基于Aforge的手势识别之二 单点手势识别

               

本文来自http://blog.csdn.net/hellogv/ ，引用必须注明出处！

       本文把Aforge的运动识别与前面介绍的手写识别融合在一起，实现单个手指的手势识别。下图演示了本文代码运行的结果，图片有点大，请稍候。。。

       我预先让程序学习了B和C这两个字母，然后通过手指的手势识别向程序绘画图形，所以点击recorgize时，就自动把图形的特征对应的字母给识别出来了。

      这个例子关键部分在于如何灵活运用Aforge的运动识别，如何判断是要画图，还是普通的移来移去呢？在这里，我判断移动对象的大小，当突然面积增大(即两个指套合并)则开始绘图(手势识别的开始)，当拆开再合并则为解除绘图(手势识别的结束)，说白了就是用一个当前状态=!当前状态去做。

      本文的代码可以到这里下载：http://download.csdn.net/source/2313846

      下面贴出运动判断的核心代码：

        private void videoSourcePlayer1_NewFrame( object sender, ref Bitmap image )        {            nowImg = (Bitmap)image.Clone();            Bitmap objectImage = colorFilter.Apply( image );            // lock image for further processing            BitmapData objectData = objectImage.LockBits( new Rectangle( 0, 0, image.Width, image.Height ),                ImageLockMode.ReadOnly, image.PixelFormat );            // grayscaling            UnmanagedImage grayImage = grayFilter.Apply( new UnmanagedImage( objectData ) );                 // unlock image            objectImage.UnlockBits( objectData );            // locate blobs             blobCounter1.ProcessImage( grayImage );            List<Rectangle> rects = new List<Rectangle>();            rects.AddRange(blobCounter1.GetObjectsRectangles());            if ( rects.Count >0 )            {                #region 去掉内部和黏在一起的对象                for (int i = 0; i < rects.Count - 1; i++)                {                    //true表示X轴上不能相交，false表示相交                    Boolean isNoTouchX = Math.Max(rects[i + 1].Right , rects[i].Right) - Math.Min(rects[i + 1].Left ,rects[i].Left) > (rects[i].Width + rects[i + 1].Width);                    //true表示Y轴上不能相交，false表示相交                    Boolean isNoTouchY = Math.Max(rects[i + 1].Bottom, rects[i].Bottom) - Math.Min(rects[i + 1].Top, rects[i].Top) > (rects[i].Height + rects[i + 1].Height);                    if (isNoTouchX == false && isNoTouchY == false)//如果两个对象相交                    {                        Rectangle rect = new Rectangle(Math.Min(rects[i].Left, rects[i + 1].Left),                            Math.Min(rects[i].Top, rects[i + 1].Top),                            Math.Max(rects[i].Right, rects[i + 1].Right) - Math.Min(rects[i].Left, rects[i + 1].Left),                            Math.Max(rects[i].Bottom, rects[i + 1].Bottom) - Math.Min(rects[i].Top, rects[i + 1].Top));                        rects.RemoveAt(i + 1);                        rects.RemoveAt(i);                        rects.Add(rect);                        i = 0;                    }                }                #endregion                #region 画出表示点                Rectangle objectRect = rects[0];                int oldSize=oldRect.Width+oldRect.Height;                int nowSize=rects[0].Width+rects[0].Height;                if (nowSize > (oldSize * 1.2))//如果突然变大，即两个指套合并                {                    isCapture =!isCapture;                    clsHandWrite.Clear();                }                Graphics g = Graphics.FromImage(image);                if (isCapture)//如果捕捉到对象                {                    Pen pen = new Pen(Color.FromArgb(255, 0, 0), 3);                    g.DrawRectangle(pen, objectRect);                    int x = (objectRect.Left + objectRect.Width / 2) * pbDraw.Width / videoSourcePlayer1.Width;                    int y = (objectRect.Top + objectRect.Height / 2) * pbDraw.Height / videoSourcePlayer1.Height;                            clsHandWrite.Draw(x,y );                }                else//如果没有捕捉到对象                {                    Pen pen = new Pen(Color.FromArgb(160, 255, 160), 3);                    g.DrawRectangle(pen, objectRect);                }                g.Dispose();                            #endregion                oldRect = rects[0];                            }            UpdateObjectPicture(objectImage );                    }

           

再分享一下我老师大神的人工智能教程吧。零基础！通俗易懂！风趣幽默！还带黄段子！希望你也加入到我们人工智能的队伍中来！https://blog.csdn.net/jiangjunshow