针对目前广义维数多重分形谱算法运算量大,难以用于预警识别的现状,提出1阶Minkowski度量的多重分形谱算法,并将该算法用于坦克噪声预警识别试验.该方法根据广义维数定义本质,通过简化测度计算,减少运算量.试验表明,该算法在信噪比较低时,预警识别性能高于目前常用的过零率检测方法.%A generalized multifractal spectrum algorithm based on 1-norm Minkowski measure was proposed to resolve the problem that present multifractal spectrum algorithm was too complicated to be applied in identification for early alarming, the experiment was implemented to prove the validity of the algorithm for tank-radiated noise identification as well. This algorithm reduced the numeration of the generalized multi-fractal spectrum by simplifies the measure according to the definition of generalized dimension. The experiment demonstrates that under low SNR the ability of early alarming identification by this method is higher than zero-pass analyzing method we use today.
展开▼