Page 211 - The Definitive Guide to Building Java Robots

Preston_5564C06.fm Page 192 Friday, September 23, 2005 5:13 AM

192 CHAPTER 6 ■ VISION

Figure 6-11. Threshold image: 0 min, 127 max

Next, I want to detect the motion in a webcam. The easiest way to do this is to subtract
frames from another. The resulting difference would be a frame of motion. Depending on the
number of frames per second, this will determine the size of the image.

Code Objective

The objective here is to detect motion.

Code Discussion

In this case, all we did was store the previous image globally, and then when the next image
comes in, we subtract the grey pixel difference. For this to work, I created a class-level field
called lastImage. This static image will be set at the end of the method for the next iteration.
Also, this will take some time to process, so I’ve reduced the frames per second to 5. See
Example 6-15 and the results in Figure 6-12.

Example 6-15. backgroundSubtract()
// get motion by subtracting background between current old image
public BufferedImage backgroundSubtract(BufferedImage srcImg) {
// make sure to set the frames per second to about 5!
int h = srcImg.getHeight();
int w = srcImg.getWidth();

// create dst image
BufferedImage dstImg = new BufferedImage(w, h, BufferedImage.TYPE_INT_RGB);

if (lastImage != null) {
for (int y = 0; y < h; ++y) {
for (int x = 0; x < w; ++x) {
// get grey of image
int srcPixel = getGrey(srcImg.getRGB(x, y));
// get color of last image
int diffPixel = getGrey(lastImage.getRGB(x, y));

97022d2480fe4a63cfdfa123a6e70098