Surface fitting is still an area of great research interest. In this paper two particular surface... more Surface fitting is still an area of great research interest. In this paper two particular surface fitting problems are discussed. First, various strategies to approximate complex free-form shapes are described, with special emphasis on the so-called functionally decomposed surface representation. This is strongly related to surface fitting when the set of data points to be approximated is incomplete; i.e. there are areas which belong to other surfaces or which are just not accessible. In these cases we have to somehow "bridge" the unknown areas, and produce an overall pleasing surface, which smoothly connects the surface portions where data is available. The algorithm described is a combination of previous approaches to minimize hybrid objective functions, with several practical improvements. The second problem aims at fitting surfaces where not only positional data but surface normal data needs to be approximated as well. After analyzing the consistency of the data and possible objective functions to be minimized, the solution, which was found to be the most suitable is presented. Both the above problems emerged due to practical demands and have been intensively tested using real data obtained from the automobile industry. Some colour pictures illustrate the results.
Abstract : Describes Dr. Tamas Varady's visit to Purdue University between 16-18 May 1993. He... more Abstract : Describes Dr. Tamas Varady's visit to Purdue University between 16-18 May 1993. He presented a seminar on the current problems of general topology surface modelling (polynomial patches for general topology surface modelling and vertex blending).
Surface fitting is still an area of great research interest. In this paper two particular surface... more Surface fitting is still an area of great research interest. In this paper two particular surface fitting problems are discussed. First, various strategies to approximate complex free-form shapes are described, with special emphasis on the so-called functionally decomposed surface representation. This is strongly related to surface fitting when the set of data points to be approximated is incomplete; i.e. there are areas which belong to other surfaces or which are just not accessible. In these cases we have to somehow "bridge" the unknown areas, and produce an overall pleasing surface, which smoothly connects the surface portions where data is available. The algorithm described is a combination of previous approaches to minimize hybrid objective functions, with several practical improvements. The second problem aims at fitting surfaces where not only positional data but surface normal data needs to be approximated as well. After analyzing the consistency of the data and possible objective functions to be minimized, the solution, which was found to be the most suitable is presented. Both the above problems emerged due to practical demands and have been intensively tested using real data obtained from the automobile industry. Some colour pictures illustrate the results.
Surface fitting is still an area of great research interest. In this paper two particular surface... more Surface fitting is still an area of great research interest. In this paper two particular surface fitting problems are discussed. First, various strategies to approximate complex free-form shapes are described, with special emphasis on the so-called functionally decomposed surface representation. This is strongly related to surface fitting when the set of data points to be approximated is incomplete; i.e. there are areas which belong to other surfaces or which are just not accessible. In these cases we have to somehow "bridge" the unknown areas, and produce an overall pleasing surface, which smoothly connects the surface portions where data is available. The algorithm described is a combination of previous approaches to minimize hybrid objective functions, with several practical improvements. The second problem aims at fitting surfaces where not only positional data but surface normal data needs to be approximated as well. After analyzing the consistency of the data and possible objective functions to be minimized, the solution, which was found to be the most suitable is presented. Both the above problems emerged due to practical demands and have been intensively tested using real data obtained from the automobile industry. Some colour pictures illustrate the results.
Abstract : Describes Dr. Tamas Varady's visit to Purdue University between 16-18 May 1993. He... more Abstract : Describes Dr. Tamas Varady's visit to Purdue University between 16-18 May 1993. He presented a seminar on the current problems of general topology surface modelling (polynomial patches for general topology surface modelling and vertex blending).
Surface fitting is still an area of great research interest. In this paper two particular surface... more Surface fitting is still an area of great research interest. In this paper two particular surface fitting problems are discussed. First, various strategies to approximate complex free-form shapes are described, with special emphasis on the so-called functionally decomposed surface representation. This is strongly related to surface fitting when the set of data points to be approximated is incomplete; i.e. there are areas which belong to other surfaces or which are just not accessible. In these cases we have to somehow "bridge" the unknown areas, and produce an overall pleasing surface, which smoothly connects the surface portions where data is available. The algorithm described is a combination of previous approaches to minimize hybrid objective functions, with several practical improvements. The second problem aims at fitting surfaces where not only positional data but surface normal data needs to be approximated as well. After analyzing the consistency of the data and possible objective functions to be minimized, the solution, which was found to be the most suitable is presented. Both the above problems emerged due to practical demands and have been intensively tested using real data obtained from the automobile industry. Some colour pictures illustrate the results.
Uploads
Papers by Tamas Varady