NMR principle analysis based object detection for intelligent measurement of crude oil moisture content

摘要

Tracking crude oil moisture content in real time is critical in the oil refinery production process. Traditional methods are time-consuming, low precision and lack of mechanism analysis. Aiming to solve these problems, we present an object detection method of crude oil moisture content based on NMR principle and deep learning, which named convex region search. It realizes the location of water information in the crude oil ¹H NMR spectrum and classification of all the sample points. Then corresponding regression models are established based on the captured features of disparate categories. The experimental results show that the proposed method can not only locate the position of water peak in NMR spectrum quickly and precisely, but also predict the moisture content value accurately.