Skip to content

A flutter app to detect potholes during the journey and report it to PWD.

Notifications You must be signed in to change notification settings

nihara75/Pothole-Alert-App

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Prototype

• The App was designed using the Flutter framework • Google Maps functionality was integrated to it using a plugin and an API key • Users could send the location of a pothole through an email if they report it

Dataset

We used Pothole dataset of Kaggle with our own pothole image colections.It consists of 10,000+ images. All these images were preprocessed using opencv. The preprocessing done includes resizing, converting to grayscale, and adaptive thresholding. All the images are treated as numpy arrays. After preprocessing each image, they’re added to the trainX list and the corresponding labels are added to trainY list.

Training

The training was done using Google Colab so that we could get TeslaK80 GPU for faster and efficient training of the network. After preprocessing dataset i.e.creating labels file , splitting up the list of training and testing data the images and the list of their locations are kept together.The yolo.cfg file was used for training configurations which include fifteen yolo layers.

Colab Link

Colab

*Before use: Need to provide your own API key to the code

About

A flutter app to detect potholes during the journey and report it to PWD.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages