基于HSV彩色空间与改进OTSU的车位线分割方法

Parking Line Segmentation Method Based On HSV Color Space And Improved Otsu

车位线分割是智能化泊车的前提,在传统的RGB彩色空间分割车位线的基础上,提出了一种基于HSV彩色空间与改进的OTSU算法的车位线分割方法。首先,将RGB图像转化到HSV空间,利用H、S分量的不同性质并结合改进的OTSU算法依次过滤掉非车位线信息,初步将车位线分割出来。然后,利用形态学处理以及Hough变换结合的方法得到一个较为完善的车位线。改进的OTSU算法引入了粒子群优化算法,相比传统的OTSU算法,大大缩短了分割时间。利用该方法对车位线比较单一环境以及相对复杂环境的图像进行实验,结果表明:该算法能够有效地将车位线区域分割出来,人眼识别效果整体上较好,并大大减少了分割时间,具有较强的实用性。

Parking line segmentation is the premise of intelligent parking.Based on the traditional RGB color space,a parking line segmentation method based on HSV color space and improved Otsu algorithm is proposed.Firstly,the RGB image is transformed into HSV space,and the non parking line information is filtered out by using the different properties of H and s components and the improved Otsu algorithm,and the parking line is initially segmented.Then,a more perfect parking line is obtained by combining morphological processing with Hough transform.The improved Otsu algorithm introduces particle swarm optimization algorithm,which greatly reduces the segmentation time compared with the traditional Otsu algorithm.The experimental results show that the algorithm can effectively segment the parking line area,the human eye recognition effect is better,and the segmentation time is greatly reduced.