Method for recognition and capture standardization for imagens of keys in maintenance protocols of Radio Base Stations

Abstract

The paper presents a dynamic key/non-key image verification problem for a Radio Base Stationmaintenance protocol and utilizes feature extractors and computational intelligence classificationmodels to develop a tool to optimize the human work involved. Image moment extractors and deepnetwork architecture are experienced as feature extractors and then a Random Forest Classifier andSupport Vector Machine to determine the class of the analyzed images. The results are presented atthe end, performing in the best case described 97% of accuracy in the problem