The Real Time Scene Matching Algorithm for Railcar Aided Navigation

In order to provide a railcar image-aided navigation technique, aimed to the characters of the railcar scene, it was presented a real time fusion matching algorithm with multi-resolution and multi-direction features. First, this algorithm preprocessed images to depress the influences of disturbs and noises with the method of fast fuzzy edge variation. Second, it utilized multi-resolution images to respectively extract the characteristics of texture and edge invariant moments. This way reduces computing complex and increases anti-distortion capability. Then it adopted normalization product-correlation of Maxima method distance measurement and fuzzy mathematics technique to realize fusion matching effectively. Simulation results indicate that this algorithm possesses high matching rate in real time and interference immunity.