提取直线特征实现机场跑道实时检测  

Real-time detection of airport runway by extracting line feature

作　　者：邸男[1,2] 朱明[1] 王毅楠[3] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院研究生院,北京10039 [3]长春轨道客车股份有限公司,吉林长春130062

出　　处：《光学精密工程》2009年第9期2336-2341,共6页Optics and Precision Engineering

基　　金：国家863高技术研究发展计划资助项目(No.2005AA778032)

摘　　要：考虑基于链码跟踪和霍夫变换的传统直线检测方法对噪声敏感且计算量大的缺点,提出了将改进的链码跟踪与分层霍夫变换相结合的方法来提取直线,成功实现了航拍图像中机场跑道的实时高精度检测。采用改进的链码跟踪剔除短的和弯曲的线段,减少霍夫变换处理的点数,同时确定直线的近似方向,缩小霍夫变换的角度搜索范围。对霍夫变换进行金字塔分层计算,减小计算量。在程序设计方面,为图像建立链表结构数组,固定内存空间大小,避免了动态分配内存方式生成链表产生的越界问题。该算法已经成功嵌入TMS320C6416硬件平台。实验结果表明,本算法比传统霍夫变换算法运算量减少了约29倍,与单纯的链码跟踪直线检测相比,能够克服图像的模糊、遮挡等干扰,成功实现了机场跑道的实时精确检测。A novel method combing an Improved Chain Code Based Edge Tracking （ICCBET） with Pyramid Hough Transform （PHT） is proposed to detect the airport runway in real time successfully , for the traditional line detection method by using chain code based edge tracking and Hough transform is sensitive to the noise and hard to satisfy the real time applications. The ICCBET is used to remove short and curving lines, reduce pixels, confirm approximate orientation and to shorten the angle range of PHT processes. Then, the computation cost is reduced considerably by using the stage of pyramid HT. Finally,a chain list array for the image is designed to fix the memory to avoid the overflow which arises from building chain list with memory dynamic allocation. The algorithm has been applied in the TMS320C6416 hardware system. The experiments on various images show that the computation cost by proposed method has decreased by 29 times as compared with that of traditional Hough method and it overcomes the blur and occlusion compared with the original chain code based edge tracking for line detection.

关 键 词：直线检测 链码跟踪 霍夫变换 机场跑道 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]