Wyniki wyszukiwania - BazTech

1

An active 3D robot vision system for robotic welding applications

Lilas T., Kollias S.

Machine Graphics and Vision

|

2000

|

Vol. 9, No. 4

743-762

EN

In this paper we present development of an active three-dimensional robot vision system, conducted on a large gantry-type welding-robot. The constraints are a complex working environment and the size of the robot system, tackled by developing and using technoques from several research fields. In particular, laser lighting is combined with stereo vision and photogrammetry in order to obtain accurate three-dimensional measurements. Computational geometry methods based on Delaunay triangulations are used to generate a three-dimensional model of the object. Pattern matching and segmentation techniques are used to evaluate the model and identify its important features. Path planning techniques are used for guiding the tool around the object and adapting trajectories to fulfill the welding requirements. The proposed active vision system significantly increases the potential of the welding robot system, constituting a sophisticated sensor system that is effective for robotic applications.

2

Compact 3D model generation based on 2D views of human faces : application to face recognition

Karpouzis K., Votsis G., Tsapatsoulis N., Kollias S.

Machine Graphics and Vision

|

1998

|

Vol. 7, No. 1/2

75-85

EN

A face recognition and synthetic-natural hybrid coding tool involving frontal and profile views is presented in this paper. The tool utilizes the above mentioned 2D images-upon which certian protuberant points are automatically detectrd- and adapts a generic 3D head model (polygon mesh) according to the predetermined information gained by the availble views. This mesh provides shape information which -combined with texture information- is important for the rebustness of a recognition system. the main advantage of the proposed approach is its ability to overcom constraints raised from arbitrary variations in scale, rotation and orientation, which are dominant deterioration factors in the identyfication task. Besides the face recognition aspect, applications which involve geometry transmistion, such as teleconferencing, can take advantage of the proposed algorithm's ability to parametrically describe a complex organic model, such as a human head. During the texture map creation process, issues related to the luminance differences and rotation varians between the avaible views, are successfully dealt with. Regarding the adaptation of the polygon topology of the 3D model, we apply a set of localized transformations, in order to preserve the continuity of the human head surface. Besides this, the problem of the minimum organic model representation required is addressed. The aspect under which this issue is verged upon, is thad of the solution of a trade-off problem between low computational complexity and high approximation quality.

3

A dynamically trained neural network for machine vision applications

Doulamis N., Doulamis A., Kollias S.

Machine Graphics and Vision

|

1998

|

Vol. 7, No. 1/2

113-123

EN

A dynamically trained neural network is proposed in this paper proper for adapting the network performance to non stationary image or video inputs. The sheme includes, on one hand, a retrieval mechanism which selects the most approprite network from the system memory and, on the other hand, a weight perturbation procedure which adapts the network weights to the current condition. If no suitable network exists in memory, new weights and network structure are crated and them stored for future use. Experimental results are provided indicating the good performance of the proposed system to computer and machine vision applications.