基于DEM的坡度坡向分析

技术分享 3年前 (2022-05-15) 0 999+

坡度坡向分析方法

坡度(slope)是地面特定区域高度变化比率的量度。坡度的表示方法有百分比法、度数法、密位法和分数法四种，其中以百分比法和度数法较为常用。本文计算的为坡度百分比数据。如当角度为45度（弧度为π/4）时，高程增量等于水平增量，高程增量百分比为100%。

坡向(aspect)是指地形坡面的朝向。坡向用于识别出从每个像元到其相邻像元方向上值的变化率最大的下坡方向。坡向可以被视为坡度方向。坡向是一个角度，将按照顺时针方向进行测量，角度范围介于 0（正东）到 360（仍是正东）之间，即完整的圆。不具有下坡方向的平坦区域将赋值为-1（arcgis处理时为-1，其他可能为0）。

坡度、坡向计算一般采用拟合曲面法。拟合曲面一般采用二次曲面，即3×3的窗口，如下图所示。每个窗口的中心为一个高程点。图中的中心点e坡度和坡向计算过程如下。

参考链接：

[1]https://blog.csdn.net/zhouxuguang236/article/details/40017219

[2]https://blog.csdn.net/weixin_45561357/article/details/106677574

[3]https://www.cnblogs.com/gispathfinder/p/5790469.html

注意：DEM的空间坐标系一定要为投影坐标系。

ArcGIS坡度坡向分析

打开DEM数据

坡度分析

坡度结果

坡向分析

坡向结果

python-gdal坡度坡向分析

from osgeo import gdal

demfile = r"D:微信公众号坡度坡向N40E117_Albers.tif"

# 获取DEM信息
infoDEM = gdal.Info(demfile)

# 计算坡度
slopfile = r"D:微信公众号坡度坡向N40E117_Albers_gdal_Slope.tif"
slope = gdal.DEMProcessing(slopfile, demfile, "slope", format='GTiff', slopeFormat="percent", zeroForFlat=1, computeEdges=True)

# 计算坡向
aspectfile = r"D:微信公众号坡度坡向N40E117_Albers_gdal_Aspect.tif"
b = gdal.DEMProcessing(aspectfile, demfile, "aspect", format='GTiff', trigonometric=0, zeroForFlat=1, computeEdges=True)

坡度结果

坡向结果

python坡度坡向分析

import gdal
import numpy as np
from scipy import ndimage as nd
from copy import deepcopy

demfile = r"D:微信公众号坡度坡向N40E117_Albers.tif"
slopefile = r"D:微信公众号坡度坡向N40E117_Albers_python_Slope.tif"

#读取DEM数据
ds = gdal.Open(demfile)
cols = ds.RasterXSize
rows = ds.RasterYSize
geo = ds.GetGeoTransform()
proj = ds.GetProjection()
dem_data = ds.ReadAsArray()
data = deepcopy(dem_data).astype(np.float32)
band = ds.GetRasterBand(1)
nodata = band.GetNoDataValue()
data[data == nodata] = np.nan
# data[data<-999]=np.nan
mask = np.isnan(data)
# 将无效值或背景值临近像元填充
if np.sum(mask) > 0:
    ind = nd.distance_transform_edt(mask, return_distances=False, return_indices=True)
    data = data[tuple(ind)]

