摘要
The wavelet transform is developed to identify the differentphases in a fermentation process. In this method, the wavelettransform modulus maxima are used to estimate the local maximumpoints of the second derivative of the growth curve in order toclassify the different phases of fermentation process. Moreover, themethod can effectively get rid of noise from the signal, making useof the different characters showed by signal and noise in the wavelettransform modulus maxima. Compared with neural network modeling, thepresented method needs less quantity of information and calculation.The results of experiments show that this method is effective.
The wavelet transform is developed to identify the different phases in a fermentation process.In this method,the wavelet transform modulus maxima are used to estimate the local maximum points of the second derivative of the growth curve in order to classify the different phases of fermentation process.Moreover,the method can effectively get rid of noise from the signal,making use of the different chacters showed by signal and nose in the wavelet transform modulus maxima.Compared with neural network modeling,the presented method needs less quantity of information and calculation.The results of experiments show that this method is effective.
基金
Supported by the Natural Science Foundation of Shandong Province(Q99B01).