35-运动物体检测1（EmguCV学习）

Record

1、运动物体检测常用的方法：背景差法、帧差法、光流法；
2、背景差法使用原图像减去背景图像，得到前景图像，也就是运动目标；
3、帧差法根据相邻两帧或三帧图像，利用像素间的差异性，判断是否有运动目标；
4、背景差法与帧差法都进行图像减法，最常使用AbsDidd()函数；
5、背景差法基本步骤：

6、帧差法基本步骤：

7、Else:

Code

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;

using System.Drawing;
using Emgu.CV;
using Emgu.CV.Util;
using Emgu.CV.CvEnum;
using Emgu.CV.Structure;
using Emgu.Util;

namespace lesson35
{
    class Program
    {
        static void Main(string[] args)
        {
            图片背景差法检测运动物体
            //Mat bgImg = CvInvoke.Imread("1.bmp");
            //Mat fgImg = CvInvoke.Imread("55.bmp");
            //CvInvoke.Imshow("bg", bgImg);
            //CvInvoke.Imshow("fg", fgImg);
            转换灰度图
            //Mat gray1 = new Mat();
            //Mat gray2 = new Mat();
            //CvInvoke.CvtColor(bgImg, gray1, ColorConversion.Bgr2Gray);
            //CvInvoke.CvtColor(fgImg, gray2, ColorConversion.Bgr2Gray);
            图片做差
            //Mat diff = new Mat();
            //CvInvoke.AbsDiff(gray2, gray1, diff);
            //CvInvoke.Imshow("diff", diff);
            二值化
            //CvInvoke.Threshold(diff, diff, 50, 255, ThresholdType.Binary);
            //CvInvoke.Imshow("threshold", diff);
            中值滤波或腐蚀去除噪点(中值滤波效果更好)
            CvInvoke.MedianBlur(diff, diff, 3);
            CvInvoke.Imshow("medianblur", diff);
            //Mat element = CvInvoke.GetStructuringElement(ElementShape.Rectangle, new Size(3, 3), new Point(-1, -1));
            //CvInvoke.Erode(diff, diff, element, new Point(-1, -1), 1, BorderType.Default, new MCvScalar());
            //CvInvoke.Imshow("erode", diff);
            膨胀连通区域
            //Mat element2 = CvInvoke.GetStructuringElement(ElementShape.Rectangle, new Size(11, 11), new Point(-1, -1));
            //CvInvoke.Dilate(diff, diff, element2, new Point(-1, -1), 1, BorderType.Default, new MCvScalar());
            //CvInvoke.Imshow("dilate", diff);
            绘制最小外接矩形
            //VectorOfVectorOfPoint contours = new VectorOfVectorOfPoint();
            //CvInvoke.FindContours(diff, contours, null, RetrType.External, ChainApproxMethod.ChainApproxNone);
            //for(int i = 0; i < contours.Size; i++)
            //{
            //    Rectangle rect = CvInvoke.BoundingRectangle(contours[i]);
            //    CvInvoke.Rectangle(fgImg, rect, new MCvScalar(0, 0, 255), 2);
            //}
            //CvInvoke.Imshow("result", fgImg);
            //CvInvoke.WaitKey(0);

            ///视频背景差法检测运动物体
            //VideoCapture cap = new VideoCapture("ball.avi");
            //if (!cap.IsOpened)
            //{
            //    Console.WriteLine("open the video failed..");
            //    return;
            //}
            //int count = 0;
            //Mat bgImg = new Mat();
            //Mat frame = new Mat();
            //while (true)
            //{
            //    cap.Read(frame);
            //    if (frame.IsEmpty)
            //    {
            //        Console.WriteLine("frame is empty...");
            //        break;
            //    }
            //    count++;
            //    if (1 == count)
            //    {
            //        bgImg = frame.Clone();  //取第一帧为背景
            //    }
            //    CvInvoke.Imshow("video", frame);
            //    Mat result = MoveDetect(bgImg, frame);
            //    CvInvoke.Imshow("move", result);
            //    if (CvInvoke.WaitKey(50) == 27)
            //    {
            //        break;
            //    }
            //}

            ///图片帧差法检测运动物体
            //Mat bgImg = CvInvoke.Imread("54.bmp");
            //Mat fgImg = CvInvoke.Imread("55.bmp");
            //CvInvoke.Imshow("bg", bgImg);
            //CvInvoke.Imshow("fg", fgImg);
            转换为灰度图
            //Mat gray = new Mat();
            //Mat gray2 = new Mat();
            //CvInvoke.CvtColor(bgImg, gray, ColorConversion.Bgr2Gray);
            //CvInvoke.CvtColor(fgImg, gray2, ColorConversion.Bgr2Gray);
            做差
            //Mat diff = new Mat();
            //CvInvoke.AbsDiff(gray2, gray, diff);
            //CvInvoke.Imshow("diff", diff);
            二值化
            //CvInvoke.Threshold(diff, diff, 45, 255, ThresholdType.Binary);
            //CvInvoke.Imshow("threshold", diff);
            //Mat element = CvInvoke.GetStructuringElement(ElementShape.Rectangle, new Size(1, 1), new Point(-1, -1));
            //CvInvoke.Erode(diff, diff, element, new Point(-1, -1), 1, BorderType.Default, new MCvScalar());
            //CvInvoke.Imshow("erode", diff);
            膨胀连通区域
            //Mat element2 = CvInvoke.GetStructuringElement(ElementShape.Rectangle, new Size(11, 11), new Point(-1, -1));
            //CvInvoke.Dilate(diff, diff, element2, new Point(-1, -1), 1, BorderType.Default, new MCvScalar());
            //CvInvoke.Imshow("dilate", diff);
            绘制最小外接矩形
            //VectorOfVectorOfPoint contours = new VectorOfVectorOfPoint();
            //CvInvoke.FindContours(diff, contours, null, RetrType.External, ChainApproxMethod.ChainApproxNone);
            //for (int i = 0; i < contours.Size; i++)
            //{
            //    Rectangle rect = CvInvoke.BoundingRectangle(contours[i]);
            //    CvInvoke.Rectangle(fgImg, rect, new MCvScalar(0, 0, 255), 2);
            //}
            //CvInvoke.Imshow("result", fgImg);
            //CvInvoke.WaitKey(0);

