Edge detection in multispectral images using the n-dimensional self-organizing map

Abstract

We propose a new method for performing edge detection in multispectral images based on the self-organizing map (SOM) concept. Previously, 1-dimensional or 2-dimensional SOMs were trained to provide a linear mapping of high-dimensional multispectral vectors. Then, edge detection was applied on that mapping. However, the 1-dimensional SOM may not converge on a suitable global order for images with rich content. Likewise, the 2-dimensional SOM introduces false edges due to linearization artifacts. Our method feeds the edge detector without linearization. Instead, it exploits directly the distances of SOM neurons. This avoids the aforementioned drawbacks and is more general, as a SOM of arbitrary dimensionality can be used. We show that our method achieves significantly better edge detection results than previous work on a high-resolution multispectral image database.

A Novel Framework for Interactive Visualization and Analysis of Hyperspectral Image Data

“…A typical SOM consists of a 1D array or 2D grid of model vectors (also referred to as neurons) that represent vectors in the original data space. The 3D SOM variant extends a 4connected 2D lattice to a 6-connected 3D lattice [54]. The third dimension often enables the SOM to learn a more accurate topology on hyperspectral data.…”

Colour edge detection based on the fusion of hue component and principal component analysis

“…Despite the fact that a great number of edge detection approaches have been proposed in some literatures so far, there is still a continuing research effort. Recently, the main interest has been directed towards algorithms applied to colour and multispectral images [2, 3]. Compared to the grey‐scale images, the colour and multispectral images contain richer information, and consequently, the edges of colour images with a more complex form of testing [4].…”

Edge Detection of Color Images Based on Improved Morphological Gradient Operators

“…It is an important attribute in image recognition [1,2]. Edge detection can sketch out the contour of the target thus make it more observable therefore make itself a significant preprocessing procedure of image processing and occupies an important position in computer vision and image processing [3,4].…”