Module 5: Unsupervised & Supervised Classification

 

        This week's lab had us utilizing ERDAS imagine again to use satellite imagery and classify the spectral signatures based on chosen pixel groups in order to minimize storage space, minimize attribute table data, and maximize important spectral signatures.

    Unsupervised classification, in my own opinion, was more tedious when it came to collecting areas of interest and assigning pixel signatures, but ultimately, simpler in carrying out its intended use. Supervised classification basically did everything for you, but for whatever reason, was difficult to yield the results I was looking for on my own. 

    I fully intend on using unsupervised classification in my final project, as it has demonstrated the ability to not only highlight the spectral signatures I'm looking for, but the ability to quantify them based on their area is an invaluable skill. 

    I look forward to using it in the coming days.