            ///视频帧差法检测运动物体
            VideoCapture cap = new VideoCapture("man.avi");
            if(!cap.IsOpened)
            {
                Console.WriteLine("Open video failed..");
                return;
            }
            int count = 0;
            Mat bgImg = new Mat();
            Mat frame = new Mat();
            while(true)
            {
                cap.Read(frame);
                if(frame.IsEmpty)
                {
                    Console.WriteLine("frame is empty..");
                    break;
                }
                count++;
                if (count == 1)
                    bgImg = frame.Clone();
                Mat result = MoveDetect2(bgImg, frame);
                CvInvoke.Imshow("move", result);
                bgImg = frame.Clone();	//更新前一帧
                if(CvInvoke.WaitKey(30)==27)
                {
                    break;
                }

            }


        }
        static Mat MoveDetect2(Mat bgImg,Mat fgImg)
        {
            Mat result = fgImg.Clone();
            Mat gray = new Mat();
            Mat gray2 = new Mat();
            CvInvoke.CvtColor(bgImg, gray, ColorConversion.Bgr2Gray);
            CvInvoke.CvtColor(fgImg, gray2, ColorConversion.Bgr2Gray);
            //做差
            Mat diff = new Mat();
            CvInvoke.AbsDiff(gray2, gray, diff);
            CvInvoke.Imshow("diff", diff);
            //二值化
            CvInvoke.Threshold(diff, diff, 45, 255, ThresholdType.Binary);
            CvInvoke.Imshow("threshold", diff);
            //中值滤波
            CvInvoke.MedianBlur(diff, diff, 5);
            CvInvoke.Imshow("median blur", diff);
            //膨胀
            Mat element = CvInvoke.GetStructuringElement(ElementShape.Rectangle, new Size(9, 9), new Point(-1, -1));
            CvInvoke.Dilate(diff, diff, element, new Point(-1, -1), 1, BorderType.Default, new MCvScalar());
            CvInvoke.Imshow("dilate", diff);
            //绘制轮廓矩形
            VectorOfVectorOfPoint contours = new VectorOfVectorOfPoint();
            CvInvoke.FindContours(diff, contours, null, RetrType.External, ChainApproxMethod.ChainApproxNone);
            for (int i = 0; i < contours.Size; i++)
            {
                Rectangle rect = CvInvoke.BoundingRectangle(contours[i]);
                CvInvoke.Rectangle(result, rect, new MCvScalar(0, 0, 255), 2);
            }
            return result;
        }


        static Mat MoveDetect(Mat bgImg, Mat fgImg)
        {
            Mat result = fgImg.Clone();
            Mat gray = new Mat();
            Mat gray2 = new Mat();
            CvInvoke.CvtColor(bgImg, gray, ColorConversion.Bgr2Gray);
            CvInvoke.CvtColor(fgImg, gray2, ColorConversion.Bgr2Gray);
            //做差
            Mat diff = new Mat();
            CvInvoke.AbsDiff(gray2, gray, diff);
            CvInvoke.Imshow("diff", diff);
            //二值化
            CvInvoke.Threshold(diff, diff, 10, 255, ThresholdType.Binary);
            CvInvoke.Imshow("threshold", diff);
            //中值滤波
            CvInvoke.MedianBlur(diff, diff, 5);
            CvInvoke.Imshow("median blur", diff);
            //膨胀
            Mat element = CvInvoke.GetStructuringElement(ElementShape.Rectangle, new Size(9, 9), new Point(-1, -1));
            CvInvoke.Dilate(diff, diff, element, new Point(-1, -1), 1, BorderType.Default, new MCvScalar());
            CvInvoke.Imshow("dilate", diff);
            //绘制轮廓矩形
            VectorOfVectorOfPoint contours = new VectorOfVectorOfPoint();
            CvInvoke.FindContours(diff, contours, null, RetrType.External, ChainApproxMethod.ChainApproxNone);
            for(int i = 0; i < contours.Size; i++)
            {
                Rectangle rect = CvInvoke.BoundingRectangle(contours[i]);
                CvInvoke.Rectangle(result, rect, new MCvScalar(0, 0, 255), 2);
            }

            return result;

        }

    }
}

效果

1. 图片背景差法：

2. 视频背景差法检测运动物体：

3. 图片帧差法检测运动物体：

4. 视频帧差法检测运动物体