Скачать с ютуб Thresholding ROI for Spectral Signature Collection & Classification Using Maximum Likelihood in ENVI в хорошем качестве

Thresholding ROI for Spectral Signature Collection & Classification Using Maximum Likelihood in ENVI

The purpose of this video is to teach how to classify the most common types of land cover in a Landsat 8 satellite image. Land cover refers to the different types of physical surfaces on the earth, like forests, water bodies, urban areas, and agricultural fields. To identify and categorize these surfaces, we’ll use various tools available in ENVI software. We will use tools like the Band Animation Tool and the Regions of Interest (ROI) Tool. The Band Animation Tool helps us visualize different spectral bands (like visible, infrared, etc.) captured by the satellite. By looking at these bands, we can better understand how different land cover types reflect light in various ways. The Regions of Interest (ROI) Tool allows us to manually select specific areas within the image that represent different land cover types (such as a patch of forest or a river). We use a technique called thresholding within the ROI Tool to identify and collect the spectral signatures of these land cover types. Thresholding helps us isolate and focus on specific ranges of pixel values that correspond to a particular land cover type, making the training data more accurate. Once we have gathered the training data using thresholding, we proceed with the classification using the Maximum Likelihood technique. This method calculates the probability of each pixel belonging to a particular land cover class based on the spectral signature data we collected. The pixel is then assigned to the class with the highest probability, resulting in a detailed map of classified land cover types. By combining these tools and methods, we can accurately identify and classify different land cover types across the Landsat 8 scene.