Research Project - Aerosol Analysis

Created by: Hevander Da Costa

Table of Contents

1. Overview
2. Video Walkthrough
3. Features
4. Updates
5. DFS Implementation
6. BFS Implementation
7. AStar Implementation
8. Edge and Area Detection
9. Image Preprocessing

Overview

Aerosol Analysis is a package designed in MATLAB for the quantitative analysis of aerosol plums and jets. The file Jet_isolation.mlx can be used to readily process raw images for background separtation, image smoothing, and image clustering. The package uses custome search-based and machine learning based algorithms for topographical analysis of plums, jets, or blobs. The various algorithms can be used to idenifity boundary velocities and sub cluster velocity characteristics; such as in a particle image velocity regime. This package was used for a study on risk analysis of various aerosol emission Link to Publication.

Video Walkthrough

DataProcessingPIV.mp4

Features

The following required functionality is completed:

• Edge detection algorithm with variable threshhold of detection
• Kmeans based image clustering with fluster percentage,
• Plum/Jet metrics: width, length, areas
• Boundary analysis of Aerosol for circumference
• Dijkstra's, Breath-first, Depth-first, AStar
• Image segmentation and smoothing

Updates

The following future features may be implemented:

• Graphing funcitonality for the percentage of various velocity ranges
• Dynamic tracing of morphing contigous blobs
  • bounding box method
  • Edge detection and continous search algormithm verification for close fit

DFS-Implementation

Depth First Search

Blob Test Depth First Search

BFS Implementation

Breath First Search

Blob Test Breath First Search

AStar Implementation

Blob Test AStar

Edge and Area Detection

Blob Area and Edge

Image Preprocessing

Blob image isolation, smoothing and clustering