API Integration and Back-end Development.
To be updated soon!
Below attached is the project report.
Final Year Project Report
Several studies have been performed in marine biology where the fish population is counted
using volunteer divers. This has several drawbacks in comparison with underwater camera
surveillance. Some fish species will hide themselves if divers appear, while, in the case of
camera surveillance, the fish get used to the cameras. In recent years,
underwater cameras have been used to study coral reefs. Several approach have been developed
to detect and recognize fish using underwater camera footage, but this work is mainly restricted
to more constrained environments. This project uses an improved version of the algorithms, where
both fish detection and tracking are performed on live video footage recorded. The objective is to
monitor the school of fishes underwater and count them. The prototype includes setting up an
environment of fish in an aquarium, and the species used for experimental purpose is Guppy Fish
The idea is to recreate the ocean environment in an aquarium and use it as a prototype to detect and track fishes. As the fish moves, the input video feed of their motion is taken and bounding box is applied onto them using background subtraction. Since there was a drawback when the camera would move, another technique named linear regression and histogram of oriented gradients is implemented where each fish is individually tracked and its path is monitored. We monitor the paths so that we can keep a count of the number of fishes to tabulate their population which is an input for methods implemented in the defined future scope.
Contact me to see other related documents and see the working of the app. Below attached is a document for basic reference.
Smart Fish Tracking
Trash Out is an Android App developed using Android Studio. It takes the user's location from the mobile phone and fetches all the nearby geo-tagged waste bins so that the user can easily dispose any trash. The co-ordinates of the waste bins is taken from a pre-fed Database and using Dijkstra's algorithm, from the user's location to the various nearby bins, the nearest bin is pin pointed and the direction is loaded in the map shown in the app, using Google Maps API. Three classifications are done based on the type of trash as :
This app was presented at Hack2Help Hackathon held at Dayanada Sagar University in
January, 2016 and was awarded with a certificate of Appreciation. This app focuses
to promote the Swatch Bharath Andolan, and aims to create a cleaner and greener India.
Contact me to read related documents and see the working of the app.