An Study On Soil Erosion Intensity In Karst Region Based On GF1 Data

The fragile ecological environment in karst region and the unreasonable utilization of human beings caused serious soil erosion and continuous deterioration of the ecological environment in this region.To accurately grasp the information of soil erosion intensity can provide reliable reference data for the comprehensive control of soil and water loss.And the remote sensing estimation method is currently the key technology to grasp regional soil erosion intensity.At present soil erosion research mainly used foreign data with expensive price and small width.But with the successful launch of the Gaofen-1satellite in our country,the data of high resolution,multi-temporal and high width can be acquired successfully.To study Gaofen-1image's influence of extracting soil erosion factor in karst area,the difference of soil erosion intensity and the change rule of the scale effect and soil erosion intensity landscape index,It not only has important significance for the application of high score data in the study of soil erosion intensity,but also can provide reference for the selection of soil erosion research image data and scale effect study in the karst region.This paper uses the Baishuihe small watershed as the research area and selects the high-resolution remote sensing data,namely: GF1-WFV(16m)image,GF1-PMS(8m)image,and fusion(2m)image,based on these images.Soil erosion factors were extracted to further analyze the effects of imagery on vegetation coverage,land use types,and soil erosion intensity.Soil erosion intensity characteristics and scale effects of different images were analyzed.At the same time,changes in soil erosion intensity landscape index and soil erosion analysis were analyzed.Suitable images.The preliminary results obtained are as follows:(1)The extraction of soil erosion factors is greatly influenced by the image resolution.With the increase of image resolution,the coverage of soil erosion factors is more obvious,the more clearly the spatial structure of the image,the smoother the texture,the more detailed information;the amount of information The extraction effect of the area is continuously enhanced and the amount of information is increased.At the same time,the distribution of vegetation coverage information shows a state ofconvergence in the middle;as the resolution increases,the mixed pixels are decomposed and the area of cultivated land increases,and the construction land and forest land area continuously decrease;GF1-WFV The accuracy of vegetation coverage estimation was 63.11%,73.88%,82.13%,GF1-PMS images,and fusion images,respectively.The Kappa coefficients for land use types were 0.6639,0.8122,and 0.8255,respectively;the overall accuracy was 72.6087% 86.0021 and 90.0303%.(2)There is a significant difference in the estimated soil erosion intensity of the image.The GF1-WFV image has the largest area of soil erosion,accounting for60.48% of the total area,with a slight proportion of 6.91%,a moderate proportion of28.04%,and a strong proportion of 11.31%,extremely strong ratio of 7.93% and intensive accounting for 5.61%;fused images had the smallest area of ? ? soil erosion,accounting for 40.52% of the total area,which was reduced by 19.96%relative to GF1-WFV images,and the ratio of mildness was 5.11.The proportions of % and moderate were 23.48%,the strong ratio was 7.52%,the extremely strong ratio was 3.19%,and the violent ratio was 1.22%.(3)The soil erosion intensity grades of GF1-WFV image,GF1-PMS image,and fusion image have obvious similarities in the spatial distribution of the watershed.Moderate erosion and strong erosion are mainly distributed in the middle of the basin with low vegetation coverage and south of the basin.The arable land areas are uniformly distributed in other areas;extremely strong and severe erosion are mainly distributed in the areas along the watershed of the river basin,around the construction land and where the ridge slope is large and the vegetation coverage is low.(4)When the intensity of soil erosion is mild or medium erosion,the influence of slope direction on soil erosion intensity is small,the proportion of soil erosion is basically the same,there is no major change;the intensity is strong erosion,extremely strong erosion,severe erosion At the same time,the difference in the proportion of soil erosion on the northern slope,the northwest slope and the southwest slope was the most obvious.(5)The scale effect of soil erosion intensity on GF1-WFV image,GF1-PMS image,and fusion image shows that,with the increase of image resolution,the proportion of soil erosion intensity in each grade gradually decreases.The greatest extent of change occurred,and the proportion distribution of each grade gradually obeyed the law of normal distribution.The spatial distribution of soil erosion intensity is gradually clear.(6)From the three types of landscape data PD,PAFRAC,LSI,etc.,the severelyeroded landscape has the smallest fragmentation degree and the most moderately eroded landscape fragmentation degree;the soil erosion intensity landscape obtained from the landscape level indices CONTAG,COHESION,AI The fragmentation degree is the largest,the landscape shape is the most complex,the dispersity of the landscape patch is the highest,and the spatial connectivity of the landscape is the highest,while the GF1-WFV image is the smallest.The ensemble obtained from the AI has the highest degree of landscape aggregation in the landscape,and the lowest GF1-WFV image;the SHDI shows that the soil erosion intensity of the fused image does not have an even distribution in the landscape,while the GF1-WFV image shows soil erosion.Intensity patches showed a trend of equalization in the landscape,there were no obvious dominant types in the landscape and the types of plaques were uniformly distributed in the landscape.(7)Comparing the vegetation coverage,land-use types,and soil erosion intensity information of GF1-WFV images,GF1-PMS images,and fusion images,it is concluded that there is a large amount of information loss in the extraction of soil erosion factors by GF1-WFV images.,in the rough space,can not guarantee the accuracy,the fusion image can meet the accuracy requirements,extraction and estimation is better.Therefore,under the premise of meeting the accuracy requirements,GF1-WFV images are not suitable for the study of soil erosion intensity in karst regions,and fusion images are the most suitable image data.