Method and device for image-assisted runway localization

A method for runway localization on the basis of a feature analysis of at least one image of the runway surroundings taken by a landing aircraft, wherein in order to determine the central axis of the runway, feature matching between image features of the image and mirrored image features of the image is carried out, wherein features of the one runway side are made to be congruent with features of the other runway side, and wherein the maximum similarity of the two feature images is attained when the minor axis, as the center axis of the runway, lies in the center of the runway, wherein the image before determining the feature images is converted into a horizontally oriented image, each time from the image or the horizontally oriented image a feature image and a mirrored feature image is generated by a feature extraction, the feature extraction is done by determining the horizontal edge thickness in the image or the horizontally oriented image, the determination of the horizontal edge thickness is done by means of convolution of the respective image with a horizontal filter mask of form [−1, 0, 1] and subsequent generation of the amplitude of the filter response, wherein to produce a feature matching image one forms a correlation between individual corresponding lines in the feature image of the image and the mirrored feature image of the image according to the following relation

• V(k)=IFFT(FFT(M1(k))*CONJ(FFT(M2(k))) ), with
• FFT:
• the fast Fourier transform;
• IFFT:
• the inverse fast Fourier transform
• M1:
• the feature image
• M2:
• the mirrored feature image
• M1(k) and M2(k):
• the k-th lines of the feature images
• CONJ:
• complex conjugation.