Study On Suspended Sediment Concentration Measurement Model Based On Image Multi Feature Fusion

Suspended sediment refers to the suspended sediment moving with the flow.For a long time,people have been trying to use various methods to measure the suspended sediment concentration(SSC).However,there are still some problems in the existing suspended solids detection methods,such as effectively obtaining sediment concentration information,accurately extracting relevant features,and reasonably constructing multi-functional data fusion model.At present,there is not a complete set of SSC measurement and processing system based on image multi feature fusion.In this paper,the method of measuring SSC by digital image is studied.On the basis of overcoming the environmental interference,the real-time online measurement system of SSC is studied,which provides a reasonable theoretical basis for real-time on-line measurement of river sediment concentration.In the online measurement system of SSC,in addition to selecting reasonable image features,it is also necessary to select the appropriate measurement model,so as to obtain more accurate information value of sediment concentration.With the development of related technologies,multi feature fusion has gradually become one of the directions of SSC measurement.Multivariate linear fusion,probabilistic neural network and convolutional neural network also contribute to the measurement of SSC.Therefore,this paper proposes a method of suspended sediment concentration measurement based on image multi feature fusion,in order to further improve the accuracy of suspended sediment concentration.The main research contents and innovations of this paper include:1.This paper discusses the method and principle of image feature extraction of SSC,summarizes the existing methods of SSC measurement,analyzes the feasibility of SSC measurement based on image multi feature data fusion by analyzing the existing methods,and improves some algorithms.2.Multi feature fusion model,Kalman+probabilistic neural network fusion model and convolution neural network+support vector machine fusion model are proposed and applied to the measurement of SSC.Firstly,the images of different SSC captured by HD cameras represent different SSC values;secondly,the SSC values of the test set images are measured by combining the models with the training set information;finally,the best measurement method of SSC is determined by error comparison.3.The software and hardware platform of SSC measurement based on image multi feature fusion is designed.The image acquisition module and information communication module are used for data acquisition and transmission,and the measurement visualization is realized by MATLAB software.Finally,this paper applies the above research results to the actual Madu project of the Yellow River,verifies the practicability and usability of the research content,and provides technical support for soil and water conservation and environmental governance of the Yellow River in the future.