Servo - July 2011

Adding Vision: Interfacing With RoboRealm

by John Blankenship and Samuel Mishal

For many hobbyists, giving their robot vision is one of their ultimate goals. RoboRealm® is an application for use in computer vision, image analysis, and robotic vision systems. It makes adding vision painless by providing easy-to-use modules that provide color filters, shape recognition, edge detection, and much, much more.

This article examines a variety of uncomplicated ways to interface your programs with RoboRealm.

www.servomagazine.com/index.php?/magazine/article/ july2011_Blankenship

Although vision is a desirable goal for many robot hobbyists, the complex low level programming and mathematical algorithms involved often force many people to assume vision is beyond their capabilities.

Let’s examine what is entailed when processing vision. An image to be analyzed can be thought of as three, two-dimensional arrays (one each for red, green, and blue information). The numbers stored in each element of these arrays represent the brightness of a respective x-y coordinate in the image. Processing and analyzing the image involves applying mathematical algorithms to these arrays. Let’s look at some examples.

An image can be made brighter by increasing all of the array elements by some percentage. A single color pixel in the image can be converted to black and white by replacing the element in each of the three arrays that is associated with that pixel, with an average of those same three elements. These operations are relatively simple but their complexity can increase quickly as more elaborate actions are needed. Converting a standard image to one showing only the edges of objects, for example, involves finding pixels that differ in color from at least one of their surrounding pixels.

The level of programming and mathematical skills needed to implement sophisticated vision algorithms makes is difficult for many hobbyists to experiment with this subject. Fortunately, a program called RoboRealm provides a solution for many situations by providing a large library of modules that makes it easy to capture images from web cams, and then process and analyze those images. Detailed information — as well as a 30 day trial copy of RoboRealm — can be obtained from www.RoboRealm.com.

Although RoboRealm is a programming language itself, it is very different from other languages you might be familiar with. Let’s look at some examples to see how RoboRealm can be used to implement a simple vision system. When you first start RoboRealm, you see the opening screen shown in Figure 1.