Use OpenCVdotNet to Capture Object on Webcam

• Download source - 58.25 KB

Introduction 

This code is based on "Webcam Simple Frame Difference" at the blog: http://haryoktav.wordpress.com/2009/02/28/webcam-c-simple-frame-difference/.

Objectives 

• Monitoring a room with a webcam
• No need to save a long movie, just capture pictures showing the change
• Easy to view the result, no need to view the entire long movie (maybe it does not have any change) 

Background

This code uses OpenCVdotNet library to:

1. download and install OpenCV 1.0 - http://sourceforge.net/projects/opencv/ (do not install 1.1)
2. download and install OpenCVDotNet 0.7 - http://code.google.com/p/opencvdotnet/

I use Visual Studio 2005 to design a form like this:

The form has 2 pictureboxes to show the current image and the captured image; 1 timer to repeat checking the difference of image in the room and save to file if needed; 1 button to active timer, 1 label to show the number of captured images.

Using the Code 

There is only one form, and the code is as given below: 

After installing OpenCVDotnet library, add it to the project:

using System;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Drawing.Imaging;
using System.Text;
using System.Windows.Forms;
using OpenCVDotNet; 	//don't forget to add opencvdotnel.dll as Reference

Declare some private variables:

namespace Objcap
{
    public partial class Form1 : Form
    {
        private CVCapture capture;            //buffer for captured image from webcam
        private CVImage backgnd;              //buffer for previous image, 
					//subtracts result and captured image
        private byte threshold=30;            //threshold value = 30
        private bool only_first = false;      //flag
        private int filename = 0;             //to make file name change after each save 

Build a function for button1_Click: to active timer: 

     public Form1()
        {
            InitializeComponent();
        }
        private void button1_Click(object sender, EventArgs e)
        {
            timer1.Enabled = !(timer1.Enabled); //toggle start/stop 
            if (timer1.Enabled)
                capture = new CVCapture(0);     //start camera when timer enabled
            else
                capture.Release();              //stop camera when timer disabled
        }        

The main function is in timer1_tick:

Every time Timer1 ticks (set in interval properties = 500 ->500 milliseconds), we need to compare the previous image (saved in backgnd) with current image (frame).
If there is a difference, it means there is an object in or out of your room. So save that image.

private void timer1_Tick(object sender, EventArgs e)
{
    using (CVImage frame = capture.QueryFrame())  //get camera frame
    {
        frame.Resize(160, 120);             	//resize image to 160x120 pixel
        if (only_first == false)            	//only first time
        {
            backgnd = new CVImage(frame);   	//copy buffer to get the same size
            only_first = true;              	//disable flag forever
        }
        //gray scale conversion
        for (int row = 0; row < frame.Height; ++row)
        {
            for (int col = 0; col < frame.Width; ++col)
            {
                CVRgbPixel pixel = frame[row, col]; 	//get current pixel
                byte bwValue = pixel.BwValue;   	//get the gray color
                frame[row, col] = new CVRgbPixel(bwValue, bwValue, bwValue);//set current
								// pixel to gray
            }
        }
        int pixelnum = 0; // number of different pixel
        for (int row = 0; row < frame.Height; ++row)
        {
            for (int col = 0; col < frame.Width; ++col)
            {
                if (Math.Abs(frame[row, col].BwValue - 
			backgnd[row, col].BwValue) > threshold)
                {                       
                    pixelnum++;
                }
                backgnd[row, col] = frame[row, col];  	// save current frame to 
						// background frame for next time
            }
        } 
         pictureBox1.Image = frame.ToBitmap();  //display image
         int objsize = 500;  	// depends on Size of Object to be captured
                             	// if number of different pixels > 
				// objsize then save image to hard drive
        if (pixelnum > objsize)
        {
            using (CVImage frame2 = capture.QueryFrame())
            {                              
                // Capture color image
                filename++;                      	//change file name   
                string path = "D:/webcam/";        	// Path to save file
                path += DateTime.Today.Day.ToString()+"."+
                	filename.ToString() + ".jpg";   	// Full path - file name with 
						// added day saved         
                frame2.ToBitmap().Save(path, ImageFormat.Jpeg);   // Save image file 
						// with max resolution of webcam
                frame2.Resize(160, 120);               // Resize image to show on 
						// picturebox3
                pictureBox3.Image = frame2.ToBitmap();	//Show captured image
            }
        }      
         label5.Text = "Pictures: " + filename.ToString();   	// Show number of 
							// captured images
    }
    GC.Collect();
}

Points of Interest

This code runs well. You can run the program for a month if you want.
The program will save only important images with file named by saving day.

The program when run: (remember to create a folder: D:\Webcam):

And the result (saved files):

History

I tried to improve it. Now it does not cost memory any more.