Segmentation Improvement of High Resolution Remote Sensing Images based on superpixels using Edge-based SLIC algorithm (E-SLIC)

Abstract

The segmentation of high resolution remote sensing images is one of the most important analyses that play a significant role in the maximal and exact extraction of information. There are different types of segmentation methods among which using superpixels is one of the most important ones. Several methods have been proposed for extracting superpixels. Among the most successful ones, we can refer to SLIC method. This method has some disadvantages among which can refer to over segmentation and noncompliance with the real objects. Here, in this study, we have tried to overcome these drawbacks and propose a novel method for segmentation of large-scale images by adding edge information to the SLIC algorithm. Three different urban data including airborn and spaceborn images with high space resolution and different objects diversity have been chosen with evaluate the proposed method. The results of the proposed method have been compared to the original SLIC algorithm and other common superpixel segmentation techniques, such as DBSCAN, and superpixel segmentation with entropy rates. The quantitative comparison of the results with the help of the standard deviation parameter within the class (WCSD) shows that in case of satellite images with an average of about 780 and 1040 units and in the case of aerial images with an average of about 220 units, the standard deviation of the produced segments in the proposed method is less than the other competing methods. The visual comparison also indicates that the components produced by the proposed method have the lowest standard deviation and are homogeneous.

Image Segmentation and Grouping Using Graph Method (SLIC Algorithm)

“…By focusing on both color similarity and spatial proximity, SLIC efficiently creates compact and uniform superpixels. It is employed in a variety of use cases, including image segmentation, fruit and cell recognition, and remote sensing image analysis (Kakhani, MOKHTARZADEH & Valadan, 2021).…”