Ah, the joys of jetlag. After getting up with jetlag at 3am, yesterday, I finally worked out what was wrong with my inconsistently performing speech recognition approach shown in this previous post.
I added a few more enhancements for Wednesday’s AU class: access to the Kinect’s elevation property (allowing you to tilt backwards and forwards using the voice commands “up” and “down”), as well as the ability to complete (“finish”) the operation as we as cancel (“cancel”) it.
All in all the application has become pretty useful. As I made these enhancements to the base KinectJig class, they have become available throughout the samples building upon it. Which is helpful, but not ideal for the purposes of my demo: I went ahead and implemented an additional system variable (KINSPEECH) to allow the behaviour to remain dormant until the appropriate part of the session (you can imagine the confusion during a presentation if what you were saying to the audience was taken as voice commands for the current demo sample :-).
using Autodesk.AutoCAD.ApplicationServices;
using Autodesk.AutoCAD.DatabaseServices;
using Autodesk.AutoCAD.EditorInput;
using Autodesk.AutoCAD.Geometry;
using Autodesk.AutoCAD.GraphicsInterface;
using Autodesk.AutoCAD.Runtime;
using Microsoft.Research.Kinect.Nui;
using Microsoft.Research.Kinect.Audio;
using Microsoft.Speech.AudioFormat;
using Microsoft.Speech.Recognition;
using System.Runtime.InteropServices;
using System.Collections.Generic;
using System.Diagnostics;
using System.Reflection;
using System.Threading;
using System.Windows.Threading;
using System.Linq;
using System.IO;
using System;
namespace KinectSamples
{
public class ColoredPoint3d
{
public double X, Y, Z;
public int R, G, B;
}
public abstract class KinectJig : DrawJig
{
// To stop the running jig by sending a cancel request
[DllImport("acad.exe", CharSet = CharSet.Auto,
CallingConvention = CallingConvention.Cdecl,
EntryPoint = "?acedPostCommand@@YAHPB_W@Z"
)]
extern static private int acedPostCommand(string strExpr);
// We need our Kinect sensor
private Runtime _kinect = null;
// And audio array
private KinectAudioSource _audio = null;
// Microsoft Speech recognition engine
private SpeechRecognitionEngine _sre;
// With the images collected by it
private ImageFrame _depth = null;
private ImageFrame _video = null;
// An offset value we use to move the mouse back
// and forth by one screen unit
private int _offset;
// Flag and property for when we want to exit
private bool _finished;
public bool Finished
{
get { return _finished; }
set { _finished = value; }
}
// Selected color, for whatever usage makes sense
// in child classes
private short _color;
internal short ColorIndex
{
get { return _color; }
}
// Elevation setting to apply to Kinect camera
private int _tilt;
// Sampling value to reduce point set on capture
internal static short Sampling
{
get
{
try
{
return (short)Application.GetSystemVariable("KINSAMP");
}
catch
{
return 50;
}
}
}
// Flag to indicate whether to recognize speech input
internal static bool Speech
{
get
{
try
{
return
(short)Application.GetSystemVariable("KINSPEECH") > 0;
}
catch
{
return false;
}
}
}
// Extents to filter points
private static Extents3d? _ext = null;
