China
Africa
Explore chapters and articles related to this topic
Thinning and Skeletonizing
Published in Edward R. Dougherty, Digital Image Processing Methods, 2020
An important morphological operation we will need later is the hit-or-miss transform, which is performed on binary images. Skeletonizing can be described in terms of this operation. Let a be a binary image and s and t be two binary templates that do not overlap. Then the hit-or-miss transform, abbreviated by HMT, is described in image algebra by a⊕(s,t)=(aΘ−s′)*(a¯Θ−t′)
Segmentation and Edge/Line Detection
Published in Scott E. Umbaugh, Digital Image Processing and Analysis, 2017
The morphological hit-or-miss transform is a fundamental method for detection of simple shapes. It is a basic pattern recognition tool that, like the previous morphological methods, uses a SE to determine the patterns or shapes it detects. In addition to the 1's (object) and 0's (background), the SE may contain “don't cares,” specified with an x. The hit-or-miss transform works by overlaying the SE on the image and requires an exact match for a “hit” to occur—a hit is marked with a 1. The following example shows a hit-or-miss transform that finds the upper right corner points of binary objects.
Quantification of Nanostructure Orientation via Image Processing
Published in Klaus D. Sattler, st Century Nanoscience – A Handbook, 2020
The preferred approach to compute the skeleton of 1D nanostructures involves the use of methods which preserve the topology of the original binary image, such as the use of the hit-or-miss transform [26]. The hit-or-miss transform involves the application of structuring elements, one associated with a local structure being present (foreground) and another being absent (background). Given the skeleton, additional applications of the hit-or-miss transform include using structuring elements which identify branchpoints and endpoints [15].
Automated Object Detection for Visual Inspection of Nuclear Reactor Cores
Published in Nuclear Technology, 2022
Michael G. Devereux, Paul Murray, Graeme West
We present two new automated methods for detecting and locating keyways. The first considers a tool from mathematical morphology known as the Hit-or-Miss Transform22 (HTM) combined with suitable pre- and post-processing. Mathematical morphology has been applied to a wide range of problems and achieves considerable success in the automated detection of various different objects in images ranging from ship detection in satellite images23 to galaxies in astronomical images.24 It is also very suitable for this type of problem due to the fact that it natively works on a pixel-by-pixel basis as opposed to an object basis. This framework is optimized using a genetic algorithm (GA) to provide a robust and transparent framework to automatically detect keyways. While prior research25,26 uses a GA to design a single structuring element (SE) to perform an image processing operation, our system designs multiple SEs, orders their use, and combines them in the most effective way to construct a classifier suitable for labeling individual pixels as keyway/not keyway.