| Summary: | Nowadays, the most influential automotive industries and other big companies such as Google have been putting efforts in order to launch autonomous vehicles into our streets. In order to that safely, a autonomous vehicle has to be capable to move from point A to point B paying attention to every pedestrian and vehicle, baring in mind that road rules must be obeyed. Moreover, obstacles may appear unexpectedly or road adhesion could be altered. Therefore, in order to fulfill the main purpose of this project, two stereo matching methods were addressed so the two of them could be compared. The first algorithm was a local block matching and the second one was a semi-global matching. After the disparity map calculation, two histograms from the map were calculated: a vertical and horizontal histogram. With these histograms obstacles positioning and dimensions could be extracted.
|
|---|