Bridge Health Monitoring System using IoT🌉

Viraj Shetty — Fri, 22 May 2020 09:38:18 +0000

In this post, I'll be describing about the Undergraduate Capstone Project that I created along with my teammates
Parth Doshi @parthketandoshi
Ajinkya @ajinkyatungare24

Infrastructures are subjected to constant corrosion making them vulnerable to degradation. Of all different types of structures, bridges are the most vulnerable as they are exposed to various natural factors and if the corrosion is not checked, it can lead to loss of Human Life and property. An IoT based Structural Health Monitoring System can allow us to collect real-time information about different parameters from the bridge and prevent a disaster from happening.

Audio Analysis using python.

Description

When we consider the sound a surface makes on impact with a hard object - the sound for a hollow material compared to a dense material will be different. Considering this we can analyze a sample of audio of a dense material and the same dense material over time when it deteriorates. We know that over time it will sound different if it has any deformities or cracks Making use of this idea we created an application.

Description of the application

This is a python based application which takes two audio recordings and compares them, outputs their respective Mel’s spectrogram
and calculates the numerical difference of the two audio files. Here the first file will be of a recording of the bridge in perfect health and in perfect condition. Over time the bridge will have a certain degree of deterioration. This time another audio sample will be taken and this time it will be compared to the initial audio file in the python based audio analysis application.
The numerical difference will be calculated and Mel’s Spectrogram of both the audio files will be plotted to visually notice any changes.

End Result

We will have the audio sample difference over the interval and the authorities can keep a note of all the changes happening in the bridge and check for any deformities or abnormalities in the bridges. For inspections too the data from the audio analysis sample can help figure out which parts of the bridge are facing deterioration. The application also has a system to send the data through email to the authority - the email will have the two spectrograms and the result data.

Future Scope:

Making use of Robotics to automate the process of capturing data on the sites
Using machine learning models to generate the desired output without difference

Image Analysis using Python