public static Extents3d? Extents
{
get { return _ext; }
set { _ext = value; }
}
public KinectJig()
{
// Initialise the various members
_offset = 1;
_color = 3;
_tilt = 0;
_finished = false;
_kinect = Runtime.Kinects[0];
_kinect.VideoFrameReady +=
new EventHandler<ImageFrameReadyEventArgs>(
OnVideoFrameReady
);
_kinect.DepthFrameReady +=
new EventHandler<ImageFrameReadyEventArgs>(
OnDepthFrameReady
);
_kinect.SkeletonFrameReady +=
new EventHandler<SkeletonFrameReadyEventArgs>(
OnSkeletonFrameReady
);
if (Speech)
{
InitializeSpeech();
}
}
public virtual void OnDepthFrameReady(
object sender, ImageFrameReadyEventArgs e
)
{
_depth = e.ImageFrame;
}
public virtual void OnVideoFrameReady(
object sender, ImageFrameReadyEventArgs e
)
{
_video = e.ImageFrame;
}
public virtual void OnSkeletonFrameReady(
object sender, SkeletonFrameReadyEventArgs e
)
{
}
void OnSpeechHypothesized(
object sender, SpeechHypothesizedEventArgs e
)
{
}
void OnSpeechRecognized(
object sender, SpeechRecognizedEventArgs e
)
{
// Ignore if we don't have a high degree of confidence
if (e.Result.Confidence < 0.8)
return;
// Set the color property based om the text input
switch (e.Result.Text.ToUpperInvariant())
{
case "RED":
_color = 1;
break;
case "YELLOW":
_color = 2;
break;
case "GREEN":
_color = 3;
break;
case "CYAN":
_color = 4;
break;
case "BLUE":
_color = 5;
break;
case "PINK":
case "MAGENTA":
_color = 6;
break;
case "UP":
_tilt = 5;
break;
case "DOWN":
_tilt = -5;
break;
case "CANCEL":
CancelJig();
break;
case "FINISH":
_finished = true;
break;
default:
break;
}
}
public void StartSensor()
{
if (_kinect != null)
{
// We still need to enable skeletal tracking
// in order to map to "real" space, even
// if we're not actually getting skeleton data
_kinect.Initialize(
RuntimeOptions.UseDepth |
RuntimeOptions.UseColor |
RuntimeOptions.UseSkeletalTracking
);
_kinect.VideoStream.Open(
ImageStreamType.Video, 2,
ImageResolution.Resolution640x480,
ImageType.Color
);
_kinect.DepthStream.Open(
ImageStreamType.Depth, 2,
ImageResolution.Resolution640x480,
ImageType.Depth
);
_kinect.NuiCamera.ElevationAngle = 0;
// We need speech recognition started on a separate,
// MTA thread
if (Speech)
{
var t = new Thread(StartSpeech);
t.Start();
}
}
}
private static RecognizerInfo GetKinectRecognizer()
{
Func<RecognizerInfo, bool> matchingFunc =
r =>
{
string value;
r.AdditionalInfo.TryGetValue("Kinect", out value);
return
"True".Equals(
value,
StringComparison.InvariantCultureIgnoreCase
) &&
"en-US".Equals(
r.Culture.Name,
StringComparison.InvariantCultureIgnoreCase
);
};
return
SpeechRecognitionEngine.InstalledRecognizers().Where(
matchingFunc
).FirstOrDefault();
}
private void InitializeSpeech()
{
RecognizerInfo ri = GetKinectRecognizer();
if (ri == null)
{
Editor ed =
Application.DocumentManager.MdiActiveDocument.Editor;
ed.WriteMessage(
"There was a problem initializing Speech Recognition. " +
"Ensure you have the Microsoft Speech SDK installed " +
"and configured."
);
}
try
{
_sre = new SpeechRecognitionEngine(ri.Id);
}
catch
{
Editor ed =
Application.DocumentManager.MdiActiveDocument.Editor;
ed.WriteMessage(
"There was a problem initializing Speech Recognition. " +
"Ensure you have the Microsoft Speech SDK installed " +
"and configured."
