Structure-context based fuzzy neural network approach for automatic target detection

This paper proposes a structure-context based fuzzy neural network (SCBFNN) approach for automatic target detection. Fuzzy neural network methods not only possess advantages as adaptivity, parallelism, robustness, ruggedness, optimality, but integrate advantages as depicting and solving system uncertainty by fuzzy set theory, accordingly, they are powerful tools for image processing and pattern recognition. Use fuzziness measures as objective function of neural network can depict uncertainty of pixel´s category validly so as to optimize image classification by minimizing the objective function. Put information constraint of structure context on neurons´ weighting processing can reduce loss of image information, especially, the rich information comprised by target edges, by which target´s attributes such as profile and shape can be retained validly, and the false detection rate can also be improved prominently. Experiments on remote sensed images of target are executed to validate SCBFNN approach, and the results exhibit that SCBFNN possesses good ability to automatic target detection, simultaneously, possesses valid abilities to eliminating uncertainty and retaining target shape compared with conventional neural network methods. At last, a brief conclusion is given.