In the previous chapter you learned about the basic building blocks for implementing GPS and sensor-based AR applications. If you tried the different examples we presented, you might have noticed that the feeling of getting digital objects in real space (registration) works but can become coarse and unstable. This is mainly due to the accuracy problems of the used sensors (GPS, accelerometer, and so on) found in smartphones or tablet devices, and the characteristics of these technologies (for example, gyroscope drifting, GPS reliance on satellite visibility, and other such technologies). In this chapter, we will introduce you to a more robust solution, with it being the second major approach for supporting mobile AR: Computer vision-based AR.
Computer vision-based AR doesn't rely on any external sensors but uses the content of the camera image to support tracking, which is analysis through a flow of different algorithms. With computer...