);
}
// Populate our word choices
Choices words = new Choices();
words.Add("red");
words.Add("green");
words.Add("blue");
words.Add("yellow");
words.Add("pink");
words.Add("magenta");
words.Add("cyan");
words.Add("up");
words.Add("down");
words.Add("finish");
words.Add("cancel");
// Create a GrammarBuilder from them
GrammarBuilder gb = new GrammarBuilder();
gb.Culture = ri.Culture;
gb.Append(words);
// Create the actual Grammar instance, and then load it
// into the speech recognizer
Grammar g = new Grammar(gb);
_sre.LoadGrammar(g);
// Attach our event handler for recognized commands
// We won't worry about rejected or hypothesized callbacks
_sre.SpeechHypothesized += OnSpeechHypothesized;
_sre.SpeechRecognized += OnSpeechRecognized;
}
[MTAThread]
private void StartSpeech()
{
try
{
// Create and setup our audio source
_audio = new KinectAudioSource();
_audio.SystemMode = SystemMode.OptibeamArrayOnly;
_audio.FeatureMode = true;
_audio.AutomaticGainControl = false;
_audio.MicArrayMode = MicArrayMode.MicArrayAdaptiveBeam;
// Get the audio stream and pass it to the
// speech recognition engine
Stream kinectStream = _audio.Start();
_sre.SetInputToAudioStream(
kinectStream,
new SpeechAudioFormatInfo(
EncodingFormat.Pcm, 16000, 16, 1,
32000, 2, null
)
);
_sre.RecognizeAsync(RecognizeMode.Multiple);
}
catch
{
Editor ed =
Application.DocumentManager.MdiActiveDocument.Editor;
ed.WriteMessage(
"There was a problem initializing the KinectAudioSource." +
" Ensure you have the Kinect SDK installed correctly."
);
}
}
public void StopSensor()
{
if (_kinect != null)
{
if (_audio != null)
{
_audio.Stop();
_sre.RecognizeAsyncCancel();
_sre.RecognizeAsyncStop();
_audio.Dispose();
_audio = null;
}
_kinect.Uninitialize();
_kinect = null;
}
}
protected virtual SamplerStatus SamplerData()
{
return SamplerStatus.Cancel;
}
protected virtual bool WorldDrawData(WorldDraw draw)
{
return false;
}
protected override SamplerStatus Sampler(JigPrompts prompts)
{
// We don't really need a point, but we do need some
// user input event to allow us to loop, processing
// for the Kinect input
PromptPointResult ppr =
prompts.AcquirePoint("\nClick to capture: ");
if (ppr.Status == PromptStatus.OK)
{
if (_finished)
{
CancelJig();
return SamplerStatus.Cancel;
}
if (_tilt != 0)
{
_kinect.NuiCamera.ElevationAngle += _tilt;
_tilt = 0;
}
return SamplerData();
}
return SamplerStatus.Cancel;
}
protected override bool WorldDraw(WorldDraw draw)
{
return WorldDrawData(draw);
}
public void ForceMessage()
{
// Let's move the mouse slightly to avoid having
// to do it manually to keep the input coming
System.Drawing.Point pt =
System.Windows.Forms.Cursor.Position;
System.Windows.Forms.Cursor.Position =
new System.Drawing.Point(
pt.X, pt.Y + _offset
);
_offset = -_offset;
}
public List<ColoredPoint3d> GeneratePointCloud(
int sampling, bool useColor = false)
{
return GeneratePointCloud(
_kinect, _depth, _video, sampling, useColor
);
}
// Generate a point cloud from depth and RGB data
internal static List<ColoredPoint3d> GeneratePointCloud(
Runtime kinect, ImageFrame depth, ImageFrame video,
int sampling, bool withColor = false
)
{
// We will return a list of our ColoredPoint3d objects
List<ColoredPoint3d> res = new List<ColoredPoint3d>();
// Let's start by determining the dimensions of the
// respective images
int depHeight = depth.Image.Height;
int depWidth = depth.Image.Width;
int vidHeight = video.Image.Height;
int vidWidth = video.Image.Width;
// For the sake of this initial implementation, we
// expect them to be the same size. But this should not
// actually need to be a requirement
if (vidHeight != depHeight || vidWidth != depWidth)
{
Application.DocumentManager.MdiActiveDocument.
Editor.WriteMessage(
"\nVideo and depth images are of different sizes."
