Recognition Method of Petroleum Fluorescence Spectra Based on Convolutional Neural Network

Abstract As an important raw material and petrochemical tool, petroleum not only brings convenience to mankind, but also huge socio-economic cultural value our social development, at the same time it causes a lot of serious damage ecological environment. The identification measurement pollutants has become main tool identify pollution sources, control their protect This paper explores fluorescence spectrum method based on convolutional neural network. Based extensive research this method, simple analysis understanding technology petroleum-related principles are carried out, then summarized according relevant data find out factors that affect recognition, prepare for experiment. feasibility is verified through recognition experiment experimental results show relative error under different concentrations both within range 9%. Through error, can be seen resolution shows downward trend with increase concentration. According above data, when network algorithm used components mixed solution, qualitative completed well. When in solution quantitatively analyzed, there certain error.