Improve segmentation of hyperspectral images using area-limit and path-minimization methods
Subject Areas : Electronics Engineering
Ahmad Keshavarz
1
,
Fatemeh Hajiani
2
1 - Bushehr Persian Gulf University
2 - Islamic Azad University, Khormuj Branch, Iran, Faculty Member
Keywords:
Abstract :
In this paper, the segmentation methods of limiting the area and the minimum path were performed in order to collect regional information on the resulting areas using the area uniformity method. In the first stage, using the image smoothing method, a number of sub-areas were converted, then a more complete segmentation was performed on each of the areas of the first stage using the methods of limiting the area and the minimum path. The proposed method identifies new sections using regular seed selection and sum spacing. The area restriction method controls the amplitude changes from one point called the seed in each area and the minimum path method controls the size of the areas. These two methods were implemented on the image and increased the compression rate by reducing the number of areas.