A new contour tracing algorithm for reconstructing 2D shape of an object template with parametric curves
Is this your thesis?
This record came from a bulk archive import. If it’s yours, link it to your profile.
2022
0 views
0 downloads
Advisor: Dr. Öğr. Üyesi Kali Gürkahraman
Abstract (EN)
Parametric curves such as Bézier and B-splines are widely used in computer aided design, image processing and computer vision. Using these parametric curves, an object in an image can be reconstructed at different scales and orientations. In this study, a novel method that is robust to scaling and rotation problems has been developed by using parametric curves, since performing the template matching is challenging in scaling and rotation conditions. The prerequisite for creating the object shape with parametric curves is to obtain sequentially the points in the point cloud belonging to the outer boundary of the object. Since traditional edge detection algorithms do not obtain the points sequentially, a novel algorithm has been developed in this study, which sequentially detects the points constituting the outer boundary of the template. Unlike other contour tracing algorithms in the literature, our algorithm is implemented using a filter containing weights. The filter, which is located at a certain position in the binary image, produces a unique numerical value that determines the next tracing direction thanks to its weights. Sequential boundary points obtained with the developed contour tracing algorithm are used both in the reconstruction of the curves representing the object shape and in the template matching process. The template matching method used in this study is carried out in three steps. In the first step, the outer boundary points of the template and matching objects are obtained sequentially with the developed contour tracing algorithm. In the second step, binary images containing the object shape are created with the B-spline curve created using these points. In the last step, a signal is obtained by calculating the distances of the boundary points of the objects in the 2D image created with B-spline curve to the shape center, and this signal is used for template matching. The performance values of all steps were determined experimentally. The experimental results of the matching method using the outputs of the contour tracing algorithm show that our tracing algorithm is successful.
Author
Nihat Arslan
Institution
How to Cite
Nihat Arslan (Master Thesis). A new contour tracing algorithm for reconstructing 2D shape of an object template with parametric curves, 2022, Sivas University of Science and Technology.
Keywords
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Sivas University of Science and Technology
- Control of surface wettability in polymethylmethacrylate films(2026)
- Effects of different endophytic bacteria on in vitro growth of fenugreek (trigonella foenum-graecum)(2026)
- Investigation of the effects of permanent magnet synchronous machines on the power system parameters within the scope of microgrid(2026)
- Smart contact lenses: A literature review on dual-use scenarios in healthcare and defense industry(2026)
- Investigation of the joining of additively manufactured aluminum to copper by friction stir welding(2026)
- 3D profiler with led source and 2D continuous wavelet transform(2026)