摘要
机场跑道的卫星图片经过处理后表现出来的骨架特征为边缘直线,在边缘图像中检测直线通常使用的方法是Hough变换(HT).由于(HT)是一种穷举式的搜索,在处理复杂图像时存在大量无效计算,实时性较差.针对机场跑道识别的实时性要求,提出一种改进的用于在二值图像中检测直线的快速Hough变换算法,此算法克服了标准Hough变换以图像边界点为扫描边界的缺点,并且能及时中断无谓的扫描,较好地解决了无效累积问题,实验证明,与标准Hough变换相比,它不仅具备Hough变换原有的高可靠性和抗干扰能力,而且具备Hough变换所不具备的高效性和低存储,克服了标准Hough变换的高计算代价和耗存储的缺点.
Since the bone features of the processed airfield runway＇s remote sensing image are edge lines, we need to extract lines from these images to recognize the airfield runway. Hough transform is a common method to extract lines in binary image, but it can not meet the requirement of real-time processing because of its high complexity in computation and massive storage space demands. In order to recognize the airfield runway fast and accurately, an advanced method based on Hough transform have been developed, in which the useless scanning can be stopped promptly and the image edges are not be used as scanning edges as they do in the standard Hough transform. By this unique way, the recognizing efficiency has been enhanced greatly. The results of experiments have proved that this algorithm performs well and has higher computing speed and less space demand compared with the original Hough transform.
出处
《小型微型计算机系统》
CSCD
北大核心
2005年第12期2200-2202,共3页
Journal of Chinese Computer Systems
基金
基金项目:国家"十五"国防预研项目(806030101)资助
