Detection of airport runway based on the orientation and width spectrums of images

For the detection of airport runway in imaging guidance, a method based on the analysis of orientation and width spectrums was proposed. Firstly, the characteristics of structure and gray of runway in infrared images were analyzed, and then the runway regions were pre-segmented based on the local gray properties. Secondly, sobel operators were used to detect the edges, based on which the orientation spectrum was constructed with the edge orientations calculated with local PCA method. According to the significant peak of orientation spectrum, the main direction of runway was obtained, as well as the mode of the projection cross width spectrum construction. Thirdly, the width spectrum was calculated, from which the projection width could be obtained. With the main orientation and projection width, the geometric width of runway was then calculated, which fulfilled the detection of the runway in infrared image. Experimental results with the infrared images indicate that the method proposed has good performance for runway detection, not only with the high detection rate, but also with nice robustness for geometric distortion.