Research On Evaluation Of Biofilm Growth Characteristics Based On Hyperspectral Technology

Microbial fuel cell(MFC)is a new generation of green energy technology,which can convert the chemical energy of sewage and waste water into electrical energy.Therefore,microbial fuel cell electricity production performance and efficiency is the key to whether the technology can be widely used.However,there are many factors that affect the performance of the battery,such as the battery structure,the type of the producing bacteria,the pH value of the solution,the temperature and so on.But the most important factor is the growth of microorganisms in the MFC,therefore,research and explore the MFC internal optimum pH,temperature,anode materials is an important task,its purpose is to better ensure the MFC internal microbial growth in the best environment.In this paper,the MFC anode biofilm was used as the research object,the hyperspectral imaging technology was used to obtain the hyperspectral image,and the growth status of MFC anode biofilm was analyzed and monitored in two aspects.The main contents of this thesis include the following.(1)Designed and fabricated two chamber microbial fuel cell reaction device,successful operation of the microbial fuel cell,acquisition in different pH,different temperature,relatively suitable for PH and the temperature of the MFC anode biofilm.(2)Based on the principle of hyperspectral imaging system,the high spectral image acquisition and acquisition system of MFC inner anode biofilm was set up based on hyperspectral imaging system.The spectral images of different PH,different temperature,and biological film at different time of PH=5 and temperature were obtained.(3)The correlation between the characteristic curve of 27 kinds of spectral images and the film thickness is analyzed.The results of the experiments and analysis show that there are strong correlation between the curve features of 5 kinds of spectral images and the thickness of the biofilm.(4)According to the experiment and analysis results we selected five spectral image characteristic curve are established on the basis of the characteristics of a single curve fitting,BP neural network,singular value decomposition(SVD)iterative,and multi feature fusion of curve fitting model and model prediction accuracy and running time from the two aspects of analyzing the advantages and disadvantages of each estimation model.To realize the estimation of the microbial fuel cell anode surface biological film thickness.(5)In this paper,using principal component analysis,feature extraction,image feature;according to different thickness of the biofilm reflection value differences,will reflect the value divided into six parts,each part with different RGB representation;and can reflect the local growth of image feature.(6)This paper from the aspects of biofilm thickness and biofilm image,of the anode biofilm growth were evaluated.The experimental results show that the lactic acid bacteria in PH=5 and temperature is 35 degrees Celsius,the growth of biofilm better;in PH=5 and temperature is 35 degrees Celsius,biological membrane after 84 h almost larger changes.