# 计算坡度
xsize = np.abs(geo[1])
ysize = np.abs(geo[5])
x = ((data[:-2, 2:] - data[:-2, :-2]) + 2 * (data[1:-1, 2:] - data[1:-1, :-2]) + (data[2:, 2:] - data[2:, :-2])) / (8 * xsize)
y = ((data[2:, :-2] - data[:-2, :-2]) + 2 * (data[2:, 1:-1] - data[:-2, 1:-1]) + (data[2:, 2:] - data[:-2, 2:])) / (8 * ysize)
s_data = np.full((rows, cols), -999, dtype=np.float32)
s_data[1:-1, 1:-1] = (np.arctan(np.sqrt((np.power(x, 2) + np.power(y, 2)))))
s_data[1:-1, 1:-1] = np.abs(np.tan(s_data[1:-1, 1:-1])) * 100
s_mask = s_data==-999
# 边缘填充
if np.sum(s_mask) > 0:
    ind = nd.distance_transform_edt(s_mask, return_distances=False, return_indices=True)
    s_data = s_data[tuple(ind)]
# 掩膜
s_data[dem_data==nodata] = -999
# 写出结果
driver = gdal.GetDriverByName("gtiff")
outds = driver.Create(slopefile, cols, rows, 1, gdal.GDT_Float32)
outds.SetGeoTransform(geo)
outds.SetProjection(proj)
outband = outds.GetRasterBand(1)
outband.WriteArray(s_data)
outband.SetNoDataValue(-999)

坡度结果

import gdal
import numpy as np
from scipy import ndimage as nd
from copy import deepcopy

demfile = r"D:微信公众号坡度坡向N40E117_Albers.tif"
aspectfile = r"D:微信公众号坡度坡向N40E117_Albers_python_Aspect.tif"

#读取DEM数据
ds = gdal.Open(demfile)
cols = ds.RasterXSize
rows = ds.RasterYSize
geo = ds.GetGeoTransform()
proj = ds.GetProjection()
dem_data = ds.ReadAsArray()
data = deepcopy(dem_data).astype(np.float32)
band = ds.GetRasterBand(1)
nodata = band.GetNoDataValue()
data[data == nodata] = np.nan
# data[data<-999]=np.nan
mask = np.isnan(data)
# 将无效值或背景值临近像元填充
if np.sum(mask) > 0:
    ind = nd.distance_transform_edt(mask, return_distances=False, return_indices=True)
    data = data[tuple(ind)]

# 计算坡向
xsize = np.abs(geo[1])
ysize = np.abs(geo[5])
x = ((data[:-2, 2:] - data[:-2, :-2]) + 2 * (data[1:-1, 2:] - data[1:-1, :-2]) + (data[2:, 2:] - data[2:, :-2])) / (8 * xsize)
y = ((data[2:, :-2] - data[:-2, :-2]) + 2 * (data[2:, 1:-1] - data[:-2, 1:-1]) + (data[2:, 2:] - data[:-2, 2:])) / (8 * ysize)
a_data = np.full((rows, cols), -999, dtype=np.float32)
a_data[1:-1, 1:-1] = np.arctan2(y, -1 * x) * 57.29578
a_data_ = deepcopy(a_data[1:-1, 1:-1])
a_data[1:-1, 1:-1][a_data_ < 0] = 90 - a_data[1:-1, 1:-1][a_data_ < 0]
a_data[1:-1, 1:-1][a_data_ >90] = 450 - a_data[1:-1, 1:-1][a_data_ > 90]
a_data[1:-1, 1:-1][(a_data_ >= 0) & (a_data_ <= 90)] = 90 - a_data[1:-1, 1:-1][(a_data_ >= 0) & (a_data_ <= 90)]
a_data[1:-1, 1:-1][(x==0.)& (y==0.)] = -1
a_data[1:-1, 1:-1][(x==0.)& (y>0.)] = 0
a_data[1:-1, 1:-1][(x==0.)& (y<0.)] = 180
a_data[1:-1, 1:-1][(x>0.)& (y==0.)] = 90
a_data[1:-1, 1:-1][(x<0.)& (y==0.)] = 270.

# 边缘填充
a_mask = a_data==-999
if np.sum(a_mask) > 0:
    ind = nd.distance_transform_edt(a_mask, return_distances=False, return_indices=True)
    a_data = a_data[tuple(ind)]

# 掩膜
a_data[dem_data==nodata] = -999
# 写出结果
driver = gdal.GetDriverByName("gtiff")
outds = driver.Create(aspectfile, cols, rows, 1, gdal.GDT_Float32)
outds.SetGeoTransform(geo)
outds.SetProjection(proj)
outband = outds.GetRasterBand(1)
outband.WriteArray(a_data)
outband.SetNoDataValue(-999)