复杂航空图像中的大边缘提取方法

The big edge detection of complicated aerial images

提出了一种用于在复杂的航空照片中抑制细小边缘,获取主要景物边缘信息的新方法.该方法首先对图像进行中值滤波和对比度展宽,以去除图像中的噪声,提高图像的清晰度.为了消除图像中的细小目标,提出了一种基于最大偏差最小的邻域平滑法.在边缘提取时采用了梯度算子,并引入模糊理论来判别和跟踪边缘.模糊边缘检测算法的特点是不需要确定门限值,具有很强的自适应性.实验结果显示了该方法在提取复杂照片中的大景物时,不但能获得较连续的边界,而且可以有效地抑制细小目标.

A new algorithm for edge detection is proposed. The main difference between this algorithm and other existing algorithms is that small edges are restrained while the edges of big objects are obtained. The first step of this algorithm is pretreatment, which includes median filtering and contrast extending. The procedure of the median filtering and the filtering window are described. Second, we develop a neighbor smoothing algorithm, which is based on the fact that the maximum deviation is minimum, to eliminate the small object in the image. The smoothing procedure includes steps 1 to 5. Then we employ the gradient operator and fuzzy theory to judge and trace the edges. The member function of fuzzy set and its analytic form are both given. The judgement function, which is used as the criterion for edge points is also described. The experiments show that when we are handling the complicated aerial images, the main edges are obtained while the small edges are restrained.