Detection of Pipeline using Machine Learning Algorithm and Analysing the Effect of Resolution Enhancement on Object Recognition Accuracy

Evaluation of defects in pipeline installed in
underwater demands smart engineered solutions. Burying the
large pipelines in underwater facilitated for the purpose of
petroleum, LPG, potable water and other gaseous requirements
are increasing every decade. Underwater pipeline route
tracking during leakage or any emergency conditions is
necessary and is challenging. The detection of pipelines in
underwater is made easy through SONAR imaging and is
evaluated through state-of-the-art technology. So heuristic
approach is formulated here using the emerging machine
learning technique. The research framework is focused on
design and analysis using efficient machine learning algorithm,
which can detect buried pipelines. Various statistical
parameters are selected here for the comparison of highly
compete algorithms named convolutional neural networks
(CNN), Fuzzy C means clustering and Apriori Algorithm. The
system design is adopted using MATLAB IDE tool and
simulated results are validated and tested.