29 March 2024


Monitoring the condition of forest areas is essential for effectively identifying tree diseases and preventing their spread. The use of high-resolution satellite images enables the regular acquisition of up-to-date information over large forested regions. The processing and analysis of such data can be automated using machine learning methods capable of detecting signs of abnormal vegetation changes, which may indicate the presence of diseases.
In this work, SWIFTT partners from the Space Research Institute of Ukraine and collaborators explore the use of the Bhattacharyya distance and Spearman’s rank correlation coefficient for feature selection from satellite images. A greedy algorithm was employed to identify a subset of weakly correlated features. Experimental results indicate that the selected features improve classification quality compared to using all spectral bands. The proposed approach demonstrates effectiveness for selecting informative and weakly correlated features and has the potential to be applied to other remote sensing tasks.
Read the paper in the link below.
-
Next Article
ACM/SIGAPP Symposium On Applied Computing -
Previous Article
Workshop Copernicus for Forestry - View All Articles