OpenCV-Quality-Degradation-Detection-in-Cameras-for-Self-Driving-Vehicle-Applications

This work consists of a convolutional neural network created with Tensorflow amd Keras libraries in Python, future work impending.

Purpose

The self-driving vehicle industry has grown exponentially over the past decade, and the impact it will make on the lives of many billions of commuters across the world everyday is undeniable. However, the current self-driving vehicles are faced with many environmental factors such as rain, dust and other obscurities that can impact their sensing capabilities and can increase the chances of an accident beyond acceptable safety limits. This project implements a trained Convolutional Neural Network that can predict the degradation in the quality of camera images during their operation, and can help avoid accidents by timely cleaning of camera sensor.

Accompanied paper

Here is the link to the accompanied paper

Link to saved Model with Trained Weights

Link to zip file for saved model to be found here here

Future Directions

Future plans are to finish the development of GAN model for performance comparison as well as more testing with larger data sets.