Detekce a odhad 3D pózy objektů v RGB-D obrazech

Abstract

The thesis deals with the implementation of the algorithm for fast detection of 3D textureless objects and subsequent determination of their pose in the image. Each object is represented by the set of images depicting the object from various angles (poses). The thesis is divided into theoretical and practical part. Theoretical part introduces used methods, dataset and related work. The practical part deals with the implementation of the algorithm itself and thoroughly describes the individual parts of the detection cascade. Procedure for fine pose estimation of detected objects is further described. Finally, a final summary is made, including experiments, evaluation and presentation of measured results.