基于深度学习算法的硝酸铵溶液析晶点检测系统OA北大核心

An automatic quality detection system for expanded ammonium nitrate aqueous solution was designed based on the on-site conditions of the production line.The accuracy of different deep learning algorithms in determining the crystallization state of ammonium nitrate solution was studied.EfficiencyNet algorithm exhibited the highest accuracy.Effi-ciencyNet algorithm was improved by increasing the number of feature channels in each layer and removing multiple MB-Conv layers in depth.The numbers of parameters were reduced,FLOPs were lowered,and the detection were accelerated.The average error between the automatically measured crystallization point temperature by the system and the manually measured temperature is less than 0.3℃.The results demonstrate that the system can accurately measure the crystalliza-tion temperature and density of ammonium nitrate solution,and automatically generate detection reports for ammonium nitrate solution.At the same time,it can trace and query data,and mark abnormal data,thereby meeting production requirements.