|
|
Detection of a Semi-Structured Semi-Finished Product from a Defined Area Using Artificial Intelligence Methods
Zmrzlý, Jan ; Škrabánek, Pavel (referee) ; Juříček, Martin (advisor)
This thesis addresses the issue of machine vision in the context of Industry 4.0, with an emphasis on the detection of semi-structured objects from surfaces. The first part of the thesis discusses the theoretical aspects of the task, including selected machine vision algorithms and the use of neural networks in this area. Furthermore, a survey of the available methods for solving this problem is conducted, as well as the current state of the art of the EDUset ONE robotic cell with respect to machine vision. Based on the analysis, a hardware solution in the form of camera, lighting and other components is proposed. Subsequently, the design and implementation of different methods for detecting multiple types of objects is carried out, with emphasis on modularity, efficiency and accuracy. Finally, the work compares these methods and verifies their functionality in interaction with a real robotic cell.
|
guest ::
