。

In order to reduce the influence of other vehicles on image,and segment and measure the outline information of the target vehicle accurately while the vehicle shape ultralimit measurement is executed,a shape angle based segmentation algo-rithm which is combined with block image and edge extraction is put forward in this paper. The frame differential method is adopted to weaken the influence of image edge information of the background,and then Canny operator edge detection is used to recognize and extract the obtained contour information on the basis of shape angle. An image segmentation method which classi-fies an image by segmenting the image into blocks is proposed. The algorithm effectively solves the influence of background on foreground during the measurement. Experimental results show that the algorithm has strong robustness and high accuracy.