China
Africa
Explore chapters and articles related to this topic
Conclusion
Published in Yusuf Arayici, John Counsell, Lamine Mahdjoubi, Gehan Nagy, Soheir Hawas, Khaled Dewidar, Heritage Building Information Modelling, 2017
Yusuf Arayici, John Counsell, Lamine Mahdjoubi
Semantic feature recognition is a very interesting development. By using rudimentary information like the direction of gravity, the relationships between clusters of points can be refined to the point that building elements are recognised. However, accuracy is still a problem when greater areas are segmented and the boundary tracing suffers. By developing the ability to create an outline or wireframe structure of the geometry, the data size can be reduced significantly in the subsequent phases. The same process of linear regression can be used to isolate normals that conform to a preset threshold when dealing with 2D data.
WatershedBased Ice Floe Segmentation
Published in Qin Zhang, Roger Skjetne, Sea Ice Image Processing with MATLAB®, 2018
This boundary tracing algorithm detects the exterior boundary of an object in a binary image. Any holes that are present or any objects within holes are ignored. To trace the boundary of a hole within an object, which represent an interior boundary of the object, a hole detection algorithm (e.g., a hole filling algorithm as introduced later in Section 7.1.3) should be used first to extract the hole, and then the boundary tracing algorithm should be applied to the hole (and any other hole found).
Casualty Identification with Dental Radiographs and Photographs
Published in IETE Journal of Research, 2023
B. Vijayakumari, M. Vasanthal, S. Dhivya Dharshini
Contour tracing plays a major role in feature extraction. Anil K. Jain and Hong [23] proposed a semi-automatic segmentation method where the user manually selects the Region of Interest (ROI) and then the image is divided into horizontal integrals and vertical strips which are used for gap valley detection. To find a smooth curve passing through the lines in each strip, the spline function is used. G. Dibeh et al. [24] proposed a novel approach for separation, which shows the robustness of the method to obtain segmentation over image artefacts, especially teeth filling, skewed and overlapping teeth. Segmentation using the active contour model [25] is also available in the literature. Boundary tracing is implemented by Moore’s neighbour tracing algorithm modified by Jacob’s stopping criteria. It works under eight-pixel connectivity. The algorithm terminates when the start pixel is visited for the second time. Two different effective alternatives to existing stopping criterion were proposed by Jacob. (i) Stop after visiting the start pixel n times, where n is greater than or equal to 2. (ii) Stop after entering the start pixel a second time in the same manner it was entered initially. Since the output of this algorithm is not clear and provides more unwanted details, another method is implemented using morphological operations. Two fundamental morphological operations are erosion and dilation. The algorithm of morphological extraction is as follows:
Generation of patterned indentations for additive manufacturing technologies
Published in IISE Transactions, 2019
Ulas Yaman, Melik Dolen, Christoph Hoffmann
On the other hand, the FDM process embodies a number of technical challenges. That is, the developed programs (i.e., M-scripts) accepting binary bitmap images as inputs generate the “indented” images for each and every layer depending on the texture function selected. Contours of the modified images, which are represented as 2D-polygons, are then obtained by boundary-tracing methods. The post-processor specifically devised for an Ultimaker 2 Go 3D printer employs the vertices of these polygons to generate the corresponding NC code. Table 2 tabulates the FDM printing parameters. To minimize the fabrication time, as well as the amount of deposited material, only the outer shells (with 1.2 mm thickness) of the artifacts are fabricated.