innerve-hackathon

This project is for the igdtu data science hackathon 2020

Project : Path detection for visually impaired

What it does?

Main objective is two things : Lane detection and object recognition.

Lane detection is done through two methods:

1. OpenCV houghes transformation
2. Using CNN

Object detection is done through a standard implementation of the YOLO algorithm.

Overview

Lane detection using Hough Transformation

• Used openCV for producing a lane detector. Inherent techniques used were Canny detection and Hough transforms.
• Used Convolutional Neural Networks to work upon the limitations of the openCV model.
• Used the standard YOLO algorithm for the implementation of the object detectors.

You can view harshit's repo to know more.

Lane detection using CNN

• Used a custom built openCV app to label frames of the dataset (a valid lane to go through)
• Made a sequential CNN that learns the dataset
• Predicts a valid path and direction to go.

Object detection using the YOLO algorithm

• Implemented the YOLO algorithm to detect object on the streets
• Detected various class of objects ranging from people , car and animals
• Used a very deep Darknet-53 model to predict the same

How to run?

1. OpenCV lane detection :

       Run the ipynb in colab

2. CNN lane detection :

       Add validation values from dataset_labeller/val_labels to your google drive

       run Untitled10.py in google colab

       Remember to upload model_1.h5 to your runspace (can be found in the respective folder)

3. YOLO implementation

       Download weights from here

       Add images to your gdrive/val2017/

             Add weights and everything in the files folder to your gdrive/files/

       Run the ipynb in google colab