University of Twente Student Theses

Login

Machine Learning for Ground Cover and Hot Target Analysis in RGB and Satellite Imagery

PINGEN, G.L.J. (2016) Machine Learning for Ground Cover and Hot Target Analysis in RGB and Satellite Imagery.

This is the latest version of this item.

[img] PDF
17MB
Abstract:Ground cover can be used in smart agriculture applications to perform effective automatic treatment of crops, for instance by finding an optimal distribution of water or pesticide administration. This is especially helpful for farms located in third world countries, that have little access to high-tech machinery and a lot to gain in terms of crop yield. Existing methods of ground cover analysis rely on converting original RGB features to a colour index. These methods are fast, but prone to misclassification. In a similar fashion, existing methods of hot target (wildfires) detection transform the original multispectral feature space using simple logic functions, and encounter similar problems. The aim of this research was to investigate how machine learning (RF, SVM, MLP, DNN) can be used to perform ground cover analysis of RGB smartphone photography, and hot target detection in multispectral satellite imagery. We obtain state-of-the-art performance in ground cover analysis using our DNN architecture, and outperform conventional methods of hot target detection using SVM/MLP. Our deep neural network implementation is less suited for hot target detection. Although we mainly see positive effects of machine learning in these domains, one disadvantage of learning-based approaches is the necessity of large quantities of labelled data.
Item Type:Essay (Master)
Clients:
TNO, The Hague, The Netherlands
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:38 earth sciences, 48 agricultural science, 54 computer science
Programme:Human Media Interaction MSc (60030)
Link to this item:http://purl.utwente.nl/essays/71278
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page