65 / 2017-02-10 17:15:02

Convolutional Neural Network for Power Grid Equipment Classification

12793,12792,10394,12791

Research on image classification
has become a hot topic in the field of computer
vision. The traditional algorithms use
feature extraction together with machine learning
to classify. However, expert knowledge is
needed to define the features and feature extraction
depends greatly on the kind of power
grid device to be classified. A different approach
for the task is to use Convolutional
Neural Network(CNN) for classifying different
kinds of equipment. CNN has the ability
to perform feature extraction and classification
in the same architecture with no need of expert
knowledge. Tested on photos taken in a power
grid, the proposed CNN method achieves better
performance than some of traditional algorithms.