【摘要】 为了解决路侧停车位数目检测问题,本文采用基于图像识别的方法,将复杂的动态视频信息转化为简单的静态图像信息。首先,采用中值滤波、伽马变换和形态学处理等方法对感兴趣区域进行图像预处理;接着采用基于PPHT与直线聚类的方法检测停车位标记线;其次,采用二次方递减图像拼接的方法将该路段所有停车位拼接到一幅图像上,既防止了停车位数目的错检,又简化了处理流程;最后,采用轮廓提取算法对拼接图上的停车位计数。使用4个不同路段的停车位对所提出的方法进行了定量评估,结果表明,该方法的准确率为98.53%,优于以往的方法。更多还原

【Abstract】 In order to solve the problem of detecting the number of parking spaces on the roadside,this paper adopts a method based on image recognition to convert complex dynamic video information into simple static image information. First,use median filtering,gamma transformation,and morphological processing to preprocess the image of the region of interest; then use PPHT and linear clustering to detect parking space marking lines; then use the method of quadratic decreasing image stitching All the parking spaces on the road section are spliced onto one image,which prevents the wrong detection of the number of parking spaces and simplifies the processing flow. Finally,the contour extraction algorithm is used to count the parking spaces on the mosaic map. The proposed method is quantitatively evaluated using parking spaces on four different road sections. The results show that the accuracy of the method is 98.53%,which is better than previous methods.更多还原