The goal of our project is to design a Pothole detection System. which assists the driver in avoiding potholes on the roads, by giving the driver prior warnings. For example, it can be like a buzzer or series of LED, if the driver is approaching a pothole driver may be warned regarding the pothole on the road. In this paper, we propose a robust and straightforward design of a portable and affordable device that can alert the driver about the detected pothole. The hardware system installed in a moving vehicle can automatically detect and report potholes via image-processing of Raspberry-Pi microcontroller. The detailed images of the pothole and its location are stored and viewed through the GOOGLE API.

Index Terms — Pothole Detection, Hardware system, Raspberry-Pi microcontroller, GOOGLE API.

In this fast-moving world, people want to reach their destinations as soon as possible. Some apps suggest the routes which can make us reach our goal early, and other apps show all possible ways with traffic congestion for whatever reason it may be, such as google maps and many more.

But there are fewer apps that tell the condition of the road, whether it is good to travel or not, whether it suggests the route to the driver based on the state of the way. So this model mainly focuses on the travel safety of the passengers and updates the passengers with the best route to travel.

• The main goal of the team is to develop a model that makes the travel safer by suggesting the best and safer route.

The embedded device processes the incoming stream of video and detects any pothole in the flow, if any, it immediately intimates the user with a signal. It also saves the information of pothole detected and coordinates in a file for cross-validation.

The below are all the stakeholders taking part in the total scenario:

• User(driver)

• DataAnalyst who takes the data provided by the devices and perform some data analysis stuff.

• Software Developer who debugs any issues after the installation

• Company which produces these systems

• IT Staff who installs and maintain the system.

People in this time need instant solutions for problems; one of the issues is that they need to reach their destinations early, and many applications can do this work. Still, what our app does differently makes us stand out from others, it not only suggests which road reaches your destination faster but also indicates which route is most safe to travel.

The work which can be done by our app is the most critical work, which makes it to find it's way into the minds of the users.

The embedded device processes the incoming stream of video and detects any pothole in the stream, if any it immediately intimates the user with a signal.

It also saves, the information of pothole detected and coordinates in a file for cross validation.

The system can also provide it with three features like:

• Detect pothole and Alert driver

• Alert the driver about the pothole

Type of software which we are going to use is python scripts, drivers are our expected users, are going to use this in embedded systems.

The user needs not to provide any of their confidential details for the module to work.

For the module to work, the car should be running and should also limit their speed to under 40kmph.

As we are using a pre-trained model for detecting the potholes, the time constraints will be on alerting the driver about the pothole and suggesting to him the best and safe possible route.

The system restarts once for every 6 hours of working, which will flush all the cache, additional to this we need to run the monthly updates to keep the software upto date. As the lifespan of the pi camera is less compared to raspberry pi we need to change it once every 5 years.

One can create an account/login using his driving license.

The following are the hardware sub-systems in the model:

• Raspberry Pi 3
• Raspberry Pi 5MP Camera Modules
• Mini Night Vision Camera 1080P
• Jumper Wires

The following are the software sub-systems in the model:

• AWS CloudBuild
• AWS Config

The sensor nodes on the car and the GPS reciever can detect or identify the potholes on the first note, or the manual recordings can also be used to update the location of the pothole to the data file which runs through a pothole detection algorithm which then can give the accurate postion of the detected potholes.

Software test automation tools enable you to simplify testing and reduce time to release by automating functional tests for your applications.

Unit testing tools which we are going to use are:

• Pytest
• Travis CI

For validation testing, we use the images which we have collected from sites to check wheather all the scripts are bug-free. This will also be carried out during the demonstartion to our stakeholders(excluding developers).

For defect testing, we record some real-world videos of road and present it, which not only includes images of road but, also images of some patterns which are similar to potholes, to measure how accurate is it detecting potholes, when both potholes and potholes like patterns are present.

Replacement of raspberry pi should be made once every five years and additional updates every month. AWS Config is a service that let's assess, audit, and evaluate the configurations of AWS resources which we are using.

The version control system which we are using is git. AWS CodeBuild is a fully managed continuous integration service that compiles source code, runs tests, and produces software packages that are ready to deploy.

To make the python scripts standalone, we are using a package called pyinstaller. AWS packaging automates the process of packaging and publishing software to managed Windows and Linux instances across the cloud landscape, as well as to on-premises servers, through a single simplified interface.

One of the risk which the device faces is external accidents/shocks which are unpredictables example:

• Someone hitting it
• Snow covering the lens
• camera being covered by some object(paper).
• Rain water going into the circuits and frying them.

The other kind of risks which the device faces is privacy and security of data. Time to time updates about the potholes to the cloud and intimate the driver about it correctly. The data should be transmitted only in a secure environment.

The above hardware risks can be avoided if the device have a dedicated slot where it is safe enough and should detect the potholes as well.

The privacy and the security of the device should be maintained by a good security analyst so that it would not be lost. And the software developer should schedule time to time checks about the cloud management and the software.

The challenges in front are, the budget as we cant take that to a high budget model so that it can be used by many and should get a high pixel camera with a low budget and the model should not do false interpretations of the pothole and make the user certain about this model.

• To make it budget friendly.
• Reduce the latency between the uploads.
• Detect potholes even when the car is at very high speed.

There could be even more possibilities of extending the model to a state at which it not only detects the potholes but also detecting the fixed potholes and updating the earlier data and detecting speed brakers where the drivers need to slow down. Another extension that we can make is to estimate the depth of the pothole and categorizing them to different deifferent level of danger and alert the driver about that danger whether he can travel on that road or not. And can be notified to higher officials for the fixing of these potholes.

There should be enough light to detect the pothole, and car should be moving with a average speed of 40-60km/hr to get a good quality of the video.

Identify who will serve on the CCB, which determines whether issues are within the current project scope and whether they should be addressed.

Dataset link: link