Pro Vision

Team

Table of Contents

  1. Introduction
  2. Problem Statement
  3. Objectives
  4. Methodology I
  5. Methodology II
  6. REFERENCES
  7. Links

Introduction

Pro Vision is a project focused on removing weather effects like rain, fog, haze, and snow from images to restore their quality and detail.. Traditional image processing methods fall short, making this enhancement crucial for applications such as autonomous driving, surveillance systems, outdoor vision-dependent systems, and satellite imaging.goal of this project is to restore the image quality by removing the weather effects

Problem Statement

Due to bad weather conditions obscuring key details in images, reducing the contrast and clarity and also adding noise and artifacts, the visual and logical output from the image is poor. Hence the reliability is not ensured;analysis is therefore redundant in many cases. The purpose of this project is to address this problem domain

Objectives

Main Goal of this project is to produce clean and clear images by removing weather effectswhich degrade the image quality in the means of a robust method ● Ensure solutions work for multiple weather conditions (general method) ● Create an efficient system for real-time and large-scale applications. ● Integrate advanced algorithms for optimal results.

Methodology I

Physics-Based Approach ● Overview:Involves modeling and estimating physical parameters like transmission maps, atmospheric light, and scene radiance to restore weather-degraded images. ● Key Parameters: ○ Observed image ○ Transmission map (light attenuation due to weather) ○ Atmospheric light (scattered light from the medium) ● Steps: ○ Estimate transmission map using DCP (Dark Channel Prior). ○ Estimate atmospheric light. ○ Restore image using a predefined formula. ○ Apply post-processing for contrast enhancement. 1 ● Challenges: ○ Estimation Accuracy: Requires precise estimation of transmission maps and atmospheric light. ○ Severe Weather Limitations: Struggles with extreme conditions without advanced techniques.

Methodology II

Specific Weather Condition Methods Overview: Tailored implementations for specific weather effects, optimizing results for unique conditions. ● Characteristics: ○ Provides more efficient outputs. ○ Optimized for intended weather conditions. ○ Requires more computational power.

Weather Condition Proposed approaches Challenges
Rain CNN, GAN, Image Decomposition Data needs, model complexity,real-time issues
Haze L0 gradient, Dictionary Learning,Guided Filtering Sparse representation, training,scalability
Snow Contrast Restoration, Multi-ScaleCNN, Histogram Stretching Physical models, parametersensitivity,computational load