WorldWind学习系列十:RendableObject中的DirectX渲染分析——ImageLayer为例

  学习WW有一段时间了,但是若想开发自己基于WW的插件,必然会遇到RendableObject中的DirectX渲染问题。所有需要渲染绘制的WW三维插件,最终是通过继承RendableObject并实现自己的Initialize()、Update()、Render()方法的。想写自己的Render()方法不是简单的事情,你必然要学习DirectX编程,否则,你连看懂示例中的底层Render()方法都很难,谈何开发自己的插件。

  为了突破DirectX编程对我学习WW插件的阻挠,我“快餐式”地突击学习了DirectX,对基本流程和基本原理有了一定认识。今天下午我也对BMNG插件中RendableObject子类——ImageLayer为例进行了分析学习。现在与大家分享一下。

注:请务必先学过Direct3D学习(资料收集),否则看下面的内容无异于浪费你宝贵的时间。

  Image Layer重载实现的Initialize()方法。

Initialize方法代码
         ///   <summary>
        
///  Layer initialization code
        
///   </summary>
         public   override   void  Initialize(DrawArgs drawArgs)
        {
            
try
            {
        //获取绘制设备对象
                
this .device  =  drawArgs.device;

                
if (_imagePath  ==   null   &&  _imageUrl  !=   null   &&  _imageUrl.ToLower().StartsWith( " http:// " ))
                {
           //根据URL地址,在本地构建同样层次的文件夹路径。 这里有个知识点 getFilePathFromUrl()方法,自己定位学习
                    _imagePath  =  getFilePathFromUrl(_imageUrl);
                    
                }

                FileInfo imageFileInfo 
=   null ;
                
if (_imagePath  !=   null )
                    imageFileInfo 
=   new  FileInfo(_imagePath);

                
if (downloadThread  !=   null   &&  downloadThread.IsAlive)
                    
return ;

                
if (_imagePath  !=   null   &&  
                    cacheExpiration 
!=  TimeSpan.MaxValue  &&
                    cacheExpiration.TotalMilliseconds 
>   0   &&
                    _imageUrl.ToLower().StartsWith(
" http:// " &&
                    imageFileInfo 
!=   null   &&  
                    imageFileInfo.Exists 
&&
                    imageFileInfo.LastWriteTime 
<  System.DateTime.Now  -  cacheExpiration)
                {
                    
// attempt to redownload it
                  //启用新线程,下载影像图片。我稍后分析DownloadImage()方法,是重点之一
                    downloadThread  =   new  Thread( new  ThreadStart(DownloadImage));
                    downloadThread.Name 
=   " ImageLayer.DownloadImage " ;
                    
                    downloadThread.IsBackground 
=   true ;
                    downloadThread.Start();

                    
return ;
                }

                
if (m_TextureStream  !=   null )
                {
            //从文件流中,更新Texture(纹理),重点分析
                    UpdateTexture(m_TextureStream, m_TransparentColor);
                    verticalExaggeration 
=  World.Settings.VerticalExaggeration;
            //创建Mesh,稍后重点分析
                    CreateMesh();
                    isInitialized 
=   true ;
                    
return ;
                }
                
else   if (imageFileInfo  !=   null   &&  imageFileInfo.Exists)
                {
          //从文件路径中更新纹理
                    UpdateTexture(_imagePath);
                    verticalExaggeration 
=  World.Settings.VerticalExaggeration;
          //创建格网Mesh,为了获取Vertex集合。
                    CreateMesh();
                    isInitialized 
=   true ;
                    
return ;
                }

                
if (_imageUrl  !=   null   &&  _imageUrl.ToLower().StartsWith( " http:// " ))
                {
                    
// download it...
                    downloadThread  =   new  Thread( new  ThreadStart(DownloadImage));
                    downloadThread.Name 
=   " ImageLayer.DownloadImage " ;
                    
                    downloadThread.IsBackground 
=   true ;
                    downloadThread.Start();

                    
return ;
                }

                
//  No image available
                Dispose();
                isOn 
=   false ;
                
return ;
            }
            
catch
            {
            }
        }

 

