Shape Matching

As you can see the camera has almost produced a binary or black and white image. This is desired as the purpose of this tutorial is to identify a part based on its shape. Having a black and white image will allow RoboRealm to know what pixels belong to the "part" and what pixels belong to the background (in this case our black conveyor belt).

A histogram shows the pixel intensity distribution of the above image. The graph shows the relative count of pixels of a certain intensity. In this case the intensity range is from 0 - 255 with a maximum pixel count of 320*240 or 76800. If you had a perfectly white image there would be a single line shown at the 255 (furthest right) place with a value of 76800. In most real images this will never happen and instead you will see a distribution of pixels throughout the histogram.

If you use the Histogram module to view the pixel intensity you will notice that the intensity distribution is almost binary but will still require some processing prior to shape matching. Matching shapes requires a perfect two color black and white image. The DBX 21BF04-Z makes this easy to achieve.

So lets get started!