In this dissertation, a heuristic method for object pose detection for automated feeding systems is presented, which was successfully tested in industrial applications. Testing of the procedure is done on the basis of extensive data from industrial applications, which were obtained at ten automated feeding systems, with three families of parts and a total of 15 types of workpieces. To detect the scene, typically a carrier filled randomly with workpieces, any sensor system can be used for depth measurement from which a point cloud is obtained. The object pose detection then is considered as a combinatorial optimization problem, for which a construction heuristic is applied. For this heuristic tree search, a finite set of possible workpiece poses is initially derived from the continuous search space. In order to use a decision tree, the elements of the search set are split into two components. The first component describes an interesting point in the search space, which is part of the workpiece surface. The second component describes possible workpiece poses relative to an interesting point. The partial search quantities obtained thereby have a significantly lower complexity compared to the original search set, since the interesting points can provide a constraint on the relative workpiece poses, thus restricting its freedom of movement. The applied tree search strategy is best-first search. Best-first search explores the search tree by always expanding the most promising nodes first. These nodes are chosen according to a heuristic evaluation score, representing the estimated distance from the node to a solution. Final evaluation of the workpiece poses is provided by a six-dimensional Hough voting procedure, which is also known as Generalized Hough Transform. The features used for Hough voting are sensor measurements considered relatively to an interesting point. For all possible constellations of a sensor measurement relative to an interesting point, a probabilistic statement about possible workpiece poses can be made. Through the superposition of all probability statements, solution candidates can be formed, which are subjected to a statistical test based on a quality rating. The obtained quality rating along with a given level of significance is used in order to decide about the acceptance of a workpiece pose. The method is able to locate three to four workpieces on average within 0.5 and 3 seconds, using a standard desktop computer.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Heuristisches Suchverfahren zur Objektlageerkennung aus Punktewolken für industrielle Zuführsysteme


    Beteiligte:

    Erscheinungsdatum :

    2014



    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Deutsch






    Planung von Just-in-Time-Belieferungen mit lokalen Suchverfahren

    Urban, Karsten P. / Urban, Karsten-Patrick | GWLB - Gottfried Wilhelm Leibniz Bibliothek | 2004


    Planung von Just-in-Time-Belieferungen mit lokalen Suchverfahren

    Urban, Karsten P. ;Urban, Karsten-Patrick | SLUB | 2004


    Planung von Just-in-Time-Belieferungen mit lokalen Suchverfahren

    Urban, Karsten P. / Urban, Karsten-Patrick | TIBKAT | 2004