398行DownloadImage()方法中关键代码分析学习

关键代码
         //WW中负责从Web下载文件(影像和配置文件)的类WebDownload 。(写的很好,可以被我们重用。这里是个知识点,感兴趣地自己学习)        
    using (WebDownload downloadReq  =   new  WebDownload( this ._imageUrl))
                {
                    downloadReq.ProgressCallback 
+=   new  DownloadProgressHandler(UpdateDownloadProgress);
                    
string  filePath  =  getFilePathFromUrl(_imageUrl);
                    
                    
if (_imagePath == null )
                    {
                        
//  Download to RAM
              //将下载的文件放到内存中,然后ImageHelper类从流对象加载转变为纹理对象。
                        downloadReq.DownloadMemory();
                        texture 
=  ImageHelper.LoadTexture(downloadReq.ContentStream);
                    }
                    
else
                    {
                     // 将加载的文件保存到——imagePath中,然后加载更新纹理
                        downloadReq.DownloadFile(_imagePath);
                        
UpdateTexture(_imagePath);
                    }
             //创建Mesh(我是自学DirectX,搞不清Mesh该翻译成啥,但我知道是干啥用的)
                    CreateMesh();
                    isInitialized 
=   true ;
                }

 

ImageHelper中加载纹理的函数真正实现是191行public static Texture LoadTexture(Stream textureStream, int colorKey)。

从流中加载纹理代码
public   static  Texture LoadTexture(Stream textureStream,  int  colorKey)
        {
            
try
            {
//Direct3D的TextureLoader类的FromStream,从流对象中创建纹理
                Texture texture 
=  TextureLoader.FromStream(DrawArgs.Device, textureStream,  0 0 ,
                    
1 , Usage.None, World.Settings.TextureFormat, Pool.Managed, Filter.Box, Filter.Box, colorKey);

                
return  texture;
            }
            
catch  (Microsoft.DirectX.Direct3D.InvalidDataException)
            {
            }

            
try
            {
                
//  DirectX failed to load the file, try GDI+
                
//  Additional formats supported by GDI+: GIF, TIFF
                
//  TODO: Support color keying.  See: System.Drawing.Imaging.ImageAttributes
          如果DirectX创建纹理失败,则使用GDI+通过创建Image来实现创建纹理
                 using  (Bitmap image  =  (Bitmap)Image.FromStream(textureStream))
                {
           //学过DirectX,下面创建纹理的方法在Demo中很常见(一定要先学DirectX编程)
                    Texture texture 
=   new  Texture(DrawArgs.Device, image, Usage.None, Pool.Managed);
                    
return  texture;
                }
            }
            
catch
            {
                
throw   new  Microsoft.DirectX.Direct3D.InvalidDataException( " Error reading image stream. " );
            }
        }

 

通过文件路径中加载纹理,最终也是调用上面的方法,只是先通过文件路径,将文件打开创建了流对象。请看ImageHelper.cs中52行public static Texture LoadTexture(string textureFileName, int colorKey)。

  public static Texture LoadTexture(string textureFileName, int colorKey)
  {
   try
   {

//将文件打开创建了流对象,从影像流中创建纹理
    using (Stream imageStream = File.OpenRead(textureFileName))
     return LoadTexture(imageStream, colorKey);
   }
   catch
   {
    throw new Microsoft.DirectX.Direct3D.InvalidDataException(string.Format("Error reading image file '{0}'.", textureFileName));
   }
  }

ImageLayer.cs中的956行代码UpdateTexture(string fileName)更新纹理方法,关键代码Texture newTexture = ImageHelper.LoadTexture(fileName);就是从文件路径中创新新的纹理对象,上面已经分析过了,不再赘述。

CreateMesh()方法分析

该方法是最关键的,它创建了Device最终渲染三角面列表所需要的点集合,即 protected CustomVertex.PositionNormalTextured[] vertices;其中CustomVertex.PositionNormalTextured是点类型,就是说所要绘制的点是带纹理的带法线的。(如果要问为什么要这样,那请你先学习好DirectX编程)