);
return null;
}
// Depth and color data for each pixel
Byte[] depthData = depth.Image.Bits;
Byte[] colorData = video.Image.Bits;
// Loop through the depth information - we process two
// bytes at a time
for (int i = 0; i < depthData.Length; i += (2 * sampling))
{
// The depth pixel is two bytes long - we shift the
// upper byte by 8 bits (a byte) and "or" it with the
// lower byte
int depthPixel = (depthData[i + 1] << 8) | depthData[i];
// The x and y positions can be calculated using modulus
// division from the array index
int x = (i / 2) % depWidth;
int y = (i / 2) / depWidth;
// The x and y we pass into DepthImageToSkeleton() need to
// be normalised (between 0 and 1), so we divide by the
// width and height of the depth image, respectively
// As we're using UseDepth (not UseDepthAndPlayerIndex) in
// the depth sensor settings, we also need to shift the
// depth pixel by 3 bits
Vector v =
kinect.SkeletonEngine.DepthImageToSkeleton(
((float)x) / ((float)depWidth),
((float)y) / ((float)depHeight),
(short)(depthPixel << 3)
);
// A zero value for Z means there is no usable depth for
// that pixel
if (v.Z > 0)
{
// Create a ColorVector3 to store our XYZ and RGB info
// for a pixel
ColoredPoint3d cv = new ColoredPoint3d();
cv.X = v.X;
cv.Y = v.Z;
cv.Z = v.Y;
// Only calculate the colour when it's needed (as it's
// now more expensive, albeit more accurate)
if (withColor)
{
// Get the colour indices for that particular depth
// pixel. We once again need to shift the depth pixel
// and also need to flip the x value (as UseDepth means
// it is mirrored on X) and do so on the basis of
// 320x240 resolution (so we divide by 2, assuming
// 640x480 is chosen earlier), as that's what this
// function expects. Phew!
int colorX, colorY;
kinect.NuiCamera.
GetColorPixelCoordinatesFromDepthPixel(
video.Resolution, video.ViewArea,
320 - (x / 2), (y / 2), (short)(depthPixel << 3),
out colorX, out colorY
);
// Make sure both indices are within bounds
colorX = Math.Max(0, Math.Min(vidWidth - 1, colorX));
colorY = Math.Max(0, Math.Min(vidHeight - 1, colorY));
// Extract the RGB data from the appropriate place
// in the colour data
int colIndex = 4 * (colorX + (colorY * vidWidth));
cv.B = (byte)(colorData[colIndex + 0]);
cv.G = (byte)(colorData[colIndex + 1]);
cv.R = (byte)(colorData[colIndex + 2]);
}
else
{
// If we don't need colour information, just set each
// pixel to white
cv.B = 255;
cv.G = 255;
cv.R = 255;
}
// Add our pixel data to the list to return
res.Add(cv);
}
}
// Apply a bounding box filter, if one is defined
if (_ext.HasValue)
{
// Use LINQ to get the points within the
// bounding box
var vecSet =
from ColoredPoint3d vec in res
where
vec.X > _ext.Value.MinPoint.X &&
vec.X < _ext.Value.MaxPoint.X &&
vec.Y > _ext.Value.MinPoint.Y &&
vec.Y < _ext.Value.MaxPoint.Y &&
vec.Z > _ext.Value.MinPoint.Z &&
vec.Z < _ext.Value.MaxPoint.Z
select vec;
// Convert our IEnumerable<> into a List<>
res = vecSet.ToList<ColoredPoint3d>();
}
return res;
}
// Save the provided point cloud to a specific file
internal static void ExportPointCloud(
List<ColoredPoint3d> vecs, string filename
)
{
if (vecs.Count > 0)
{
using (StreamWriter sw = new StreamWriter(filename))
{
// For each pixel, write a line to the text file:
// X, Y, Z, R, G, B
foreach (ColoredPoint3d pt in vecs)
{
sw.WriteLine(
"{0}, {1}, {2}, {3}, {4}, {5}",
pt.X, pt.Y, pt.Z, pt.R, pt.G, pt.B
);
}
}
}
}
// Translate from Skeleton Space to WCS
internal static Point3d PointFromVector(Vector v)
{
// Rather than just return a point, we're effectively
// transforming it to the drawing space: flipping the
// Y and Z axes (which makes it consistent with the
// point cloud, and makes sure Z is actually up - from
// the Kinect's perspective Y is up), and reversing
// the X axis (which is the result of choosing UseDepth
// rather than UseDepthAndPlayerIndex)
return new Point3d(-v.X, v.Z, v.Y);
}
// Cancel the running jig
internal static void CancelJig()
{
acedPostCommand("CANCELCMD");
}
// Write the provided point cloud to file, then chain
// the commands needed to import it into AutoCAD
public void WriteAndImportPointCloud(
Document doc, List<ColoredPoint3d> vecs
)
{
Editor ed = doc.Editor;
// We'll store most local files in the temp folder.
