| Summary: | In this Master’s thesis, algorithms for autonomous lunar crater detection for a satellite manoeuvre simulation vehicle (SMSV) has been investigated. The SMSV is driving on a zero gravity surface under an artificial moon surface. The vehicle has a camera attached on the top. That camera will be used for the detection of craters. There are a great amount of different approaches towards crater detection in miscellaneous articles. Such as, Hough circle transform, Canny edge detection, Ellipse fitting and the training of a cascade classifier by using Haar-like features. All the different approaches were implemented and compared, until a final version of the best algorithm was found. It was established, that the best way of detecting craters at an artificial moon surface, containing irregular shaped craters, is to train a cascade classifier using Haar-like features. This approach was also compared at different stages of the classifiers and different classifiers were also compared.
|
|---|
