论文部分内容阅读
针对传统遥感系统对特定建筑物提取难的问题,该文提出了一种高空摄影系统下基于改进霍夫变换的城市枢纽立交桥检测方法。该方法首先对图像进行预处理和分割,对分割后图像进行形态学处理提取直线和曲线道路,弥补了单一提取直线道路的不足;然后对提取后的道路进行基于改进霍夫变换的近似圆形立交桥匝道的识别;最后依据圆形完整度门限判断可识别的圆形匝道个数,并根据识别数目来判定是否存在立交桥。经过实验比较分析,该方法对于国内外常规的立交桥系统识别效果都较好。
Aiming at the problem that the traditional remote sensing system is difficult to extract a specific building, this paper presents a detection method of urban hub overpass based on improved Hough transform in aerial photography system. Firstly, the image is preprocessed and segmented. Morphological processing of the segmented images is used to extract straight lines and curvilinear roads to make up for the deficiencies of single straight line extraction. Then, the extracted road is approximated by a modified Hough Transform At last, the number of recognizable circular ramp is judged according to the circular integrity threshold, and whether there is an overpass according to the identification number is determined. After comparative analysis of the experiment, this method has good recognition effect on the conventional overpass system both at home and abroad.