Active And Passive Remote Sensing Of Inland Waters Cyanobacteria Bloom

In recent years, with the development of socioeconomic and the intensification of human activities in lake basins, too much domestic sewage of industry is discharged into the water. Therefore, Water eutrophication and cyanobacterial bloom are becoming more and more serious. Remote sensing could acquire the information of large area quickly, which is becoming more and more significant in monitoring the location, area and dynamic change of algae. So far, MODIS and HJ-CCD are the main remote sensor monitoring the cyanobacteria bloom in Taihu Lake. It is difficult for optical remote sensing to extract algal bloom information, especially in the cloudy weather conditions which could not meet the acquirement of accuracy and timeliness in algal bloom monitoring using remote sensing. Therefore, it is very important for cyanobacteria bloom monitoring using passive optical remote sensing and active microwave remote sensing-SAR.In this paper, Lake Taihu, the typical large inland lake of China is taken as an example for the monitoring of Cyanobacteria bloom. Through the analysis of optical remote sensing image of cyanobacteria bloom, it was found that there had typical characteristics of transition zones on the boundary of algal bloom. This paper proposed a new concept:gradient complexity and a new method:maximum gradient complexity method for the acquisition of segmentation threshold automatically. Monitoring algal blooms with optical remote sensing is limited by cloud cover. SAR systems can provide images regardless of weather conditions because of their ability to penetrate cloud cover. With the aim of monitoring the cyanobacteria-dominant algal blooms in Lake Taihu in cloudy weather, synthetic aperture radar (SAR) was deployed for monitoring algal blooms. The cyanobacterium Microcystis forms a layer of scum on the water surface that reduces radar backscatter and appears as dark regions in the radar images. Low backscatter may also be associated with other factors such as low wind speeds, resulting in interferences when monitoring algal blooms with SAR data alone. After feature extraction and selection, the dark regions were classified using the support vector machine method, and the ability of SAR in monitoring cyanobacteria bloom in Lake Taihu was assessed. The validity check in monitoring cyanobacteria bloom was conducted with Radarsat-2 SAR imagery and the ground truth; Applying the maximum gradient complexity method to MODIS remote imagery, this paper analysed the spatial distribution characteristics, law and development trend of the inter-annual variation, seasonal and monthly changes of Lake Taihu cyanobacteria bloom in the last ten years. Finally, the application of combined monitoring algal blooms in Lake Taihu with MODIS and SAR was discussed.The main construction of this paper include the following aspects:(1) Propose a concept of remote sensing image segmentation:The complexity of the gradient, propose an automatic extraction method for cyanobacteria bloom monitoring with remote sensing:the maximum gradient complexity method which solve the key problem of automatic processing in cyanobacteria bloom monitoring with remote sensing. (2) Propose a method for cyanobacteria bloom of inland lake monitoring using SAR remote sensing, SVM method was used for bloom pattern recognition. Through the ground synchronous experiment, the physical and chemical properties of cyanobacterial blooms was analyzed in order to reveal mechanism of remote sensing monitoring based on SAR images. (3) Using the maximum gradien-t complexity method, remote sensing data in the last ten years was processed, then the distribution characteristics of the segmentation threshold was acquired. Finally, spatial-temporal variation rules and development trend of cyanobacteria bloom in Taihu Lake was analyzed.The results of this paper could provide a new theory and method for temporal and spatial integration of cyanobacteria bloom dynamics in China inland lakes.