航空图像中线状目标的自动识别

本文构造了一个航空图像中线状目标的自动识别系统，其中包括低层目标的细化边缘提取，中层目标中心链的特征表达和高层的目标识别定位。整个过程充分利用人类关于线状目标的知识，提出了一系列针对线状目标特征的识别算法，解决了高噪声航拍图片中线状目标的边缘提取及边缘提取不连续所带来的识别难题，并用可靠的实验结果证明了系统的有效性和普适性。

Automatic Recognition of the Line-featured Objects in Aerial Images

An automatic system of recognizing line-featured objects (LFO) in the aerialimages is described in this paper, including thinned edge detection in low level processing,feature representation using the object center chain in intermedium level processing and object recognition and locating in high level processing. A set of algorithms, especially for theLFOs, are put forward by taking advantage of the knowledge about LFOs. Successful edgedetection of LFOs was realized in the highly noisy aerial images and the difficulties caused bydiscontinuous edges were also solved. The reliable experiment results demonstrate the effectiveness and generality of the system.