An Efficient Algorithm for Earth Surface Interpretation from Satellite Imagery

Abstract

Abstract.Many image segmentation algorithms are available but most of them are not fit for interpretation of satellite images. Mean-shift algorithm has been used in many recent researches as a promising image segmentation technique, which has the speed at O(kn 2 ) where n is the number of data points and k is the number of average iteration steps for each data point. This method computes using a brute-force in the iteration of a pixel to compare with the region it is in. This paper proposes a novel algorithm named First-order Neighborhood Mean-shift (FNM) segmentation, which is enhanced from Mean-shift segmentation. This algorithm provides information about the relationship of a pixel with its neighbors; and makes them fall into the same region which improve the speed to O(kn). In this experiment, FNM was compared to well-known algorithms, i.e., K-mean (KM), Constrained K-mean (CKM), Adaptive K-mean (AKM), Fuzzy C-mean (FCM) and Meanshift (MS) using the reference map from the Landsat. FNM provided better results in terms of overall error and correctness criteria.

Thaiwrittennet: Thai Handwritten Script Recognition Using Deep Neural Networks

“…Then, the horizontal and vertical SIFT gradient ( ) is computed by Laplacian in x and y neighborhood pixels (Soimart & Ketcham, 2016a)…”

Location Estimation of a Photo: A Geo-signature MapReduce Workflow

“…However, feature detection produces only horizontal and vertical co-ordinates of the key-points from a photo [12] which are not enough for the landmark retrieval. It needs a descriptive vector (known as feature description) to describe the region surrounding the key-point [11][12][13] (also known as spatial information [14][15][16][17]). Feature descriptions of a photo are in term of a vector that represents the important visual contents of the architectural place or geography from a single photo [18][19][20].…”