基于仿真模板的Hough变换直线检测

Simple Model Based Line Detection with Hough Transform

Hough变换直线检测是遥感图像理解、分割和处理中常用的直线特征自动检测技术之一,其运算过程中所需数组的大小决定了参数空间的采样多少,直接关系到计算精度和速度。在此提出的模型仿真、模板调用方法,简便快捷地实现不同数组大小的多次试验和比较,找到满足实际精度和速度要求的数组参数;将仿真模型、仿真过程和所得参数做成易于调用的模板,供检测系统使用。应用中只需将遥感图像像素坐标映射到模型的定义区间,调用后再将结果映射回图像空间。此方法使Hough变换直线检测的调试、修改和移植变得容易,经实际应用证实本方法方便高效。

Line detection with Hough transform provides one of important automatic recognition methods for remote-sensing image processing. In using, Hough transform needs an array whose size decides the sample number and directly influences the computational precision and speed. By setting up a simple model, the testing and comparing of different size of array lead to finding good array's parameters easily. The used model, the calculating process and result of simulation are packed into a template. What else have to do is to transfer image coordinate to the model's zone before calling the template, and do it back after recogniting lines. The advantages of this method of simple model, such as making it easier to modify and transplant software, have been approved by practice.