You are on page 1of 1

Abstract

Garbage cleaning along the railway tracks has become a tedious task as manual workers can’t trace
where the trash has been located. But many of the tourists have complained about the same issue of
unnecessary waste along the railway tracks, so we thought about the same and decided to solve the
problem by detecting trash along the railway tracks and give a report which consists of an amount of
trash located in different locations along the track so that it is easy for railway department to send
manpower to a particular location where more amount of trash is located using this we reduce time and
manpower and we achieve garbage-free environment. We use machine learning technique’s especially
YOLO V4 tiny model in order to identify the scrap present in the inputted video or image and we use
OpenCV to process the video or an image and we create a simple GUI (Graphical user interface) in order
to input our video using .NET Framework. We have compared the results of the yolov3 model to a
yolov4-tiny model which has given interesting results about the accuracy and speed of execution. The
results show that the detection of scrap can even be done on simple CPU systems with decent accuracy
and we can give a solution to a real-world problem using simple classification models.

Key words: OpenCV, WinForms, .NET Framework, YOLO (You Only Look Once)

You might also like