| Mangrove ecosystem is an important ecosystem in the coastal zone,which has the functions of maintaining biodiversity,promoting silting and beach consolidation,purifying water quality and maintaining ecological security in the coastal zone.At the same time,it is of great significance to the global environment and climate.Guangxi Zhuang Autonomous Region,as the province with the largest distribution area of mangroves on the coastline per unit length in China,has unique mangrove wetland landscape,which plays an important role in maintaining biodiversity and the balance of coastal wetland ecosystem,and has important research significance.It is helpful to investigate,utilize and protect mangrove resources to know mangrove species in different areas accurately.At present,the study of mangrove remote sensing identification and monitoring in Guangxi mainly focuses on the investigation of mangrove distribution area in major scale and the study of mangrove biophysical parameters in small scale.It can be seen that there is a lack of large-scale interspecific fine classification of mangroves in Guangxi.Therefore,it is urgent to use high-resolution remote sensing image data to carry out the study of interspecific fine classification of mangroves,so as to provide technical and data support for mangrove management,protection and reconstruction in Guangxi Zhuang Autonomous Region.Based on the multi-source remote sensing data of GF-1,GF-2,GF-5 and DJI,this paper studies the mangrove species in Guangxi.Taking GF-2 data in 2018 as the data source,in the study of mangrove interspecific fine classification method,an on-site sample and segmentation object combination of mangrove interspecific classification method(OBSO)is proposed.Based on the GF-5 hyperspectral data,this paper studies the method of reducing the dimension of hyperspectral data from two aspects of feature extraction and band selection,and proposes the method of feature extraction(SMNF)based on the combination of subspace division and minimum noise separation,and the method of band selection(SOIFID)based on the combination of subspace division,the best index factor and image derivative,and verifies it with mangrove in Maoweihai Sea as an example effectiveness.On the basis of OBSO classification method,taking the 2017-2019 GF-2 PMS multispectral remote sensing image data as the main data source,SVM classifier is used to carry out fine classification and spatial distribution research on typical vegetation of Mangrove Wetland in Guangxi.The main research results and conclusions are as follows:1.Based on GF-2 and DJI data,it is found that the OBSO classification method proposed in this paper has the best classification effect and the overall accuracy is 93.15%.Compared with pixel based classification method,objectoriented classification method is more suitable for high spatial resolution remote sensing image.2.Based on GF-5 and DJI remote sensing data,through the comparative analysis of the classification results between the four dimensionality reduction methods and before and after dimensionality reduction,it is found that automatic subspace division can effectively reduce information redundancy,and the best combination of index factor band dimensionality reduction method and image derivative operation can help to improve the accuracy of hyperspectral mangrove species identification.Among them,the overall accuracy of the classification results of the SOIFID combined method is the highest,94.30%,which is 3.94% higher than that of SMNF method.It can be seen that the band selection method is more suitable than the feature extraction method for the fine classification of mangrove species in typical wetland ecosystem.3.Based on the remote sensing data of GF-2,GF-1 and DJI,the comprehensive monitoring results of Mangrove in Guangxi are as follows: 1)As of November 2019,the total area of existing mangroves in Guangxi is 7376 ha.It is mainly distributed in estuaries and harbors,with uniform overall distribution and large local differences.2)Mangroves in Guangxi can be divided into 8 types,namely,Aegiceras corniculatum,Avicennia marina,Rhizophora stylosa,Sonneratia apetala,Kandelia candel,Bruguiera gymnorrhiza,Acanthus ilicifolius and a salt marsh herbaceous plant Cyperus malaccensis.As the dominant species of Mangrove in Guangxi,the area of Aegiceras corniculatum and Avicennia marina is 90.74% of the total area of Mangrove in Guangxi.3)Aegiceras corniculatum and Avicennia marina are evenly distributed in the coastal zone of Guangxi,but they are quite different in some parts.As alien species,Sonneratia apetala is mainly distributed in the north coast of Maoweihai Sea and the east side of Nanliujiang River Estuary of Lianzhou Bay.Large area of pure growth area of Rhizophora stylosa exists in Shankou Mangrove National conservation area.The distribution area of Kandelia candel is very small,only a small piece of pure forest grows very well along the coast of Pearl Harbor.As a late successional species of mangrove,Bruguiera gymnorrhiza is mainly distributed in the inner beach of mangrove,mainly in Pearl Harbor;Acanthus ilicifolius is only distributed in a small area on both sides of the channel affected by tide along Daxinwei village,Jianshan town,Maoweihai Sea.As a salt marsh plant,Cyperus malaccensis is mainly distributed in Maoweihai Sea,and widely distributed in Maoweihai Sea,often mixed with Aegiceras corniculatum.4)The total area of mangroves in Beilun estuary and Shankou Mangrove National Nature Reserve is 1050.14 ha and 635.43 ha respectively,and that in Maoweihai Sea autonomous region is 1557.00 ha.5)Taking Shankou mangrove reserve as a typical area,this paper analyzes the changes of mangrove species.From 2007 to 2017,there were no changes in mangrove species in Shankou nature reserve,which were Avicennia marina,Aegiceras corniculatum and Rhizophora stylosa.Compared with 2007,the total mangrove area in Shankou nature reserve decreased by 46.68 ha,including 116.50 ha in Avicennia marina,68.32 ha in Aegiceras corniculatum,and only 1.40 ha in Rhizophora stylosa area.The main changes of mangrove species were as follows: It’s going to happen in the Dandouhai Sea area.The inner patches of mangrove expand obviously,the degree of fragmentation is aggravating,and the landscape aggregation and connectivity are decreasing.The main causes of mangrove change are extreme weather and human activities. |