// We get a temp filename, delete the file and
// use the name for our folder
string localPath = Path.GetTempFileName();
File.Delete(localPath);
Directory.CreateDirectory(localPath);
localPath += "\\";
// Paths for our temporary files
string txtPath = localPath + "points.txt";
string lasPath = localPath + "points.las";
// Our PCG file will be stored under My Documents
string outputPath =
Environment.GetFolderPath(
Environment.SpecialFolder.MyDocuments
) + "\\Kinect Point Clouds\\";
if (!Directory.Exists(outputPath))
Directory.CreateDirectory(outputPath);
// We'll use the title as a base filename for the PCG,
// but will use an incremented integer to get an unused
// filename
int cnt = 0;
string pcgPath;
do
{
pcgPath =
outputPath + "Kinect" +
(cnt == 0 ? "" : cnt.ToString()) + ".pcg";
cnt++;
}
while (File.Exists(pcgPath));
// The path to the txt2las tool will be the same as the
// executing assembly (our DLL)
string exePath =
Path.GetDirectoryName(
Assembly.GetExecutingAssembly().Location
) + "\\";
if (!File.Exists(exePath + "txt2las.exe"))
{
ed.WriteMessage(
"\nCould not find the txt2las tool: please make sure " +
"it is in the same folder as the application DLL."
);
return;
}
// Export our point cloud from the jig
ed.WriteMessage(
"\nSaving TXT file of the captured points.\n"
);
ExportPointCloud(vecs, txtPath);
// Use the txt2las utility to create a .LAS
// file from our text file
ed.WriteMessage(
"\nCreating a LAS from the TXT file.\n"
);
ProcessStartInfo psi =
new ProcessStartInfo(
exePath + "txt2las",
"-i \"" + txtPath +
"\" -o \"" + lasPath +
"\" -parse xyzRGB"
);
psi.CreateNoWindow = false;
psi.WindowStyle = ProcessWindowStyle.Hidden;
// Wait up to 20 seconds for the process to exit
try
{
using (Process p = Process.Start(psi))
{
p.WaitForExit();
}
}
catch
{ }
// If there's a problem, we return
if (!File.Exists(lasPath))
{
ed.WriteMessage(
"\nError creating LAS file."
);
return;
}
File.Delete(txtPath);
ed.WriteMessage(
"Indexing the LAS and attaching the PCG.\n"
);
// Index the .LAS file, creating a .PCG
string lasLisp = lasPath.Replace('\\', '/'),
pcgLisp = pcgPath.Replace('\\', '/');
doc.SendStringToExecute(
"(command \"_.POINTCLOUDINDEX\" \"" +
lasLisp + "\" \"" +
pcgLisp + "\")(princ) ",
false, false, false
);
// Attach the .PCG file
doc.SendStringToExecute(
"_.WAITFORFILE \"" +
pcgLisp + "\" \"" +
lasLisp + "\" " +
"(command \"_.-POINTCLOUDATTACH\" \"" +
pcgLisp +
"\" \"0,0\" \"1\" \"0\")(princ) ",
false, false, false
);
doc.SendStringToExecute(
"_.-VISUALSTYLES _C _Realistic ",
false, false, false
);
}
}
public class KinectCommands
{
// Set the clipping volume for the current point cloud
[CommandMethod("ADNPLUGINS", "KINBOUNDS", CommandFlags.Modal)]
public void SetBoundingBox()
{
Document doc =
Autodesk.AutoCAD.ApplicationServices.