 

构建Mesh代码
     protected   virtual   void  CreateMesh()
        {
            
int  upperBound  =  meshPointCount  -   1 ;
            
float  scaleFactor  =  ( float ) 1 / upperBound;
            
double  latrange  =  Math.Abs(maxLat  -  minLat);
            
double  lonrange;
            
if (minLon  <  maxLon)
                lonrange 
=  maxLon  -  minLon;
            
else
                lonrange 
=   360.0f   +  maxLon  -  minLon;

            
int  opacityColor  =  System.Drawing.Color.FromArgb( this .m_opacity, 0 , 0 , 0 ).ToArgb();
           //创建点集合对象vertices,大小为64*64个
            vertices 
=   new  CustomVertex.PositionNormalTextured[meshPointCount  *  meshPointCount];
            
for ( int  i  =   0 ; i  <  meshPointCount; i ++ )
            {
                
for ( int  j  =   0 ; j  <  meshPointCount; j ++ )
                {    
                    
double  height  =   0 ;
                    
if ( this ._terrainAccessor  !=   null )
                        height 
=   this .verticalExaggeration  *   this ._terrainAccessor.GetElevationAt(
                            (
double )maxLat  -  scaleFactor  *  latrange  *  i,
                            (
double )minLon  +  scaleFactor  *  lonrange  *  j,
                            (
double )upperBound  /  latrange);
          
//将空间坐标转换为矢量坐标(即经纬度和半径,转为X\Y\Z)
                    
Vector3 pos  =  MathEngine.SphericalToCartesian( 
                        maxLat  -  scaleFactor * latrange * i,
                        minLon  +  scaleFactor * lonrange * j, 
                        layerRadius  +  height);
                    
//获取点的X,Y,Z Tu Tv
                    vertices[i
* meshPointCount  +  j].X  =  pos.X;
                    vertices[i
* meshPointCount  +  j].Y  =  pos.Y;
                    vertices[i
* meshPointCount  +  j].Z  =  pos.Z;
                    
                    vertices[i
* meshPointCount  +  j].Tu  =  j * scaleFactor;
                    vertices[i
* meshPointCount  +  j].Tv  =  i * scaleFactor;
                
//     vertices[i*meshPointCount + j].Color = opacityColor;
                }
            }
       
 //下面的代码是将上面的vertices集合,构建成6个一组。(注:DirectX中绘制面是绘制两个三角形,是要存六个点的
            indices 
=   new   short [ 2   *  upperBound  *  upperBound  *   3 ];
            
for ( int  i  =   0 ; i  <  upperBound; i ++ )
            {
                
for ( int  j  =   0 ; j  <  upperBound; j ++ )
                {
                    indices[(
2 * 3 * i * upperBound)  +   6 * j]  =  ( short )(i * meshPointCount  +  j);
                    indices[(
2 * 3 * i * upperBound)  +   6 * +   1 =  ( short )((i + 1 ) * meshPointCount  +  j);
                    indices[(
2 * 3 * i * upperBound)  +   6 * +   2 =  ( short )(i * meshPointCount  +  j + 1 );
    
                    indices[(
2 * 3 * i * upperBound)  +   6 * +   3 =  ( short )(i * meshPointCount  +  j + 1 );
                    indices[(
2 * 3 * i * upperBound)  +   6 * +   4 =  ( short )((i + 1 ) * meshPointCount  +  j);
                    indices[(
2 * 3 * i * upperBound)  +   6 * +   5 =  ( short )((i + 1 ) * meshPointCount  +  j + 1 );
                }
            }
          
//计算面的法向量(是重点,高等数学好的话,可以自己学习一下,看看如何求的单位法向量),不想看,调用该方法就行,知道干啥用和怎么用就行啦
            calculate_normals(
ref  vertices, indices);
        }

 Render()方法分析

 ImageLayer.cs中651行

 

