In view of the aerial reconnaissance image under complex environment, a new airport runways extraction method is put forward. Firstly, it is utilized the fuzzy enhancement arithmetic to extrude the runways in the image. And then the continuous edges of the runways are obtained after the edge extraction and connection processing etc;some curve shape edges are be decomposed into smaller linear edges after the non linear edges in the image are to dis-connect operation. Finally, the airport runways image is obtained through calculate edge size and remove the small edge. The experimental result shows that the proposed method can rapidly and accurately extract the airport runways in the aerial reconnaissance image,and has good reliability and applicability.%为了更好地在保证图像配准的速度前提下，提高配准的精度，本文提出一种新的基于SURF的图像配准改进算法。改进算法将单向匹配与方向一致性约束两者结合起来，先对待配准图像进行单向匹配，再计算出各个匹配点对之间的特征向量方向差值，当差值在一定阈值范围内时认为匹配成功，从而剔除错配点对，提高匹配的精度。实验结果表明，该算法在保证不增加匹配的时间前提下，提高了匹配的精度。
展开▼
