Remote Sensing Detection Of Lakeside Wetland Changes In Middle And High Latitudes Of Asia

The special natural ecosystem formed by the interaction between water and land,wetland,is one of the three major natural ecosystems in the world.As a natural transition zone between land and water,wetland ecosystems are easily affected by surrounding environmental factors such as hydrological fluctuations.An in-depth understanding of wetland changes and drivers of wetland change is essential and important for maintaining the balance and benign development of wetland ecosystems.In this paper,Landsat series remote sensing image data were used to detect the changes of lakeside wetlands in five typical regions of Asia.The Principal Component Analysis(PCA)and multi-exponential differential PCA methods were compared in the multiindex principal component analysis.Five algorithms for interannual wetland change detection: 1.Superpixel segmentation KT transform(Kauth-Thomas Transformation,K-T transform)method;2.Superpixel segmentation KT clustering difference method;3.Superpixel Segmentation exponential difference method;4.Superpixel segmentation window correlation analysis method;5.Support Vector Machine(SVM)postclassification comparison method(for verification method)algorithm in two verification area data detection,and evaluation detection The accuracy of the results,by verifying the accuracy of the classification results of different data sets,determine the classification scheme.Using the determined classification scheme,the changes of lakeside wetlands in five research areas were detected,combined with the driving factors such as temperature,rainfall,human activities(cultivated land,bare land change),and the trend was summarized and analyzed.The change detection algorithm and wetland detection for Asian lakeside wetlands were used.And the driving force analysis research provides a certain reference.The main results of this paper are as follows:(1)In the simple linear iterative clustering(SLIC)superpixel segmentation algorithm,when the scale is set to 30,the number of iterations is 10,and the classification result is optimal.In the annual variation detection algorithm,the differential principal component method is superior to the principal component difference method.Among the five algorithms for interannual variation detection,the super-pixel segmentation K-T clustering difference method has the highest accuracy,that is,the classified scheme is better than the remaining algorithms of 0.69 and 0.66,respectively.(2)By comparing the SLIC segmentation image classification results with the unsegmented results,it is concluded that the SLIC segmentation image classification results are better than the unsegmented classification results.In the data after SLIC segmentation,in the three-index,KT conversion,time-phase data mean,standard deviation,and median superimposed classification results,the time-phase data mean,standard deviation,median superimposed classification results and classification product data precision comparison,overall The accuracy is up to 0.9,the Kappa coefficient is 0.84,and the accuracy of the classification result is optimal.The objectoriented two algorithms are compared with the classification results of the watershed and SLIC algorithm and the accuracy of the product data.The accuracy of the classification result of the SLIC algorithm is slightly higher than that of the watershed algorithm.(3)From 1990 to 2015,in the selected mid-high latitude lakeside wetlands in Asia,except for the Oroghu lakeside wetland,the area of the four lakeside wetlands basically showed an increasing trend.Compared with the total wetland area in 1990,Wuliangsuhai Lakeside wetland The increase was at least 3%,and the largest increase was 17.49% of the Salikameishihu Lakeside Wetland,with an average increase of 11.34% of the total 1990 wetland area.The area of Oroghu Lakeside Wetland increased from 1990 to 1995,and the area decreased from 1995 to 2005,mainly due to the drying up of the lake.From 2005 to 2015,the increase in wetland area gradually returned to the 1990 state.Xingkai Lake Lakeside Wetland,from 1990 to 2015,the water body change area of Xingkai Lake is about 85,000 hectares.Before 1995,the area of natural wetland was larger than that of constructed wetland.After 1995,the area of constructed wetland increased,the area of natural wetland decreased,and the area of constructed wetland began to be larger than that of natural wetland.After that,constructed wetland and natural wetland remained relatively stable.From 1990 to 2015,the total area of water bodies,natural wetlands and constructed wetlands increased.The Wuliangsuhai Lakeside wetland,water area and wetland area showed a negative correlation between 1990 and 2010,and was positively correlated from 2010 to 2015.In 2005,the wetland area reached a maximum and the water area was reduced to a minimum.From 2005 to 2010,the wetland area and water area changed positively.In 2015,the wetland water body changed negatively.The Oroughu Lakeside Wetland,as a whole,has a negative change in water area and wetland area before 2000.From 2005 to 2010,there was a positive change,and after 2010,it was a negative change.In 1995,it reached a maximum of 17,127 hectares,and then began to decrease.In 2005,it reached a minimum of 4,380.12 hectares.It began to increase until 2015,reaching 13,335.12 hectares.From 1990 to 2016,the area of water bodies and wetlands in Ala Lake did not change much.The wetland area increased from 1990 to 1995.The wetland area changed little from 1995 to 2016,and it was basically maintained at around 120,000 hectares.The water area and wetland area showed an increasing trend,but there was a slight decrease from 2011 to 2016.From 1990 to 2015,the water area of Saleikameishihu gradually increased;the area of wetland increased from 1990 to 1995,reached its maximum value around 1995,reached a minimum around 2005,and then began to increase.The total area of wetlands increased first,reaching a maximum in 1995,decreasing between 1995 and 2005,reaching a minimum in 2005,and gradually increasing the wetland area after 2005.(4)Areas dominated by natural factors: Through the analysis of digital elevations and rivers,and the distribution of glaciers,it can be seen that the Oroghu and Saleikameishihu wetland areas are glacial recharge-type lakeside wetlands,mainly relying on glacial meltwater and snowmelt replenishment.Therefore,the change of wetland is relatively close to the temperature change.The increase in temperature leads to an increase in the ablation area of the glaciers,and the flow of snow-melting water increases,which in turn increases the amount of water supplied to the lakeside wetlands.According to the driving force,it is divided into natural factors;human factors and natural influence factors.The area dominated by natural and human factors: It can be seen that the area of natural wetland and human activities in Xingkai Lake,Wuliangsuhai and Alahu Lakes are negatively changing.The change of wetland area is mainly affected by factors such as cultivated land and reclamation.The increase in human activities such as cultivation and reclamation has reduced the area of natural wetlands.Wuliangsuhai is mainly dominated by economic crops,reed wetlands,while the area of cultivated land is also increasing.Changes in cultivated land cannot reflect the impact of farming human factors on wetland area.Therefore,the change of bare land represents human factors,and it is found that the area of bare land and wetland changes in the opposite direction,the bare land reduces the wet ground,and the bare land area increases the wetland area.