洪涝模拟仿真的实现方法主要有两种：一种是基于水动力学的洪水演进模型；还有一种是基于DEM的洪水淹没分析。详细分析例如以下：
我是GIS从业者，从我们的专业角度出发，选择基于DEM的洪水淹没分析来做洪涝的模拟仿真。而基于DEM的洪水淹没分析方法主要分为有源淹没和无源淹没。
本篇博客採用有源淹没算法实现洪涝的模拟，算法为八领域种子扩散算法。
採用C#版本号GDAL编写了FloodSimulation类，以下给出所有源码：
class FloodSimulation
{
#region 类成员变量
//点结构体
public struct Point
{
public int X; //行号
public int Y; //列号
public int Elevation; //像素值(高程值)
public bool IsFlooded; //淹没标记
};
private bool[,] IsFlood; //淹没区域标记二维数组，用于标记淹没栅格
private List<Point> m_FloodBufferList; //淹没缓冲区堆栈
public Dataset m_DEMDataSet; //DEM数据集
public Dataset m_FloodSimulatedDataSet; //洪涝淹没范围数据集
public int m_XSize; //数据X方向栅格个数
public int m_YSize; //数据Y方向栅格个数
public OSGeo.GDAL.Driver driver; //影像格式驱动
public int[] m_FloodBuffer; //填充缓冲区（洪涝淹没范围）
public int[] m_DEMdataBuffer; //DEM数据（存储高程值）
public double m_AreaFlooded; //水面面积
public double m_WaterVolume; //淹没水体体积
/* 这里的GeoTransform(影像坐标变换參数)的定义是：通过像素所在的行列值得到其左上角点空间坐标的运算參数
比如:某图像上(P,L)点左上角的实际空间坐标为：
Xp = GeoTransform[0] + P * GeoTransform[1] + L * GeoTransform[2];
Yp = GeoTransform[3] + P * GeoTransform[4] + L * GeoTransform[5]; */
public double[] m_adfGeoTransform;
#endregion
//构造函数
public FloodSimulation()
{
m_adfGeoTransform = new double[6];
m_FloodBufferList = new List<Point>();
}
/// <summary>
/// 载入淹没区DEM,并创建淹没范围影像
/// </summary>
/// <param name="m_DEMFilePath">DEM文件路径</param>
/// <returns></returns>
public void loadDataSet(string m_DEMFilePath)
{
//读取DEM数据
m_DEMDataSet = Gdal.Open(m_DEMFilePath, Access.GA_ReadOnly);
m_XSize = m_DEMDataSet.RasterXSize;
m_YSize = m_DEMDataSet.RasterYSize;
//获取影像坐标转换參数
m_DEMDataSet.GetGeoTransform(m_adfGeoTransform);
//读取DEM数据到内存中
Band m_DEMBand = m_DEMDataSet.GetRasterBand(1); //获取第一个波段
m_DEMdataBuffer = new int [m_XSize * m_YSize];
m_DEMBand.ReadRaster(0, 0, m_XSize, m_YSize, m_DEMdataBuffer, m_XSize, m_YSize, 0, 0);
//淹没范围填充缓冲区
m_FloodBuffer = new int[m_XSize * m_YSize];
IsFlood=new bool[m_XSize,m_YSize];
//初始化二维淹没区bool数组
for (int i = 0; i < m_XSize; i++)
{
for (int j = 0; j < m_YSize; j++)
{
IsFlood[i, j] = false;
}
}
//创建洪涝淹没范围影像
string m_FloodImagePath = System.IO.Path.GetDirectoryName(System.Windows.Forms.Application.ExecutablePath) + "\FloodSimulation\FloodedRegion.tif";
if (System.IO.File.Exists(m_FloodImagePath))
{
System.IO.File.Delete(m_FloodImagePath);
}
//在GDAL中创建影像,先须要明白待创建影像的格式,并获取到该影像格式的驱动
driver = Gdal.GetDriverByName("GTiff");
//调用Creat函数创建影像
// m_FloodSimulatedDataSet = driver.CreateCopy(m_FloodImagePath, m_DEMDataSet, 1, null, null, null);
m_FloodSimulatedDataSet = driver.Create(m_FloodImagePath, m_XSize, m_YSize, 1, DataType.GDT_Float32, null);
//设置影像属性
m_FloodSimulatedDataSet.SetGeoTransform(m_adfGeoTransform); //影像转换參数
m_FloodSimulatedDataSet.SetProjection(m_DEMDataSet.GetProjectionRef()); //投影
//m_FloodSimulatedDataSet.WriteRaster(0, 0, m_XSize, m_YSize, m_FloodBuffer, m_XSize, m_YSize, 1, null, 0, 0, 0);
//输出影像
m_FloodSimulatedDataSet.GetRasterBand(1).WriteRaster(0, 0, m_XSize, m_YSize, m_FloodBuffer, m_XSize, m_YSize, 0, 0);
m_FloodSimulatedDataSet.GetRasterBand(1).FlushCache();
m_FloodSimulatedDataSet.FlushCache();
}
/// <summary>
/// 种子扩散算法淹没分析
/// </summary>
/// <param name="m_Lat">种子点纬度</param>
/// <param name="m_Log">种子点经度</param>
/// <param name="m_FloodLevel">淹没水位</param>
public void FloodFill8Direct(double m_Lat,double m_Log,double m_FloodLevel)
{
//首先依据种子点经纬度获取其所在行列号
Point pFloodSourcePoint = new Point();