Render()代码
         public   override   void  Render(DrawArgs drawArgs)
        {
            
if (downloadThread  !=   null   &&  downloadThread.IsAlive)
//DirectX知识点,可以学习一下
               
 RenderProgress(drawArgs );

            
if ( ! this .isInitialized)
                
return ;

            
try
            {
                
                
if (texture  ==   null   ||  m_SurfaceImage  !=   null )
                    
return ;
        
//这里是DirectX渲染的关键点之一, 设置渲染纹理。
                drawArgs.device.SetTexture(
0this .texture);

                
if ( this ._disableZbuffer)
                {
                    
if (drawArgs.device.RenderState.ZBufferEnable)
                        drawArgs.device.RenderState.ZBufferEnable 
=   false ;
                }
                
else
                {
                    
if ( ! drawArgs.device.RenderState.ZBufferEnable)
                        drawArgs.device.RenderState.ZBufferEnable 
=   true ;
                }

                drawArgs.device.RenderState.ZBufferEnable 
=   true ;
                drawArgs.device.Clear(ClearFlags.ZBuffer, 
0 1.0f 0 );
         
//设置device.Transform.World
                drawArgs.device.Transform.World  =  Matrix.Translation(
                        (
float ) - drawArgs.WorldCamera.ReferenceCenter.X,
                        (
float ) - drawArgs.WorldCamera.ReferenceCenter.Y,
                        (
float ) - drawArgs.WorldCamera.ReferenceCenter.Z
                        );
          //设置顶点格式为CustomVertex.PositionNormalTextured.Format
                device.VertexFormat 
=  CustomVertex.PositionNormalTextured.Format;
         
//device.DeviceCaps.PixelShaderVersion.Major 获取DirectX中Device的PixelShaderVersion实现主版本号,可以学习一下(device.DeviceCaps
                
if  ( ! RenderGrayscale  ||  (device.DeviceCaps.PixelShaderVersion.Major  <   1 ))
                {
                    
if  (World.Settings.EnableSunShading)
                    {
         //获取太阳在三维空间的照射位置点
                        Point3d sunPosition 
=  SunCalculator.GetGeocentricPosition(TimeKeeper.CurrentTimeUtc);
               //转换为笛卡尔坐标系下的矢量点
                        Vector3 sunVector 
=   new  Vector3(
                            (
float )sunPosition.X,
                            (
float )sunPosition.Y,
                            (
float )sunPosition.Z);
//以下都是设置Device的渲染参数的,DirectX编程中很常见的
              
//使用灯光
                        device.RenderState.Lighting 
=   true ;
                        Material material 
=   new  Material();
                        material.Diffuse 
=  System.Drawing.Color.White;
                        material.Ambient 
=  System.Drawing.Color.White;
              
//设置材料Material
                        device.Material 
=  material;
                        device.RenderState.AmbientColor 
=  World.Settings.ShadingAmbientColor.ToArgb();                //使用法向量
                        device.RenderState.NormalizeNormals 
=   true ;
                        device.RenderState.AlphaBlendEnable 
=   true ;
              
//设置灯光参数
                        device.Lights[
0 ].Enabled  =   true ;
                        device.Lights[
0 ].Type  =  LightType.Directional;
                        device.Lights[
0 ].Diffuse  =  System.Drawing.Color.White;
               //灯光方向,来自太阳
                        device.Lights[
0 ].Direction  =  sunVector;
              //设置纹理参数
                        device.TextureState[
0 ].ColorOperation  =  TextureOperation.Modulate;
                        device.TextureState[
0 ].ColorArgument1  =  TextureArgument.Diffuse;
                        device.TextureState[
0 ].ColorArgument2  =  TextureArgument.TextureColor;
                    }
                    
else
                    {
                        device.RenderState.Lighting 
=   false ;
                        device.RenderState.Ambient 
=  World.Settings.StandardAmbientColor;

