The Research And Application Of The Method Of Extracting Cyanobacteria Bloom In Taihu Lake Based On Sentinel 2A Image

The phenomenon of blue algae bloom in taihu lake has gradually attracted people’s attention.In order to obtain the distribution range of cyanobacteria bloom in taihu lake and study the outbreak rule of cyanobacteria bloom,it is urgent to monitor the cyanobacteria bloom in taihu lake by remote sensing.At present,some studies have been made on the extraction of lake cyanobacteria blooms from low resolution remote sensing images,but there are still some defects in the extraction of lake cyanobacteria blooms from medium resolution remote sensing images.First,in order to extract the bloom information of cyanobacteria more accurately and obtain its area,the calculation method of the canopy degree of mixed pixels of cyanobacteria in medium and high resolution images needs to be further discussed and studied.Second,the medium and high resolution image information is more complex,and the segmentation threshold of cyanobacteria bloom directly and manually selected from a single image has the disadvantages of uncertainty and complexity,which makes some "filamentous" cyanobacteria bloom easy to be wrongly separated from eutrophication water.To solve the above problems,it is of great significance to study the extraction technology of cyanobacteria bloom in taihu lake based on medium-high resolution remote sensing image.Based on the sentry 2A image with a spatial resolution of 10 m,this paper studied the extraction technology and application of cyanobacteria bloom in taihu lake,including the applicability of the extraction model and the segmentation technology of cyanobacteria bloom based on the extraction model.(1)the applicability of three extraction models of chlorophyll a concentration inversion extraction model,NDVI extraction model and FAI extraction model to sentry 2A satellite image was established and analyzed,and the influence of band parameters of sentry 2A image on the applicability of the model was analyzed.Through the verification of measured chlorophyll a concentration data in water containing algae and the stability analysis of the threshold value,the best model for cyanobacteria bloom extraction is FAI(B4/B5/B10).(2)in extraction model based on the index of the image,in view of the high resolution image information more complex problems,the introduction of the gradient image representation "filamentous cyanobacteria blooms,gradient-growing point selection method was put forward and improved the algae canopy of mixed pixels,calculation,eventually improve the APA algorithm(algal blooms like growing point algorithm).The algorithm was applied to extract sentry 2 a image of algae bloom in taihu lake area computation and algal blooms,and use the days with sentry 2 a image of GF2 image with GF6 image to improve the APA algorithm to extract of cyanobacteria bloom area and scope verification,the results show that improved APA algorithm improved in high resolution images to extract the accuracy of cyanobacterial blooms.The validity and accuracy of the improved APA algorithm for extraction of cyanobacteria bloom from taihu lake were proved.(3)based on the above work,a remote sensing monitoring system for cyanobacteria bloom in taihu lake based on sentry 2A image was established to dynamically monitor the cyanobacteria bloom in taihu lake during the three years from 2017 to 2019,and to analyze the spatial and temporal variation rules of cyanobacteria bloom in taihu lake.