In this lab, we explored the application of spatial enhancements, which serve as a critical preprocessing step for preparing imagery. We focused on understanding the properties of multispectral imagery and learned how to create band indices, essential for analyzing and interpreting the data effectively.
In exercise 1 we learned how to acquire satellite data by utilizing a web page (http://glovis.usgs.goc/) to identify and download the correct image.
In exercise 2, the focus was on spatial enhancement, which plays a crucial role in improving the interpretation of remote sensing data by uncovering hidden features and correcting sensor inaccuracies. The lab involved the use of ArcGIS Pro to implement various spatial and spectral enhancement techniques on a satellite image of the Olympic Mountains' Forest. A significant aspect of this exercise was the calculation of the NDVI index, a method employed to evaluate vegetation health.
Moreover, spatial enhancement involves filtering images using surrounding pixel values through a kernel to adjust digital number values. The lab covers three spatial filters: low pass filters, which smooth images by allowing low-frequency data to pass while blurring finer details; high pass filters, which enhance edges by allowing high-frequency data to pass; and range filters. Each method serves distinct purposes in image processing, enhancing different features for effective analysis. We also used the Convolution tool from the rater function to enhance and sharpen the image.
Exercise 3 introduced histograms, which are statistical graphs used to display the frequency distributions of cell values in image data, enhancing visualization and interpretation. Histograms can be accessed via the Symbology tab in a raster layer.
Exercise 4 focused on spectral bands and their combinations for feature identification, covering combinations such as TM True Color, TM False Color IR, and TM False Natural Color. We also created various spectral indices, including NDVI and MNDWI, to improve the visualization of certain features through specific band calculations.
Lastly, we are tasked to find three different features based on different given criteria. One of these features to identify where certain areas of water, layers 1-3 (blue, green, red) become much brighter than normal, layer 4 (near infrared) becomes somewhat brighter, and layers 5-6 (SWIR1 and SWIR2) remain unchanged. I used the stretch symbology for layers 1-3 (blue, green, red) this allow me to see the brightness in certain areas of water. This spectral response occurs due to a high concentration of suspended sediments or heavy turbidity in the water. While clear water lets visible light penetrate deeply, suspended particles (like silt, clay, sand, or organic debris) strongly scatter light back to the satellite sensor. Furthermore, examining the grayscale for Layer 4 (NIR) shows a slight increase in brightness; clear water absorbs NIR light, appearing black in standard imagery, but moderate to high sediment concentrations allow some NIR light to scatter back before full absorption. In contrast, examining the grayscale for Layers 5–6 (SWIR1 and SWIR2) they remain unchanged, as the high absorption coefficient of water in the shortwave infrared spectrum effectively reduces the signal, leaving these bands dark and unaffected even in turbid conditions.
