A New Unsupervised Change Detection Approach Based on PCA Based Blocking and GMM Clustering for Detecting Flood Damage

08/22/2020 ∙ by makifgunen, et al. ∙ 0

In this paper, a new and effective unsupervised change detection technique is presented for Synthetic Aperture Radar Images (SAR) images. Unsupervised change detection is one of the important research topics in many application areas such as agricultural applications, monitoring of forests and urbanization, damage assessment after any disaster. SAR images are often used in change detection applications because they can provide images in all weather conditions unlike optical images. In this paper, a new change detection approach based on PCA blocking and Gaussian Mixture Models (GMM) clustering for SAR images was proposed. The approach firstly computes logarithmic difference image. Then, wiener and median filter are applied to reduce the effect of speckle noise in difference image and to preserve edge information. After that, the approach calculates eigenvector space from kk× blocks with PCA. Lastly, change detection map is computed using GMM clustering. Bern and Ottawa datasets were used to evaluate the proposed approach. Proposed method was compared with PCADS, PCA-FCM and PCA-Kmeans methods. Quantitative and qualitative analysis was conducted based on the Ground Truth (GT) images belong same geographical regions. According to results, proposed approach is quite successful.



There are no comments yet.


page 2

page 4

page 5

page 6

page 9

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.