                        drawArgs.device.TextureState[
0 ].ColorOperation  =  TextureOperation.SelectArg1;
                        drawArgs.device.TextureState[
0 ].ColorArgument1  =  TextureArgument.TextureColor;
                    }

                    device.RenderState.TextureFactor 
=  System.Drawing.Color.FromArgb(m_opacity,  0 0 0 ).ToArgb();
                    device.TextureState[
0 ].AlphaOperation  =  TextureOperation.BlendFactorAlpha;
                    device.TextureState[
0 ].AlphaArgument1  =  TextureArgument.TextureColor;
                    device.TextureState[
0 ].AlphaArgument2  =  TextureArgument.TFactor;

                    drawArgs.device.VertexFormat 
=  CustomVertex.PositionNormalTextured.Format;
          
  //Device实现渲染绘制,调用的是DrawIndexedUserPrimitives。(如果看不懂,请参考DirectX编程相关知识)
                   
 drawArgs.device.DrawIndexedUserPrimitives(PrimitiveType.TriangleList,  0 ,
                        vertices.Length, indices.Length  /   3 , indices,  true , vertices);

                    
                }
                
else
                {
          //以下是使用Effect绘制的
                    
if  (grayscaleEffect  ==   null )
                    {
                        device.DeviceReset 
+=   new  EventHandler(device_DeviceReset);
                        device_DeviceReset(device, 
null );
                    }
           
//设置Effect对象参数
                    grayscaleEffect.Technique 
=   " RenderGrayscaleBrightness " ;
                    grayscaleEffect.SetValue(
" WorldViewProj " , Matrix.Multiply(device.Transform.World, Matrix.Multiply(device.Transform.View, device.Transform.Projection)));
                    grayscaleEffect.SetValue(
" Tex0 " , texture);
                    grayscaleEffect.SetValue(
" Brightness " , GrayscaleBrightness);
                    
float  opacity  =  ( float )m_opacity  /   255.0f ;
                    grayscaleEffect.SetValue(
" Opacity " , opacity);
                 //以下是Effect渲染的关键
                    
int numPasses = grayscaleEffect.Begin(0 );
                    
for  ( int  i  =   0 ; i  <  numPasses; i ++ )
                    {
                       
  grayscaleEffect.BeginPass(i);

                       
  drawArgs.device.DrawIndexedUserPrimitives(PrimitiveType.TriangleList,  0 ,
                        vertices.Length, indices.Length  /   3 , indices,  true , vertices );

                        
grayscaleEffect.EndPass();
                    }

                    
grayscaleEffect.End();
                }

                drawArgs.device.Transform.World 
=  drawArgs.WorldCamera.WorldMatrix;
                
            }
            
finally
            {
                
if  (m_opacity  <   255 )
                {
                    
//  Restore alpha blend state
                    device.RenderState.SourceBlend  =  Blend.SourceAlpha;
                    device.RenderState.DestinationBlend 
=  Blend.InvSourceAlpha;
                }

                
if ( this ._disableZbuffer)
                    drawArgs.device.RenderState.ZBufferEnable 
=   true ;
            }
        }

 

我想通过分析上面的,让大家看看Render()方法渲染的流程,里面的具体渲染知识点大家可以参看一些DirectX编程。因为Update()函数中调用的方法在上面都有介绍,所以就不再另外分析啦。希望对大家深入了解学习WW插件渲染有帮助。

 

 本系列其他部分:

WorldWind学习系列九:Blue Marble插件学习

WorldWind学习系列八:Load/Unload Plugins——直捣黄龙篇

WorldWind学习系列七:Load/Unload Plugins——投石问路篇

WorldWind学习系列六:渲染过程解析篇

WorldWind学习系列五:插件加载过程全解析

WorldWind学习系列四:功能分析——Show Planet Axis、Show Position 、Show Cross Hairs功能

WorldWind学习系列三:简单功能分析——主窗体的键盘监听处理及拷贝和粘贴位置坐标功能

WorldWind学习系列三:功能分析——截屏功能和“关于”窗体分析

WorldWind学习系列二:擒贼先擒王篇2

WorldWind学习系列二:擒贼先擒王篇1

WorldWind学习系列一:顺利起航篇

 

 

 

 

 

你可能感兴趣的:(object)