Image-based Indoor positioning system (Backend API)

Introduction

Image-based indoor positioning system allowed user to know their location in HKUST by simply take a photo or upload a existing photo. This system consist of the following project:

Web Application Frontend
Mobile Application
Backend API
Image Extraction Program

This project is the Backend API of the system.

Getting Started

Software Requirement

Node
Anaconda

Environment Setup

Download the project
Download the model, FLANN database, test set and training set image here and put it in the project directory.
Create a directory Query/Testing
Activate an anaconda environment in the project directory using python version 3.7
Install packages using the following command

conda install opencv tensorflow-gpu matplotlib keras pillow

If you do not have a GPU please install tensorflow instead of tensorflow-gpu.

Run the Backend API

You can run by the following command :

node index.js