Remote Sensing Portfolio

ENVI Software was utilized for Imagery below

Change Detection

Final Lab_ Change Detection (1).pdf

Change Detection in Washington DC

Here is my Final Report for my Remote Sensing: Digital Image Processing and Analysis class. Using two Landsat 5 images from 1991 and 1999, I conducted a Change Detection using: Write Function Memory Insertion, Supervised Image Classification, and a Change Detection Matrix.

This change detection makes areas undergoing Urbanization far easier to visualize. Land Use Changes, with these methods can be quantified to see how much urbanization has undergone in a 9-year time frame.

Write Function Memory Insertion

Supervised Classification

Change Detection for Specific Classification Types

Final Change Detection

Warped Image (to start)

Color Composite Comparing a Base and Warped Image

Filtering and Stretching for Image Processing

Low Pass Filter
7x7 Kernel

Emphasizes large homogeneous areas

High Pass Filter
7x7 Kernel

Edges and Boundaries are emphasized. Larger areas are deemphasized.

Laplacian Filter

Meant to highlight points, lines, and edges in imagery

Directional Filter with Angle Value of 135

Also used to detect edges and enhances linear features oriented in a specific direction

Roberts Filter

Similar to Laplacian, used for edge detection.

FFT Forward & Circular Pass/Cut Filters

FFT Inverse with Low Pass Filter

FFT Inverse with High Pass Filter

PC1, PC2, and PC3 RGB Composite Image

Mosaic Image using Nearest Neighbor Resampling

PCA Bands

Classification

Supervised

Maximum Likelihood Classification

Minimum Distance Classification

Unsupervised

ISODATA Algorithm

Unsupervised Classification - Intro Remote Sensing

PCA ISODATA

NDVI

NDVI Color Slice Classification

Using the NDVI layer, a new Raster Color Slice was created to create a new form of Unsupervised Classification. The NDVI ranges I selected for each class were: Water (-0.202971 to -0.025000), High Density Urban (-0.025000 to -0.045000), Low Density Urban (0.045000 to 0.14500), Dark Vegetation (0.145500 to 0.300000), Very Green Vegetation (0.300000 to 0.621503).

Page updated

Report abuse