Font Size: a A A

Study On Rapid Recognition Of Marine Microplastics Based On Raman Spectroscopy

Posted on:2022-08-25Degree:MasterType:Thesis
Country:ChinaCandidate:S J YangFull Text:PDF
GTID:2491306572968259Subject:Marine science
Abstract/Summary:PDF Full Text Request
In recent years,due to the extensive use and discharge of plastics,these plastics are broken into microplastics through physical,chemical and biological effects in the environment and gather in the ocean in large numbers,resulting in the accumulation of a large number of microplastics in the ocean,damaging the Marine ecological environment and harming Marine organisms.Microplastics are small in shape,and the identification method is complicated,which takes a long time and takes a lot of time.Studies have found that common micro plastic samples in seawater has obvious characteristics of Raman spectroscopy peak "fingerprint" can be used to quickly identify,micro plastic sample sea and laser Raman detection technology is fast,don’t need a tablet,you can direct detection in aqueous solution or glass material,etc,this thesis will laser Raman detection and recognition algorithm combined with artificial intelligence.The rapid detection technology of seawater microplastics samples was studied.In this paper,based on Raman spectroscopic detection technology,combined with wavelet processing and random forest algorithm,the microplastics in seawater spectroscopic test system was established.Based on the Qt development platform,the upper computer measurement software was designed,and a rapid identification system for seawater microplastics samples was established.A laser Raman spectroscopy measurement system was built independently.Twenty-one kinds of typical microplastics samples were measured by this system,and the Raman spectroscopy database of typical microplastics samples was established.In order to improve the efficiency of spectral recognition,the spectral data were preprocessed by data compression.The data compression points of 64,128,256,512 and 1024 points were compared respectively.The experimental results show that the Raman spectroscopy compression point of microplastics is 512 points,which is the best compression point for efficiency and precision.The results can provide a reference for microplastics Raman data compression in practical engineering applications.Raman spectral identification of microplastics samples was studied by decision tree and random forest algorithms.The results show that the accuracy of cross validation based on the random forest algorithm is always higher than that of the decision tree algorithm for 21 kinds of typical microplastics.In order to further improve the identification accuracy,the key parameters of the model(max_depth,the number of trees in the random forest,and the fold times k)were optimized.The optimized model parameters(max_depth =10,the number of trees =100,k=20)were adopted.The cross validation accuracy of random forest algorithm for microplastics identification can reach 97.24%.The results can provide technical reference for the rapid identification of microplastics in actual seawater.Based on the Qt development environment,the microplastic sample Raman spectroscopy measurement and processing software was developed,and the upper computer software was used to control the microplastic spectroscopy measuring instrument to realize the collection of Raman spectroscopy data.The collected Raman spectroscopy data was optimized online,and the microplastic spectroscopy in the database was compared online with the real-time monitoring microplastic spectroscopy.Based on Raman spectroscopic microplastic detection technology,the actual seawater samples in the surrounding sea area of Yantai were analyzed and tested.Ten kinds of seawater samples were collected from the sea area around Yantai,and the samples were pretreated.The actual tests were carried out by establishing an optimized microplastics identification system.In order to verify the correctness of the identification results,the Raman spectroscopy of the actual seawater microplastics samples were compared with the Raman spectroscopy of the standard samples.The results showed that near the characteristic peaks,the actual seawater samples and the standard samples had similar characteristic peaks.The identification results were further verified by Micro-Fourier Transform Infrared Spectrometer.The results show that the identification technology of Raman test system established in this paper can effectively identify microplastics samples,which can provide a technical reference for the rapid identification of microplastics in actual seawater.
Keywords/Search Tags:microplastics, laser Raman, decision tree, random forest, Spectral measurement software
PDF Full Text Request
Related items