Application.DocumentManager.MdiActiveDocument;
Editor ed = doc.Editor;
// Ask the user to select an entity
PromptEntityOptions peo =
new PromptEntityOptions(
"\nSelect entity to define bounding box"
);
peo.AllowNone = true;
peo.Keywords.Add("None");
peo.Keywords.Default = "None";
PromptEntityResult per = ed.GetEntity(peo);
if (per.Status != PromptStatus.OK)
return;
// If "None" selected, clear the bounding box
if (per.Status == PromptStatus.None ||
per.StringResult == "None")
{
KinectJig.Extents = null;
ed.WriteMessage("\nBounding box cleared.");
return;
}
// Otherwise open the entity and gets its extents
Transaction tr =
doc.TransactionManager.StartTransaction();
using (tr)
{
Entity ent =
tr.GetObject(per.ObjectId, OpenMode.ForRead)
as Entity;
if (ent != null)
KinectJig.Extents = ent.Bounds;
ed.WriteMessage(
"\nBounding box set to {0}", KinectJig.Extents
);
tr.Commit();
}
}
// A command which waits for a particular PCG file to exist
[CommandMethod(
"ADNPLUGINS", "WAITFORFILE", CommandFlags.NoHistory
)]
public void WaitForFileToExist()
{
Document doc =
Application.DocumentManager.MdiActiveDocument;
Editor ed = doc.Editor;
HostApplicationServices ha =
HostApplicationServices.Current;
PromptResult pr = ed.GetString("Enter path to PCG: ");
if (pr.Status != PromptStatus.OK)
return;
string pcgPath = pr.StringResult.Replace('/', '\\');
pr = ed.GetString("Enter path to LAS: ");
if (pr.Status != PromptStatus.OK)
return;
string lasPath = pr.StringResult.Replace('/', '\\');
ed.WriteMessage(
"\nWaiting for PCG creation to complete...\n"
);
// Check the write time for the PCG file...
// if it hasn't been written to for at least half a second,
// then we try to use a file lock to see whether the file
// is accessible or not
const int ticks = 50;
TimeSpan diff;
bool cancelled = false;
// First loop is to see when writing has stopped
// (better than always throwing exceptions)
while (true)
{
if (File.Exists(pcgPath))
{
DateTime dt = File.GetLastWriteTime(pcgPath);
diff = DateTime.Now - dt;
if (diff.Ticks > ticks)
break;
}
System.Windows.Forms.Application.DoEvents();
if (HostApplicationServices.Current.UserBreak())
{
cancelled = true;
break;
}
}
// Second loop will wait until file is finally accessible
// (by calling a function that requests an exclusive lock)
if (!cancelled)
{
int inacc = 0;
while (true)
{
if (IsFileAccessible(pcgPath))
break;
else
inacc++;
System.Windows.Forms.Application.DoEvents();
if (HostApplicationServices.Current.UserBreak())
{
cancelled = true;
break;
}
}
ed.WriteMessage("\nFile inaccessible {0} times.", inacc);
try
{
CleanupTmpFiles(lasPath);
}
catch
{ }
}
}
// Return whether a file is accessible
internal bool IsFileAccessible(string filename)
{
// If the file can be opened for exclusive access it means
// the file is accesible
try
{
FileStream fs =
File.Open(
filename, FileMode.Open,
FileAccess.Read, FileShare.None
);
using (fs)
{
return true;
}
}
catch (IOException)
{
return false;
}
}
// Remove any temporary files from the point cloud import
internal void CleanupTmpFiles(string txtPath)
{
if (File.Exists(txtPath))
File.Delete(txtPath);
Directory.Delete(
Path.GetDirectoryName(txtPath)
);
}
}
}
Here are some results of running the code with much finer grained control over the colour changes, to give you some idea of the possibilities:
Subscribe via RSS