int x, y;
geoToImageSpace(m_adfGeoTransform, m_Log, m_Lat, out x, out y);
pFloodSourcePoint.X = x;
pFloodSourcePoint.Y = y;
//获取种子点高程值
pFloodSourcePoint.Elevation = GetElevation(pFloodSourcePoint);
m_FloodBufferList.Add(pFloodSourcePoint);
//推断堆栈中时候还有未被淹没点，如有继续搜索。没有则淹没分析结束。
while (m_FloodBufferList.Count!=0)
{
Point pFloodSourcePoint_temp = m_FloodBufferList[0];
int rowX = pFloodSourcePoint_temp.X;
int colmY = pFloodSourcePoint_temp.Y;
//标记可淹没,并从淹没堆栈中移出
IsFlood[rowX, colmY] = true;
m_FloodBuffer[getIndex(pFloodSourcePoint_temp)] = 1;
m_FloodBufferList.RemoveAt(0);
//向中心栅格单元的8个临近方向搜索连通域
for (int i = rowX - 1; i < rowX + 2; i++)
{
for (int j = colmY - 1; j < colmY + 2; j++)
{
//推断是否到达栅格边界
if (i<=m_XSize&&j<=m_YSize)
{
Point temp_point = new Point();
temp_point.X = i;
temp_point.Y = j;
temp_point.Elevation = GetElevation(temp_point);
//搜索能够淹没且未被标记的栅格单元
if ((temp_point.Elevation<m_FloodLevel||temp_point.Elevation <= pFloodSourcePoint_temp.Elevation) && IsFlood[temp_point.X, temp_point.Y] == false)
{
//将符合条件的栅格单元增加堆栈，标记为淹没，避免反复运算
m_FloodBufferList.Add(temp_point);
IsFlood[temp_point.X, temp_point.Y] = true;
m_FloodBuffer[getIndex(temp_point)] = 1;
}
}
}
}
}
//统计淹没网格数
int count = 0;
for (int i = 0; i < m_XSize; i++)
{
for (int j = 0; j < m_YSize; j++)
{
if (IsFlood[i,j]==true)
{
count++;
}
}
}
}
/// <summary>
/// 输出洪涝淹没范围图
/// </summary>
public void OutPutFloodRegion()
{
m_FloodSimulatedDataSet.GetRasterBand(1).WriteRaster(0, 0, m_XSize, m_YSize, m_FloodBuffer, m_XSize, m_YSize, 0, 0);
// m_FloodSimulatedDataSet.WriteRaster(0, 0, m_XSize, m_YSize, m_FloodBuffer, m_XSize, m_YSize, 1, null, 0, 0, 0);
m_FloodSimulatedDataSet.GetRasterBand(1).FlushCache();
m_FloodSimulatedDataSet.FlushCache();
}
// public void OutPutFloodedInfo()
// {
// }
/// <summary>
/// 获取第x行第y列栅格索引
/// </summary>
/// <param name="point">栅格点</param>
/// <returns>该点在DEM内存数组中的索引</returns>
private int getIndex(Point point)
{
return point.Y* m_XSize + point.X;
}
/// <summary>
/// 获取高程值
/// </summary>
/// <param name="m_point">栅格点</param>
/// <returns>高程值</returns>
private int GetElevation(Point m_point)
{
return m_DEMdataBuffer[getIndex(m_point)];
}
/// <summary>
/// 从像素空间转换到地理空间
/// </summary>
/// <param name="adfGeoTransform">影像坐标变换參数</param>
/// <param name="pixel">像素所在行</param>
/// <param name="line">像素所在列</param>
/// <param name="x">X</param>
/// <param name="y">Y</param>
public void imageToGeoSpace(double[] m_GeoTransform, int pixel, int line, out double X, out double Y)
{
X = m_GeoTransform[0] + pixel * m_GeoTransform[1] + line * m_GeoTransform[2];
Y = m_GeoTransform[3] + pixel * m_GeoTransform[4] + line * m_GeoTransform[5];
}
/// <summary>
/// 从地理空间转换到像素空间
/// </summary>
/// <param name="adfGeoTransform">影像坐标变化參数</param>
/// <param name="x">X(经度)</param>
/// <param name="y">Y（纬度）</param>
/// <param name="pixel">像素所在行</param>
/// <param name="line">像素所在列</param>
public void geoToImageSpace(double[] m_GeoTransform, double x, double y, out int pixel, out int line)
{
line = (int)((y * m_GeoTransform[1] - x * m_GeoTransform[4] + m_GeoTransform[0] * m_GeoTransform[4] - m_GeoTransform[3] * m_GeoTransform[1]) / (m_GeoTransform[5] * m_GeoTransform[1] - m_GeoTransform[2] * m_GeoTransform[4]));
pixel = (int)((x - m_GeoTransform[0] - line * m_GeoTransform[2]) / m_GeoTransform[1]);
}
}
模拟结果在ArcGlobe中的显示效果图：
欢迎大家留言交流。