Application and Design of the Spectral Identification System Based on Neural Network and Fuzzy Control

Abstract

The spectral identification technology is a spectral basis of qualitative analysis. With the development of pattern recognition, spectral identification technology has become an important tool for rapid detection of medicine, environmental protection, petrochemical and other industries. The neural network nonlinear mapping, adaptive learning, robustness and fault tolerance features, has a wide range of applications in signal processing, knowledge engineering, pattern recognition and other fields. This paper meets the Lambert Beer law of spectral signals for the study, outlines the basic principles of neural networks for pattern recognition, and then according to the specific requirements of the spectrum recognition, multi-feature-based and neural network spectral identification programs, and conducts system design, the establishment of the basic model framework. Finally, an instance of the method is described.