Feature extraction using unmanned aerial vehicle

Abstract

The aim of this paper is to capture images of real time scenarios where frequent monitoring is impossible by human beings, with the help of an UAV, and analyse it through image processing. The processing time involved in Image Processing of previously proposed model, using Bag of Words is very high. In order to improve this, we are using a pre-processing method called Pseudo-Feature Extraction, on the captured image before passing it to Bag of Features. By using the above method, it can be observed that time efficiency was improved for most of the test cases. Further improvements can help industrial and agricultural activities and wildlife monitoring.