Representing Data in 3-D

Before you start producing 3-D models with your Kinect, you must know how to represent and store those models. Most developers are familiar with how 2-D raster images are stored. The image consists of a regular grid of pixels that sample the 2-D space. The image resolution (i.e., how densely the 2-D space is sampled) determines not only the quality of the image but also the processing and memory requirements. Take the Kinect color image, for example (see Figure 7-1). It has VGA resolution (640 × 480), which means it has 307,200 pixels. But if we double the resolution (1024 × 960), the pixel count is four times higher. Fortunately, modern computers are able to process images of several megapixels in real time without problems. Yet, when we deal with 3-D data, there are many representation options, each with its own advantages and disadvantages.