Classic Computer Vision approaches to finding objects and performing tracking in images are based on finding corresponding keypoint-descriptor pairs (SIFT, SURF, BRIEF etc) between images. This however relies on object texture in which the keypoints can be uniquely identified. There are many situations where object does not have any texture, e.g., a coffee mug. In industrial applications this is also frequently the case for metal parts. Another example is when depth imaging cameras are used (Time-of-Flight, Stereo) which does not capture texture at all.
There is a recent Computer Vision research area known as Textureless Object Detection that uses edge features for the object localization instead of keypoints. The task of this Master Thesis is to survey this literature and evaluate methods for industrial applications such as Robot Random Bin-Picking and object detection in Time-of-Flight images. Experiments in this thesis topic can be carried out in Matlab, Python or C++
Video of Textureless Object Detection:
For more information, contact R&D manager Ola Friman (email@example.com) and apply below at "ANSÖKAN" before the 19th of October. Please notice: Application must include CV, Cover letter and LADOK (study results including courses).